Modern Control Systems Thirteenth Edition Solutions

Instructor's Solutions Manual For Modern Control Systems, 12th Ed Systems Edition

User Manual: Pdf

Open the PDF directly: View PDF .
Page Count: 754

          © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  MODERN CONTROL SYSTEMS SOLUTION MANUAL  Richard C. Dorf  Robert H. Bishop  University of California, Davis  Marquette University  A companion to MODERN CONTROL SYSTEMS TWELFTH EDITION Richard C. Dorf Robert H. Bishop  Prentice Hall Upper Saddle River Boston Columbus San Francisco New York Indianapolis London Toronto Sydney Singapore Tokyo Montreal Dubai Madrid Hong Kong Mexico City Munich Paris Amsterdam Cape Town  Educator Home | eLearning & Assessment | Support/Contact Us | Find your rep | Exam copy bookbag  Instructor's Solutions Manual for Modern Control Systems, 12/E Richard C. Dorf, University of California, Davis Robert H. Bishop, University of Texas at Austin ISBN-10: 013602498X ISBN-13: 9780136024989 Publisher: Prentice Hall Copyright: 2011 Format: On-line Supplement Published: 08/16/2010  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  P R E F A C E  In each chapter, there are five problem types: Exercises Problems Advanced Problems Design Problems/Continuous Design Problem Computer Problems In total, there are over 1000 problems. The abundance of problems of increasing complexity gives students confidence in their problem-solving ability as they work their way from the exercises to the design and computer-based problems. It is assumed that instructors (and students) have access to MATLAB and the Control System Toolbox or to LabVIEW and the MathScript RT Module. All of the computer solutions in this Solution Manual were developed and tested on an Apple MacBook Pro platform using MATLAB 7.6 Release 2008a and the Control System Toolbox Version 8.1 and LabVIEW 2009. It is not possible to verify each solution on all the available computer platforms that are compatible with MATLAB and LabVIEW MathScript RT Module. Please forward any incompatibilities you encounter with the scripts to Prof. Bishop at the email address given below. The authors and the staff at Prentice Hall would like to establish an open line of communication with the instructors using Modern Control Systems. We encourage you to contact Prentice Hall with comments and suggestions for this and future editions. Robert H. Bishop  rhbishop@marquette.edu  iii  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  T A B L E - O F - C O N T E N T S  1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13.  iv  Introduction to Control Systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 Mathematical Models of Systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22 State Variable Models . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 85 Feedback Control System Characteristics . . . . . . . . . . . . . . . . . . . . . . . 133 The Performance of Feedback Control Systems . . . . . . . . . . . . . . . . . 177 The Stability of Linear Feedback Systems . . . . . . . . . . . . . . . . . . . . . . 234 The Root Locus Method . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 277 Frequency Response Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 382 Stability in the Frequency Domain . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 445 The Design of Feedback Control Systems . . . . . . . . . . . . . . . . . . . . . . . 519 The Design of State Variable Feedback Systems . . . . . . . . . . . . . . . . 600 Robust Control Systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 659 Digital Control Systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 714  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  C H A P T E R  1  Introduction to Control Systems  There are, in general, no unique solutions to the following exercises and problems. Other equally valid block diagrams may be submitted by the student.  Exercises E1.1  A microprocessor controlled laser system: Controller  Desired power output  Error  -  Microprocessor  Current i(t)  Laser  Power Sensor  power  A driver controlled cruise control system: Controller  Process  Foot pedal Desired speed  Power out  Measurement  Measured  E1.2  Process  -  Driver  Car and Engine  Actual auto speed  Measurement  Visual indication of speed  E1.3  Speedometer  Although the principle of conservation of momentum explains much of the process of fly-casting, there does not exist a comprehensive scientific explanation of how a fly-fisher uses the small backward and forward motion of the fly rod to cast an almost weightless fly lure long distances (the 1  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  2  CHAPTER 1  Introduction to Control Systems  current world-record is 236 ft). The fly lure is attached to a short invisible leader about 15-ft long, which is in turn attached to a longer and thicker Dacron line. The objective is cast the fly lure to a distant spot with deadeye accuracy so that the thicker part of the line touches the water first and then the fly gently settles on the water just as an insect might. Fly-fisher Desired position of the fly  Controller  -  Wind disturbance  Mind and body of the fly-fisher  Process  Rod, line, and cast  Actual position of the fly  Measurement  Visual indication of the position of the fly  E1.4  Vision of the fly-fisher  An autofocus camera control system: One-way trip time for the beam  Conversion factor (speed of light or sound)  K1 Beam Emitter/ Receiver Beam return  Distance to subject  Subject Lens focusing motor  Lens  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  3  Exercises  E1.5  Tacking a sailboat as the wind shifts:  Error  Desired sailboat direction  -  Controller  Actuators  Sailor  Rudder and sail adjustment  Wind  Process  Sailboat  Actual sailboat direction  Measurement Measured sailboat direction  Gyro compass  E1.6  An automated highway control system merging two lanes of traffic: Controller  Error  Desired gap  -  Embedded computer  Actuators  Brakes, gas or steering  Process  Active vehicle  Actual gap  Measurement Measured gap  Radar  E1.7  Using the speedometer, the driver calculates the difference between the measured speed and the desired speed. The driver throotle knob or the brakes as necessary to adjust the speed. If the current speed is not too much over the desired speed, the driver may let friction and gravity slow the motorcycle down. Controller  Desired speed  Error  -  Driver  Actuators  Throttle or brakes  Measurement Visual indication of speed  Speedometer  Process  Motorcycle  Actual motorcycle speed  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  4  CHAPTER 1  E1.8  Introduction to Control Systems  Human biofeedback control system: Controller  Desired body temp  Process  Hypothalumus  -  Message to blood vessels  Actual body temp  Human body  Measurement Visual indication of body temperature  E1.9  TV display  Body sensor  E-enabled aircraft with ground-based flight path control: Corrections to the flight path  Desired Flight Path  -  Controller  Aircraft  Gc(s)  G(s)  Flight Path Health Parameters  Meteorological data  Location and speed  Optimal flight path  Ground-Based Computer Network Optimal flight path Meteorological data  Desired Flight Path  E1.10  Specified Flight Trajectory  Health Parameters  Corrections to the flight path  Gc(s)  G(s)  Controller  Aircraft  Location and speed  Flight Path  Unmanned aerial vehicle used for crop monitoring in an autonomous mode: Trajectory error  -  Controller  UAV  Gc(s)  G(s)  Flight Trajectory  Sensor Location with respect to the ground  Map Correlation Algorithm  Ground photo  Camera  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  5  Exercises  E1.11  An inverted pendulum control system using an optical encoder to measure the angle of the pendulum and a motor producing a control torque: Actuator  Voltage  Error  Desired angle  -  Controller  Process  Torque  Motor  Pendulum  Angle  Measurement  Measured angle  E1.12  In the video game, the player can serve as both the controller and the sensor. The objective of the game might be to drive a car along a prescribed path. The player controls the car trajectory using the joystick using the visual queues from the game displayed on the computer monitor. Controller  Desired game objective  Optical encoder  Error  -  Player  Actuator  Joystick  Measurement  Player (eyesight, tactile, etc.)  Process  Video game  Game objective  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  6  CHAPTER 1  Introduction to Control Systems  Problems P1.1  Desired temperature set by the driver  An automobile interior cabin temperature control system block diagram:  Error  -  Controller  Process  Thermostat and air conditioning unit  Automobile cabin  Automobile cabin temperature  Measurement Measured temperature  P1.2  Temperature sensor  A human operator controlled valve system: Controller  Process  Error *  Desired fluid output *  -  Tank  Valve  Fluid output  Measurement Visual indication of fluid output *  Meter * = operator functions  P1.3  A chemical composition control block diagram: Controller  Process  Error Desired chemical composition  -  Mixer tube  Valve  Measurement Measured chemical composition  Infrared analyzer  Chemical composition  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  7  Problems  P1.4  A nuclear reactor control block diagram: Controller  Process  Error Desired power level  Reactor and rods  Motor and amplifier  -  Output power level  Measurement Measured chemical composition  P1.5  A light seeking control system to track the sun:  Measurement  Light source  Dual Photocells  P1.6  Ionization chamber  Controller  Ligh intensity  Trajectory Planner  Desired carriage position  Controller  -  Motor, carriage, and gears  K  Photocell carriage position  If you assume that increasing worker's wages results in increased prices, then by delaying or falsifying cost-of-living data you could reduce or eliminate the pressure to increase worker's wages, thus stabilizing prices. This would work only if there were no other factors forcing the cost-of-living up. Government price and wage economic guidelines would take the place of additional "controllers" in the block diagram, as shown in the block diagram. Controller  Process Market-based prices  Initial wages  Process  Motor inputs  Error  -  Industry  Government price guidelines  Controller  Wage increases  Government wage guidelines  Cost-of-living  K1  Prices  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  8  CHAPTER 1  P1.7  Introduction to Control Systems  Assume that the cannon fires initially at exactly 5:00 p.m.. We have a positive feedback system. Denote by ∆t the time lost per day, and the net time error by ET . Then the follwoing relationships hold: ∆t = 4/3 min. + 3 min. = 13/3 min. and ET = 12 days × 13/3 min./day . Therefore, the net time error after 15 days is ET = 52 min.  P1.8  The student-teacher learning process: Process  Controller  Lectures  Error Desired knowledge  -  Teacher  Knowledge  Student  Measurement  Exams  Measured knowledge  P1.9  A human arm control system: Process  Controller u Desired arm location  e  y  s Brain  Nerve signals  z Measurement  Visual indication of arm location  Pressure Eyes and pressure receptors  Arm & muscles  d  Arm location  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  9  Problems  P1.10  An aircraft flight path control system using GPS: Controller  Desired flight path from air traffic controllers  Actuators  Computer Auto-pilot  Error  -  Process  Ailerons, elevators, rudder, and engine power  Flight path  Aircraft  Measurement Measured flight path  P1.11  The accuracy of the clock is dependent upon a constant flow from the orifice; the flow is dependent upon the height of the water in the float tank. The height of the water is controlled by the float. The control system controls only the height of the water. Any errors due to enlargement of the orifice or evaporation of the water in the lower tank is not accounted for. The control system can be seen as:  Desired height of the water in float tank  P1.12  Global Positioning System  -  Controller  Process  Float level  Flow from upper tank to float tank  Actual height  Assume that the turret and fantail are at 90◦ , if θw 6= θF -90◦ . The fantail operates on the error signal θw - θT , and as the fantail turns, it drives the turret to turn.  y  Wind  qW = Wind angle qF = Fantail angle qT = Turret angle  Controller  *  qW qF qT  qW  *  Turret  x  -  Process Torque  Error  Fantail  Fantail  Gears & turret  qT  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  10  CHAPTER 1  P1.13  Introduction to Control Systems  This scheme assumes the person adjusts the hot water for temperature control, and then adjusts the cold water for flow rate control. Controller  Error  Desired water temperature  Process  Hot water system  Valve adjust  -  Hot water  Actual water temperature and flow rate Desired water flow rate  Cold water system  Valve adjust  -  Cold water  Measurement  Measured water flow Measured water temperature  P1.14  Human: visual and touch  If the rewards in a specific trade is greater than the average reward, there is a positive influx of workers, since q(t) = f1 (c(t) − r(t)). If an influx of workers occurs, then reward in specific trade decreases, since c(t) = −f2 (q(t)). Controller  Average rewards r(t)  P1.15  Desired Fuel Pressure  Error  -  f1(c(t)-r(t))  Process q(t)  - f2(q(t))  Total of rewards c(t)  A computer controlled fuel injection system:  -  Controller  Process  Electronic Control Unit  High Pressure Fuel Supply Pump and Electronic Fuel Injectors  Measurement Measured fuel pressure  Fuel Pressure Sensor  Fuel Pressure  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  11  Problems  P1.16  With the onset of a fever, the body thermostat is turned up. The body adjusts by shivering and less blood flows to the skin surface. Aspirin acts to lowers the thermal set-point in the brain. Controller  Desired temperature or set-point from body thermostat in the brain  Process  Adjustments within the body  -  Body temperature  Body  Measurement Measured body temperature  Internal sensor  P1.17  Hitting a baseball is arguably one of the most difficult feats in all of sports. Given that pitchers may throw the ball at speeds of 90 mph (or higher!), batters have only about 0.1 second to make the decision to swing—with bat speeds aproaching 90 mph. The key to hitting a baseball a long distance is to make contact with the ball with a high bat velocity. This is more important than the bat's weight, which is usually around 33 ounces (compared to Ty Cobb's bat which was 41 ounces!). Since the pitcher can throw a variety of pitches (fast ball, curve ball, slider, etc.), a batter must decide if the ball is going to enter the strike zone and if possible, decide the type of pitch. The batter uses his/her vision as the sensor in the feedback loop. A high degree of eye-hand coordination is key to success—that is, an accurate feedback control system.  P1.18  Define the following variables: p = output pressure, fs = spring force = Kx, fd = diaphragm force = Ap, and fv = valve force = fs - fd . The motion of the valve is described by ÿ = fv /m where m is the valve mass. The output pressure is proportional to the valve displacement, thus p = cy , where c is the constant of proportionality.  Constant of proportionality  Spring  Screw displacement x(t)  K  fs  -  Valve position  fv  Valve  c  y  Diaphragm area  fd  A  Output pressure p(t)  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  12  CHAPTER 1  P1.19  Introduction to Control Systems  A control system to keep a car at a given relative position offset from a lead car:  Throttle  Position of follower  Follower car  Actuator  u  -  Controller  Relative position  -  Position of lead  Lead car  Fuel throttle (fuel)  Video camera & processing algorithms  Reference photo  Desired relative position  P1.20  A control system for a high-performance car with an adjustable wing:  Desired road adhesion  -  Process  Actuator  Controller  Computer  Adjustable wing  Road conditions  Race Car  Road adhesion  Measurement  Measured road adhesion  P1.21  K  Tire internal strain gauges  A control system for a twin-lift helicopter system: Measurement Measured separation distance  Desired separation distance  -  Controller  Process Separation distance  Pilot Desired altitude  Radar  Helicopter Altitude  Measurement Measured altitude  Altimeter  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  13  Problems  P1.22  The desired building deflection would not necessarily be zero. Rather it would be prescribed so that the building is allowed moderate movement up to a point, and then active control is applied if the movement is larger than some predetermined amount. Process Controller  Desired deflection  Hydraulic stiffeners  -  Building  Deflection  Measurement  Measured deflection  P1.23  Strain gauges on truss structure  K  The human-like face of the robot might have micro-actuators placed at strategic points on the interior of the malleable facial structure. Cooperative control of the micro-actuators would then enable the robot to achieve various facial expressions. Controller  Process  Error Desired actuator position  -  Voltage  Electromechanical actuator  Amplifier  Actuator position  Measurement  Position sensor  Measured position  P1.24  We might envision a sensor embedded in a "gutter" at the base of the windshield which measures water levels—higher water levels corresponds to higher intensity rain. This information would be used to modulate the wiper blade speed. Process  Controller  Desired wiper speed  Wiper blade and motor  Electronic Control Unit  -  Measurement  K  Measured water level  Water depth sensor  Wiper blade speed  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  14  CHAPTER 1  Introduction to Control Systems  A feedback control system for the space traffic control:  P1.25  Controller  Error  Desired orbit position  Control law  -  Actuator Jet commands  Process Applied forces  Reaction control jets  Satellite  Actual  orbit position  Measurement Measured orbit position  Radar or GPS  Earth-based control of a microrover to point the camera:  P1.26  Microrover Camera position command  Receiver/ Transmitter  Controller  G(s)  Gc(s)  Rover position  Camera  Camera Position  m Ca ap er  Sensor  ea  iti os  M  Measured camera position  on  d re su  d an  m m co  ap er  m ca  on  iti os  P1.27  Desired Charge Level  Control of a methanol fuel cell:  -  Controller  Recharging System  Gc(s)  GR(s)  Methanol water solution  G(s) Sensor  Measured charge level  Fuel Cell  H(s)  Charge Level  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  15  Advanced Problems  Advanced Problems AP1.1  Control of a robotic microsurgical device:  Microsurgical robotic manipulator  Controller Desired End-effector Position  -  G(s)  Gc(s)  End-effector Position  Sensor  H(s)  AP1.2  An advanced wind energy system viewed as a mechatronic system: AERODYNAMIC DESIGN STRUCTURAL DESIGN OF THE TOWER ELECTRICAL AND POWER SYSTEMS  SENSORS Rotor rotational sensor Wind speed and direction sensor ACTUATORS Motors for manipulatiing the propeller pitch  Physical System Modeling  CONTROL SYSTEM DESIGN AND ANALYSIS ELECTRICAL SYSTEM DESIGN AND ANALYSIS POWER GENERATION AND STORAGE  Sensors and Actuators WIND ENERGY SYSTEM  Software and Data Acquisition  CONTROLLER ALGORITHMS DATA ACQUISTION: WIND SPEED AND DIRECTION ROTOR ANGULAR SPEED PROPELLOR PITCH ANGLE  AP1.3  Signals and Systems  Computers and Logic Systems  COMPUTER EQUIPMENT FOR CONTROLLING THE SYSTEM SAFETY MONITORING SYSTEMS  The automatic parallel parking system might use multiple ultrasound sensors to measure distances to the parked automobiles and the curb. The sensor measurements would be processed by an on-board computer to determine the steering wheel, accelerator, and brake inputs to avoid collision and to properly align the vehicle in the desired space.  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  16  CHAPTER 1  Introduction to Control Systems  Even though the sensors may accurately measure the distance between the two parked vehicles, there will be a problem if the available space is not big enough to accommodate the parking car. Controller  Desired automobile position  Error  Actuators  On-board computer  -  Steering wheel, accelerator, and brake  Process  Actual automobile position  Automobile  Measurement  Position of automobile relative to parked cars and curb  Ultrasound  There are various control methods that can be considered, including placing the controller in the feedforward loop (as in Figure 1.3). The adaptive optics block diagram below shows the controller in the feedback loop, as an alternative control system architecture.  AP1.4  Process  Astronomical object Uncompensated image  Astronomical telescope mirror  Compensated image  Measurement  Wavefront reconstructor  Wavefront corrector  Wavefront sensor  Actuator & controller  AP1.5  Desired floor  Error  -  The control system might have an inner loop for controlling the acceleration and an outer loop to reach the desired floor level precisely.  Controller #2  Outer Loop  Desired acceleration  Error  -  Controller #1  Elevator motor, cables, etc.  Inner Loop Measured acceleration  Acceleration Measurement  Elevator  Floor  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  17  Advanced Problems  An obstacle avoidance control system would keep the robotic vacuum cleaner from colliding with furniture but it would not necessarily put the vacuum cleaner on an optimal path to reach the entire floor. This would require another sensor to measure position in the room, a digital map of the room layout, and a control system in the outer loop.  AP1.6  Process Desired distance from obstacles  Error  -  Controller  Measured distance from obstacle  Motors, wheels, etc.  Infrared sensors  Robotic vacuum cleaner  Distance from obstacles  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  18  CHAPTER 1  Introduction to Control Systems  Design Problems CDP1.1  The machine tool with the movable table in a feedback control configuration: Controller  Error  Desired position x  Amplifier  -  Actuator  Process  Machine tool with table  Positioning motor  Actual position x  Measurement  Position sensor  Measured position  DP1.1  Use the stereo system and amplifiers to cancel out the noise by emitting signals 180◦ out of phase with the noise. Process  Controller Noise signal Desired noise = 0  Shift phase by 180 deg  -  Machine tool with table  Positioning motor  Noise in cabin  Measurement  Microphone  DP1.2  Desired speed of auto set by driver  1/K  An automobile cruise control system: Controller  Desired shaft speed  -  Electric motor  Process  Automobile and engine  Valve  Measurement  Measured shaft speed  Shaft speed sensor  Drive shaf t speed  K  Actual speed of auto  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  19  Design Problems  DP1.3  An automoted cow milking system: Measurement Cow location  Vision system  Motor and gears  -  Desired cup location  Process  Actuator  Controller  Location of cup  Robot arm and cup gripper  Cow and milker  Milk  Measurement  Vision system  Measured cup location  DP1.4  A feedback control system for a robot welder: Controller  Desired position  Process  Computer and amplifier  Error  -  Voltage  Motor and arm  Weld top position  Measurement  Vision camera  Measured position  DP1.5  A control system for one wheel of a traction control system: Antislip controller  Engine torque  +  -  Wheel dynamics  +  -  Wheel speed  Sensor  + Actual slip  1/Rw  Vehicle dynamics  Brake torque  +  Vehicle speed  Antiskid controller  Rw = Radius of wheel  Sensor  Measured slip  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  20  CHAPTER 1  Introduction to Control Systems  A vibration damping system for the Hubble Space Telescope:  DP1.6  Controller Desired jitter = 0  Error  Computer  -  Actuators  Gyro and reaction wheels  Process Signal to cancel the jitter  Spacecraft dynamics  Jitter of vibration  Measurement  Measurement of 0.05 Hz jitter  DP1.7  A control system for a nanorobot: Controller  Desired nanorobot position  Rate gyro sensor  Error  -  Biocomputer  Actuators  Plane surfaces and propellers  Process  Nanorobot  Actual nanorobot position  Measurement  External beacons  Many concepts from underwater robotics can be applied to nanorobotics within the bloodstream. For example, plane surfaces and propellers can provide the required actuation with screw drives providing the propulsion. The nanorobots can use signals from beacons located outside the skin as sensors to determine their position. The nanorobots use energy from the chemical reaction of oxygen and glucose available in the human body. The control system requires a bio-computer–an innovation that is not yet available. For further reading, see A. Cavalcanti, L. Rosen, L. C. Kretly, M. Rosenfeld, and S. Einav, "Nanorobotic Challenges n Biomedical Application, Design, and Control," IEEE ICECS Intl Conf. on Electronics, Circuits and Systems, Tel-Aviv, Israel, December 2004. DP1.8  The feedback control system might use gyros and/or accelerometers to measure angle change and assuming the HTV was originally in the vertical position, the feedback would retain the vertical position using commands to motors and other actuators that produced torques and could move the HTV forward and backward.  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  21  Design Problems  Process Desired angle from vertical (0o)  Error  -  Controller  Measured angle from vertical  Motors, wheels, etc.  Gyros & accelerometers  HTV  Angle from vertical  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  C H A P T E R  2  Mathematical Models of Systems  Exercises E2.1  We have for the open-loop y = r2 and for the closed-loop e = r − y and y = e2 . So, e = r − e2 and e2 + e − r = 0 . 16  14  12  y  10  8  open-loop  6  4  closed-loop  2  0  0  0.5  1  1.5  2 r  2.5  3  3.5  4  FIGURE E2.1 Plot of open-loop versus closed-loop.  For example, if r = 1, then e2 + e − 1 = 0 implies that e = 0.618. Thus, y = 0.382. A plot y versus r is shown in Figure E2.1. 22  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  23  Exercises  E2.2  Define f (T ) = R = R0 e−0.1T and ∆R = f (T ) − f (T0 ) , ∆T = T − T0 . Then, ∆R = f (T ) − f (T0 ) =  ∂f ∂T  T =T0 =20◦  ∆T + · · ·  where ∂f ∂T  T =T0 =20◦  = −0.1R0 e−0.1T0 = −135,  when R0 = 10, 000Ω. Thus, the linear approximation is computed by considering only the first-order terms in the Taylor series expansion, and is given by ∆R = −135∆T . The spring constant for the equilibrium point is found graphically by estimating the slope of a line tangent to the force versus displacement curve at the point y = 0.5cm, see Figure E2.3. The slope of the line is K ≈ 1. 2 1.5 Spring breaks  1 0.5 0  Force (n)  E2.3  -0.5 -1 -1.5 -2 -2.5 -3 -2  Spring compresses -1.5  -1  -0.5  0  0.5  1  y=Displacement (cm)  FIGURE E2.3 Spring force as a function of displacement.  1.5  2  2.5  3  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  24  CHAPTER 2  E2.4  Mathematical Models of Systems  Since R(s) =  1 s  we have Y (s) =  4(s + 50) . s(s + 20)(s + 10)  The partial fraction expansion of Y (s) is given by Y (s) =  A1 A2 A3 + + s s + 20 s + 10  where A1 = 1 , A2 = 0.6 and A3 = −1.6 . Using the Laplace transform table, we find that y(t) = 1 + 0.6e−20t − 1.6e−10t . The final value is computed using the final value theorem: 4(s + 50) lim y(t) = lim s =1. 2 t→∞ s→0 s(s + 30s + 200)   E2.5    The circuit diagram is shown in Figure E2.5. R2  v+  A +  vin -  FIGURE E2.5 Noninverting op-amp circuit.  With an ideal op-amp, we have vo = A(vin − v − ),  + v0 -  R1  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  25  Exercises  where A is very large. We have the relationship R1 vo . R1 + R2  v− = Therefore,  vo = A(vin −  R1 vo ), R1 + R2  and solving for vo yields vo =  A 1+  AR1 R1 +R2  1 Since A ≫ 1, it follows that 1 + RAR ≈ 1 +R2 vo simplifies to  vo = E2.6  vin .  AR1 R1 +R2 .  Then the expression for  R1 + R2 vin . R1  Given y = f (x) = ex and the operating point xo = 1, we have the linear approximation y = f (x) = f (xo ) +  ∂f ∂x  x=xo  (x − xo ) + · · ·  where df dx  f (xo ) = e,  = e, x=xo =1  and x − xo = x − 1.  Therefore, we obtain the linear approximation y = ex. E2.7  The block diagram is shown in Figure E2.7.  R(s)  Ea(s)  +  G1(s)  G2(s)  -  H(s) FIGURE E2.7 Block diagram model.  I(s)  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  26  CHAPTER 2  Mathematical Models of Systems  Starting at the output we obtain I(s) = G1 (s)G2 (s)E(s). But E(s) = R(s) − H(s)I(s), so I(s) = G1 (s)G2 (s) [R(s) − H(s)I(s)] . Solving for I(s) yields the closed-loop transfer function G1 (s)G2 (s) I(s) = . R(s) 1 + G1 (s)G2 (s)H(s) E2.8  The block diagram is shown in Figure E2.8. H2(s) -  R(s)  K -  E(s)  -  G1(s)  W(s) -  A(s)  G2(s)  Z(s)  1 s  Y(s)  H3(s)  H1(s)  FIGURE E2.8 Block diagram model.  Starting at the output we obtain Y (s) =  1 1 Z(s) = G2 (s)A(s). s s  But A(s) = G1 (s) [−H2 (s)Z(s) − H3 (s)A(s) + W (s)] and Z(s) = sY (s), so 1 Y (s) = −G1 (s)G2 (s)H2 (s)Y (s) − G1 (s)H3 (s)Y (s) + G1 (s)G2 (s)W (s). s Substituting W (s) = KE(s) − H1 (s)Z(s) into the above equation yields Y (s) = −G1 (s)G2 (s)H2 (s)Y (s) − G1 (s)H3 (s)Y (s) 1 + G1 (s)G2 (s) [KE(s) − H1 (s)Z(s)] s  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  27  Exercises  and with E(s) = R(s) − Y (s) and Z(s) = sY (s) this reduces to Y (s) = [−G1 (s)G2 (s) (H2 (s) + H1 (s)) − G1 (s)H3 (s) 1 1 − G1 (s)G2 (s)K]Y (s) + G1 (s)G2 (s)KR(s). s s Solving for Y (s) yields the transfer function Y (s) = T (s)R(s), where T (s) = E2.9  KG1 (s)G2 (s)/s . 1 + G1 (s)G2 (s) [(H2 (s) + H1 (s)] + G1 (s)H3 (s) + KG1 (s)G2 (s)/s  From Figure E2.9, we observe that Ff (s) = G2 (s)U (s) and FR (s) = G3 (s)U (s) . Then, solving for U (s) yields U (s) =  1 Ff (s) G2 (s)  FR (s) =  G3 (s) U (s) . G2 (s)  and it follows that  Again, considering the block diagram in Figure E2.9 we determine Ff (s) = G1 (s)G2 (s)[R(s) − H2 (s)Ff (s) − H2 (s)FR (s)] . But, from the previous result, we substitute for FR (s) resulting in Ff (s) = G1 (s)G2 (s)R(s)−G1 (s)G2 (s)H2 (s)Ff (s)−G1 (s)H2 (s)G3 (s)Ff (s) . Solving for Ff (s) yields G1 (s)G2 (s) Ff (s) = R(s) . 1 + G1 (s)G2 (s)H2 (s) + G1 (s)G3 (s)H2 (s)     © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  28  CHAPTER 2  Mathematical Models of Systems  H2(s) + -  R(s)  U(s)  G2(s)  Ff (s)  U(s)  G3(s)  FR(s)  G1(s)  -  H2(s) FIGURE E2.9 Block diagram model.  E2.10  The shock absorber block diagram is shown in Figure E2.10. The closedloop transfer function model is T (s) =  Gc (s)Gp (s)G(s) . 1 + H(s)Gc (s)Gp (s)G(s)  Controller  Gear Motor  Plunger and Piston System  Gc(s)  Gp(s)  G(s)  + R(s) Desired piston travel  -  Y(s) Piston travel  Sensor  H(s)  Piston travel measurement  FIGURE E2.10 Shock absorber block diagram.  E2.11  Let f denote the spring force (n) and x denote the deflection (m). Then K=  ∆f . ∆x  Computing the slope from the graph yields: (a) xo = −0.14m → K = ∆f /∆x = 10 n / 0.04 m = 250 n/m (b) xo = 0m → K = ∆f /∆x = 10 n / 0.05 m = 200 n/m (c) xo = 0.35m → K = ∆f /∆x = 3n / 0.05 m = 60 n/m  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  29  Exercises  E2.12  The signal flow graph is shown in Fig. E2.12. Find Y (s) when R(s) = 0.  -K  Td(s) 1  1 K2  G(s) Y (s)  -1 FIGURE E2.12 Signal flow graph.  The transfer function from Td (s) to Y (s) is Y (s) =  G(s)(1 − K1 K2 )Td (s) G(s)Td (s) − K1 K2 G(s)Td (s) = . 1 − (−K2 G(s)) 1 + K2 G(s)  If we set K1 K2 = 1 , then Y (s) = 0 for any Td (s). E2.13  The transfer function from R(s), Td (s), and N (s) to Y (s) is K K 1 R(s)+ 2 Td (s)− 2 N (s) Y (s) = 2 s + 10s + K s + 10s + K s + 10s + K             Therefore, we find that Y (s)/Td (s) = E2.14  s2  1 + 10s + K  and  Y (s)/N (s) = −  s2  K + 10s + K  Since we want to compute the transfer function from R2 (s) to Y1 (s), we can assume that R1 = 0 (application of the principle of superposition). Then, starting at the output Y1 (s) we obtain Y1 (s) = G3 (s) [−H1 (s)Y1 (s) + G2 (s)G8 (s)W (s) + G9 (s)W (s)] , or [1 + G3 (s)H1 (s)] Y1 (s) = [G3 (s)G2 (s)G8 (s)W (s) + G3 (s)G9 (s)] W (s). Considering the signal W (s) (see Figure E2.14), we determine that W (s) = G5 (s) [G4 (s)R2 (s) − H2 (s)W (s)] ,  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  30  CHAPTER 2  Mathematical Models of Systems  H1(s)  +  G1(s)  R1(s)  +  G7(s)  R2(s)  G4(s)  -  +  G2(s)  G3(s) +  Y1(s)  G9(s)  G8(s)  +  +  G6(s)  G5(s)  Y2(s)  W(s)  -  H2(s) FIGURE E2.14 Block diagram model.  or [1 + G5 (s)H2 (s)] W (s) = G5 (s)G4 (s)R2 (s). Substituting the expression for W (s) into the above equation for Y1 (s) yields Y1 (s) G2 (s)G3 (s)G4 (s)G5 (s)G8 (s) + G3 (s)G4 (s)G5 (s)G9 (s) = . R2 (s) 1 + G3 (s)H1 (s) + G5 (s)H2 (s) + G3 (s)G5 (s)H1 (s)H2 (s) E2.15  For loop 1, we have di1 1 R1 i1 + L1 + dt C1  Z  (i1 − i2 )dt + R2 (i1 − i2 ) = v(t) .  And for loop 2, we have 1 C2 E2.16  Z  di2 1 i2 dt + L2 + R2 (i2 − i1 ) + dt C1  Z  (i2 − i1 )dt = 0 .  The transfer function from R(s) to P (s) is P (s) 4.2 = 3 . 2 R(s) s + 2s + 4s + 4.2 The block diagram is shown in Figure E2.16a. The corresponding signal flow graph is shown in Figure E2.16b for P (s)/R(s) =  s3  +  4.2 . + 4s + 4.2  2s2  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  31  Exercises  v1(s)  R(s)  v2(s)  7  -  q(s)  0.6 s  1 s2+2s+4  P(s)  (a)  R(s )  1  V1  7  1 s2 + 2 s + 4  0.6 s  V2  P (s)  -1  (b) FIGURE E2.16 (a) Block diagram, (b) Signal flow graph.  E2.17  A linear approximation for f is given by ∆f =  ∂f ∂x  ∆x = 2kxo ∆x = k∆x x=xo  where xo = 1/2, ∆f = f (x) − f (xo ), and ∆x = x − xo . E2.18  The linear approximation is given by ∆y = m∆x where m=  ∂y ∂x  . x=xo  (a) When xo = 1, we find that yo = 2.4, and yo = 13.2 when xo = 2. (b) The slope m is computed as follows: m=  ∂y ∂x  = 1 + 4.2x2o . x=xo  Therefore, m = 5.2 at xo = 1, and m = 18.8 at xo = 2.  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  32  CHAPTER 2  E2.19  Mathematical Models of Systems  The output (with a step input) is Y (s) =  15(s + 1) . s(s + 7)(s + 2)  The partial fraction expansion is 18 1 3 1 15 − + . 14s 7 s+7 2s+2  Y (s) =  Taking the inverse Laplace transform yields y(t) = E2.20  15 18 −7t 3 −2t − e + e . 14 7 2  The input-output relationship is A(K − 1) Vo = V 1 + AK where K=  Z1 . Z1 + Z2  Assume A ≫ 1. Then, Vo K−1 Z2 = =− V K Z1 where Z1 =  R1 R1 C 1 s + 1  and Z2 =  R2 . R2 C 2 s + 1  Therefore, Vo (s) R2 (R1 C1 s + 1) 2(s + 1) =− =− . V (s) R1 (R2 C2 s + 1) s+2 E2.21  The equation of motion of the mass mc is mc ẍp + (bd + bs )ẋp + kd xp = bd ẋin + kd xin . Taking the Laplace transform with zero initial conditions yields [mc s2 + (bd + bs )s + kd ]Xp (s) = [bd s + kd ]Xin (s) . So, the transfer function is bd s + kd 0.7s + 2 Xp (s) = = 2 . Xin (s) mc s2 + (bd + bs )s + kd s + 2.8s + 2  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  33  Exercises  E2.22  The rotational velocity is ω(s) =  2(s + 4) 1 . 2 (s + 5)(s + 1) s  Expanding in a partial fraction expansion yields ω(s) =  81 1 1 3 13 1 1 + − − . 2 5 s 40 s + 5 2 (s + 1) 8 s+1  Taking the inverse Laplace transform yields ω(t) = E2.23  8 1 3 13 + e−5t − te−t − e−t . 5 40 2 8  The closed-loop transfer function is Y (s) K1 K2 = T (s) = 2 . R(s) s + (K1 + K2 K3 + K1 K2 )s + K1 K2 K3  E2.24  The closed-loop tranfser function is Y (s) 10 = T (s) = 2 . R(s) s + 21s + 10  E2.25  Let x = 0.6 and y = 0.8. Then, with y = ax3 , we have 0.8 = a(0.6)3 . Solving for a yields a = 3.704. A linear approximation is y − yo = 3ax2o (x − xo ) or y = 4x − 1.6, where yo = 0.8 and xo = 0.6.  E2.26  The equations of motion are m1 ẍ1 + k(x1 − x2 ) = F m2 ẍ2 + k(x2 − x1 ) = 0 . Taking the Laplace transform (with zero initial conditions) and solving for X2 (s) yields X2 (s) =  (m2  s2  k F (s) . + k)(m1 s2 + k) − k 2  Then, with m1 = m2 = k = 1, we have X2 (s)/F (s) =  1 . s2 (s2 + 2)  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  34  CHAPTER 2  E2.27  Mathematical Models of Systems  The transfer function from Td (s) to Y (s) is Y (s)/Td (s) =  E2.28  G2 (s) . 1 + G1 G2 H(s)  The transfer function is R2 R4 C R2 R4 Vo (s) = s+ = 24s + 144 . V (s) R3 R1 R3  E2.29  (a) If G(s) =  s2  1 + 15s + 50  and  H(s) = 2s + 15 ,  then the closed-loop transfer function of Figure E2.28(a) and (b) (in Dorf & Bishop) are equivalent. (b) The closed-loop transfer function is T (s) =  (a) The closed-loop transfer function is T (s) =  G(s) 1 10 = 2 1 + G(s) s s(s + 2s + 20)  where G(s) =  0.8 0.7 0.6 Amplitude  E2.30  1 . s2 + 17s + 65  0.5 0.4 0.3 0.2 0.1 0  0  1  2  3 Time sec  FIGURE E2.30 Step response.  4  5  6  s2  10 . + 2s + 10  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  35  Exercises  (b) The output Y (s) (when R(s) = 1/s) is Y (s) =  0.5 −0.25 + 0.0573j −0.25 − 0.0573j − + . s s + 1 − 4.3589j s + 1 + 4.3589j  (c) The plot of y(t) is shown in Figure E2.30. The output is given by √ √ 1 1 y(t) = 1 − e−t cos 19t − √ sin 19t 2 19     E2.31    The partial fraction expansion is V (s) =  a b + s + p1 s + p2  where p1 = 4 − 19.6j and p2 = 4 + 19.6j. Then, the residues are a = −10.2j  b = 10.2j .  The inverse Laplace transform is v(t) = −10.2je(−4+19.6j)t + 10.2je(−4−19.6j)t = 20.4e−4t sin 19.6t .  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  36  CHAPTER 2  Mathematical Models of Systems  Problems P2.1  The integrodifferential equations, obtained by Kirchoff's voltage law to each loop, are as follows: R1 i1 +  1 C1  R3 i2 +  1 C2  Z  i1 dt + L1  d(i1 − i2 ) + R2 (i1 − i2 ) = v(t) dt  (loop 1)  and  P2.2  Z  i2 dt + R2 (i2 − i1 ) + L1  d(i2 − i1 ) =0 dt  (loop 2) .  The differential equations describing the system can be obtained by using a free-body diagram analysis of each mass. For mass 1 and 2 we have M1 ÿ1 + k12 (y1 − y2 ) + bẏ1 + k1 y1 = F (t) M2 ÿ2 + k12 (y2 − y1 ) = 0 . Using a force-current analogy, the analagous electric circuit is shown in Figure P2.2, where Ci → Mi , L1 → 1/k1 , L12 → 1/k12 , and R → 1/b .  FIGURE P2.2 Analagous electric circuit.  P2.3  The differential equations describing the system can be obtained by using a free-body diagram analysis of each mass. For mass 1 and 2 we have M ẍ1 + kx1 + k(x1 − x2 ) = F (t) M ẍ2 + k(x2 − x1 ) + bẋ2 = 0 . Using a force-current analogy, the analagous electric circuit is shown in Figure P2.3, where C→M  L → 1/k  R → 1/b .  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  37  Problems  FIGURE P2.3 Analagous electric circuit.  (a) The linear approximation around vin = 0 is vo = 0vin , see Figure P2.4(a). (b) The linear approximation around vin = 1 is vo = 2vin − 1, see Figure P2.4(b).  (a)  (b)  0.4  4  3.5  0.3  3 0.2 2.5 0.1  2 vo  vo  P2.4  0  1.5  linear approximation 1  -0.1  0.5 -0.2 0 -0.3  -0.4 -1  linear approximation  -0.5  -0.5  0 vin  0.5  FIGURE P2.4 Nonlinear functions and approximations.  1  -1 -1  0  1 vin  2  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  38  CHAPTER 2  P2.5  Mathematical Models of Systems  Given Q = K(P1 − P2 )1/2 . Let δP = P1 − P2 and δPo = operating point. Using a Taylor series expansion of Q, we have Q = Qo +  ∂Q ∂δP  (δP − δPo ) + · · ·  δP =δPo  where Qo = KδPo1/2  ∂Q ∂δP  and  = δP =δPo  K −1/2 . δP 2 o  Define ∆Q = Q − Qo and ∆P = δP − δPo . Then, dropping higher-order terms in the Taylor series expansion yields ∆Q = m∆P where m= P2.6  K 1/2  2δPo  .  From P2.1 we have R1 i1 +  1 C1  R3 i2 +  1 C2  Z  i1 dt + L1  d(i1 − i2 ) + R2 (i1 − i2 ) = v(t) dt  and Z  i2 dt + R2 (i2 − i1 ) + L1  d(i2 − i1 ) =0. dt  Taking the Laplace transform and using the fact that the initial voltage across C2 is 10v yields [R1 +  1 + L1 s + R2 ]I1 (s) + [−R2 − L1 s]I2 (s) = 0 C1 s  and [−R2 − L1 s]I1 (s) + [L1 s + R3 +  1 10 + R2 ]I2 (s) = − . C2 s s  Rewriting in matrix form we have    R1 +  1 C1 s  + L 1 s + R2  −R2 − L1 s  −R2 − L1 s L 1 s + R3 +  1 C2 s  + R2     I1 (s) I2 (s)      =  0 −10/s     .  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  39  Problems  Solving for I2 yields           1 0 R2 + L 1 s 1  L1 s + R3 + C2 s + R2  =   . 1 ∆ −10/s I2 (s) R2 + L 1 s R1 + C1 s + L1 s + R2  I1 (s)  or I2 (s) =  −10(R1 + 1/C1 s + L1 s + R2 ) s∆  where ∆ = (R1 + P2.7  1 1 + L1 s + R2 )(L1 s + R3 + + R2 ) − (R2 + L1 s)2 . C1 s C2 s  Consider the differentiating op-amp circuit in Figure P2.7. For an ideal op-amp, the voltage gain (as a function of frequency) is V2 (s) = −  Z2 (s) V1 (s), Z1 (s)  where Z1 =  R1 1 + R1 Cs  and Z2 = R2 are the respective circuit impedances. Therefore, we obtain V2 (s) = −  Z    R2 (1 + R1 Cs) V1 (s). R1   Z  1  C  + R1  2  R2  +  +  V1(s)  V2(s)  -  -  FIGURE P2.7 Differentiating op-amp circuit.  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  40  CHAPTER 2  Let  ∆=  G2 + Cs  −Cs  −G2  −Cs  G1 + 2Cs  −Cs  −G2  −Cs  Cs + G2  .  Then, Vj =  ∆ij I1 ∆  or  or  V3 ∆13 I1 /∆ = . V1 ∆11 I1 /∆  Therefore, the transfer function is −Cs 2Cs + G1 T (s) =  ∆13 V3 = = V1 ∆11  −G2  −Cs  2Cs + G1  −Cs  −Cs  Cs + G2  Pole-zero map (x:poles and o:zeros) 3  2  o  1  Imag Axis  P2.8  Mathematical Models of Systems  0  x  x  -1  -2  -3 -8  o  -7  -6  -5  -4 Real Axis  FIGURE P2.8 Pole-zero map.  -3  -2  -1  0  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  41  Problems  =  C 2 R1 R2 s2 + 2CR1 s + 1 . C 2 R1 R2 s2 + (2R1 + R2 )Cs + 1  Using R1 = 0.5, R2 = 1, and C = 0.5, we have T (s) =  s2 + 4s + 8 (s + 2 + 2j)(s + 2 − 2j) √ √ . = 2 s + 8s + 8 (s + 4 + 8)(s + 4 − 8)  The pole-zero map is shown in Figure P2.8. From P2.3 we have M ẍ1 + kx1 + k(x1 − x2 ) = F (t) M ẍ2 + k(x2 − x1 ) + bẋ2 = 0 . Taking the Laplace transform of both equations and writing the result in matrix form, it follows that   M s2 + 2k    −k  M s2 + bs + k  −k     X1 (s) X2 (s)      =  F (s) 0     ,  Pole zero map 0.4  0.3  0.2  0.1 Imag Axis  P2.9  0  - 0.1  -0.2  -0.3  -0.4 -0.03  FIGURE P2.9 Pole-zero map.  -0.025  -0.02  -0.015 Real Axis  -0.01  -0.005  0  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  42  CHAPTER 2  Mathematical Models of Systems  or           k F (s) 1  M s2 + bs + k  =   2 ∆ X2 (s) k M s + 2k 0 X1 (s)  where ∆ = (M s2 + bs + k)(M s2 + 2k) − k 2 . So, G(s) =  M s2 + bs + k X1 (s) = . F (s) ∆  When b/k = 1, M = 1 , b2 /M k = 0.04, we have G(s) =  s2 + 0.04s + 0.04 . s4 + 0.04s3 + 0.12s2 + 0.0032s + 0.0016  The pole-zero map is shown in Figure P2.9. P2.10  From P2.2 we have M1 ÿ1 + k12 (y1 − y2 ) + bẏ1 + k1 y1 = F (t) M2 ÿ2 + k12 (y2 − y1 ) = 0 . Taking the Laplace transform of both equations and writing the result in matrix form, it follows that   or  M1 s2 + bs + k1 + k12     M2 s2 + k12  −k12   −k12     Y1 (s) Y2 (s)      =    F (s)    0      k12 F (s) 1  M2 s2 + k12  =   ∆ Y2 (s) k12 M1 s2 + bs + k1 + k12 0 Y1 (s)  where  2 ∆ = (M2 s2 + k12 )(M1 s2 + bs + k1 + k12 ) − k12 .  So, when f (t) = a sin ωo t, we have that Y1 (s) is given by Y1 (s) =  aM2 ωo (s2 + k12 /M2 ) . (s2 + ωo2 )∆(s)  For motionless response (in the steady-state), set the zero of the transfer function so that (s2 +  k12 ) = s2 + ωo2 M2  or  ωo2 =  k12 . M2  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  43  Problems  P2.11  The transfer functions from Vc (s) to Vd (s) and from Vd (s) to θ(s) are: K1 K2 , and (Lq s + Rq )(Lc s + Rc ) Km . θ(s)/Vd (s) = 2 (Js + f s)((Ld + La )s + Rd + Ra ) + K3 Km s  Vd (s)/Vc (s) =  The block diagram for θ(s)/Vc (s) is shown in Figure P2.11, where θ(s)/Vc (s) =  K1 K2 Km θ(s) Vd (s) = , Vd (s) Vc (s) ∆(s)  where ∆(s) = s(Lc s + Rc )(Lq s + Rq )((Js + b)((Ld + La )s + Rd + Ra ) + Km K3 ) .  Vc  1 L cs+R c  Ic  K1  Vq  1 L qs+R q  Iq K2  Vd +  1 (L d+L a)s+R d+R a  Id  Tm Km  -  1 Js+f  w  1 s  q  Vb K3  FIGURE P2.11 Block diagram.  P2.12  The open-loop transfer function is Y (s) K = . R(s) s + 20 With R(s) = 1/s, we have Y (s) =  K . s(s + 20)  The partial fraction expansion is K Y (s) = 20    1 1 − , s s + 20   and the inverse Laplace transform is y(t) =   K 1 − e−20t , 20  As t → ∞, it follows that y(t) → K/20. So we choose K = 20 so that y(t)  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  44  CHAPTER 2  Mathematical Models of Systems  approaches 1. Alternatively we can use the final value theorem to obtain y(t)t→∞ = lim sY (s) = s→0  K =1. 20  It follows that choosing K = 20 leads to y(t) → 1 as t → ∞. P2.13  The motor torque is given by Tm (s) = (Jm s2 + bm s)θm (s) + (JL s2 + bL s)nθL (s) = n((Jm s2 + bm s)/n2 + JL s2 + bL s)θL (s) where n = θL (s)/θm (s) = gear ratio . But Tm (s) = Km Ig (s) and Ig (s) =  1 Vg (s) , (Lg + Lf )s + Rg + Rf  and Vg (s) = Kg If (s) =  Kg Vf (s) . Rf + L f s  Combining the above expressions yields θL (s) Kg Km = . Vf (s) n∆1 (s)∆2 (s) where ∆1 (s) = JL s2 + bL s +  Jm s2 + bm s n2  and ∆2 (s) = (Lg s + Lf s + Rg + Rf )(Rf + Lf s) . P2.14  For a field-controlled dc electric motor we have ω(s)/Vf (s) =  Km /Rf . Js + b  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  45  Problems  With a step input of Vf (s) = 80/s, the final value of ω(t) is 80Km = 2.4 Rf b  ω(t)t→∞ = lim sω(s) = s→0  or  Km = 0.03 . Rf b  Solving for ω(t) yields 80Km −1 1 ω(t) = L Rf J s(s + b/J)     =  80Km (1−e−(b/J)t ) = 2.4(1−e−(b/J)t ) . Rf b  At t = 1/2, ω(t) = 1, so ω(1/2) = 2.4(1 − e−(b/J)t ) = 1 implies  b/J = 1.08 sec .  Therefore, ω(s)/Vf (s) = P2.15  0.0324 . s + 1.08  Summing the forces in the vertical direction and using Newton's Second Law we obtain ẍ +  k x=0. m  The system has no damping and no external inputs. Taking the Laplace transform yields X(s) =  s2  x0 s , + k/m  where we used the fact that x(0) = x0 and ẋ(0) = 0. Then taking the inverse Laplace transform yields x(t) = x0 cos P2.16  s  k t. m  Using Cramer's rule, we have   1 1.5  x1      or   2  4      x1 x2      =  6 11        1  4 −1.5   6   = ∆ −2 x2 1 11  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  46  CHAPTER 2  Mathematical Models of Systems  where ∆ = 4(1) − 2(1.5) = 1 . Therefore, x1 =  4(6) − 1.5(11) = 7.5 1  and x2 =  −2(6) + 1(11) = −1 . 1  The signal flow graph is shown in Figure P2.16. 11 1/4  6  1  -1/2  X2  X1 -1.5  FIGURE P2.16 Signal flow graph.  So, x1 = P2.17  6(1) − 1.5( 11 4 ) = 7.5 3 1− 4  and x2 =  11( 41 ) + 1−  −1 2 (6) 3 4  = −1 .  (a) For mass 1 and 2, we have M1 ẍ1 + K1 (x1 − x2 ) + b1 (ẋ3 − ẋ1 ) = 0 M2 ẍ2 + K2 (x2 − x3 ) + b2 (ẋ3 − ẋ2 ) + K1 (x2 − x1 ) = 0 . (b) Taking the Laplace transform yields (M1 s2 + b1 s + K1 )X1 (s) − K1 X2 (s) = b1 sX3 (s) −K1 X1 (s) + (M2 s2 + b2 s + K1 + K2 )X2 (s) = (b2 s + K2 )X3 (s) . (c) Let G1 (s) = K2 + b2 s G2 (s) = 1/p(s) G3 (s) = 1/q(s) G4 (s) = sb1 , where p(s) = s2 M2 + sf2 + K1 + K2 and q(s) = s2 M1 + sf1 + K1 .  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  47  Problems  The signal flow graph is shown in Figure P2.17.  G4 G3  X3 G1  G2  X1  K1 K1  FIGURE P2.17 Signal flow graph.  (d) The transfer function from X3 (s) to X1 (s) is X1 (s) K1 G1 (s)G2 (s)G3 (s) + G4 (s)G3 (s) = . X3 (s) 1 − K12 G2 (s)G3 (s) P2.18  The signal flow graph is shown in Figure P2.18. I1  V1  Va  Z2  Y3  Ia  Z4 V2  Y1 -Z 2  -Y 1  -Y 3  FIGURE P2.18 Signal flow graph.  The transfer function is V2 (s) Y 1 Z2 Y 3 Z4 = . V1 (s) 1 + Y 1 Z2 + Y 3 Z2 + Y 3 Z4 + Y 1 Z2 Z4 Y 3 P2.19  For a noninerting op-amp circuit, depicted in Figure P2.19a, the voltage gain (as a function of frequency) is Vo (s) =  Z1 (s) + Z2 (s) Vin (s), Z1 (s)  where Z1 (s) and Z2 (s) are the impedances of the respective circuits. In the case of the voltage follower circuit, shown in Figure P2.19b, we have  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  48  CHAPTER 2  Mathematical Models of Systems  Z2 Z1  +  vin  v0  +  vin  (a)  v0  (b)  FIGURE P2.19 (a) Noninverting op-amp circuit. (b) Voltage follower circuit.  Z1 = ∞ (open circuit) and Z2 = 0. Therefore, the transfer function is Vo (s) Z1 = = 1. Vin (s) Z1 P2.20  (a) Assume Rg ≫ Rs and Rs ≫ R1 . Then Rs = R1 + R2 ≈ R2 , and vgs = vin − vo , where we neglect iin , since Rg ≫ Rs . At node S, we have vo = gm vgs = gm (vin − vo ) or Rs  vo gm Rs = . vin 1 + gm Rs  (b) With gm Rs = 20, we have vo 20 = = 0.95 . vin 21 (c) The block diagram is shown in Figure P2.20.  vin(s)  gmRs  -  FIGURE P2.20 Block diagram model.  P2.21  From the geometry we find that ∆z = k  l1 − l2 l2 (x − y) − y . l1 l1  vo(s)  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  49  Problems  The flow rate balance yields A  dy = p∆z dt  which implies  Y (s) =  p∆Z(s) . As  By combining the above results it follows that l2 p l1 − l2 Y (s) = k (X(s) − Y (s)) − Y (s) . As l1 l1        Therefore, the signal flow graph is shown in Figure P2.21. Using Mason's -1 (l 1 - l 2)/l 1  k X  DZ  p/As Y  1 -l 2 / l 1  FIGURE P2.21 Signal flow graph.  gain formula we find that the transfer function is given by Y (s) = X(s) 1+  k(l1 −l2 )p l1 As k(l1 −l2 )p l2 p l1 As + l1 As  =  K1 , s + K2 + K1  where K1 = P2.22  k(l1 − l2 )p p l1 A  and K2 =  l2 p . l1 A  (a) The equations of motion for the two masses are L 2 L M L θ¨1 + M gLθ1 + k (θ1 − θ2 ) = f (t) 2 2  2 L M L2 θ¨2 + M gLθ2 + k (θ2 − θ1 ) = 0 . 2    2  With θ˙1 = ω1 and θ˙2 = ω2 , we have g k k f (t) ω˙1 = − + θ1 + θ2 + L 4M 4M 2M L   k g k ω˙2 = θ1 − + θ2 . 4M L 4M     © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  50  CHAPTER 2  Mathematical Models of Systems  a  -  F (t)  w1 1/s  1/s  1/2ML  (a)  q1  b  w2  1/s  1/s  q2  a  Imag(s) + j  + j  (b)  g k L + 4M  g k L + 2M  X O  X  + j  g L  Re(s) FIGURE P2.22 (a) Block diagram. (b) Pole-zero map.  (b) Define a = g/L + k/4M and b = k/4M . Then θ1 (s) 1 s2 + a = . F (s) 2M L (s2 + a)2 − b2 (c) The block diagram and pole-zero map are shown in Figure P2.22. P2.23  The input-output ratio, Vce /Vin , is found to be β(R − 1) + hie Rf Vce = . Vin −βhre + hie (−hoe + Rf )  P2.24  (a) The voltage gain is given by vo RL β1 β2 (R1 + R2 ) . = vin (R1 + R2 )(Rg + hie1 ) + R1 (R1 + R2 )(1 + β1 ) + R1 RL β1 β2  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  51  Problems  (b) The current gain is found to be ic2 = β1 β2 . ib1 (c) The input impedance is vin (R1 + R2 )(Rg + hie1 ) + R1 (R1 + R2 )(1 + β1 ) + R1 RL β1 β2 = , ib1 R1 + R2 and when β1 β2 is very large, we have the approximation vin RL R1 β1 β2 ≈ . ib1 R1 + R2 P2.25  The transfer function from R(s) and Td (s) to Y (s) is given by   Y (s) = G(s) R(s) −  1 (G(s)R(s) + Td (s)) + Td (s) + G(s)R(s) G(s)   = G(s)R(s) . Thus, Y (s)/R(s) = G(s) . Also, we have that Y (s) = 0 .  when R(s) = 0. Therefore, the effect of the disturbance, Td (s), is eliminated. P2.26  The equations of motion for the two mass model of the robot are M ẍ + b(ẋ − ẏ) + k(x − y) = F (t) mÿ + b(ẏ − ẋ) + k(y − x) = 0 . Taking the Laplace transform and writing the result in matrix form yields    M s2 + bs + k −(bs + k)  −(bs + k)  ms2  + bs + k     X(s) Y (s)      k m   .  =  Solving for Y (s) we find that 1 Y (s) mM (bs +k) = m b F (s) s2 [s2 + 1 + M ms +  ]  F (s) 0     .  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  52  CHAPTER 2  P2.27  Mathematical Models of Systems  The describing equation of motion is mz̈ = mg − k  i2 . z2  Defining f (z, i) = g −  ki2 mz 2  leads to z̈ = f (z, i) . The equilibrium condition for io and zo , found by solving the equation of motion when ż = z̈ = 0 , is ki2o = zo2 . mg We linearize the equation of motion using a Taylor series approximation. With the definitions ∆z = z − zo  and ∆i = i − io ,  ˙ = ż and ∆z ¨ = z̈. Therefore, we have ∆z ¨ = f (z, i) = f (zo , io ) + ∂f ∆z ∂z  z=zo i=io  ∆z +  ∂f ∂i  z=zo i=io  ∆i + · · ·  But f (zo , io ) = 0, and neglecting higher-order terms in the expansion yields 2 ¨ = 2kio ∆z − 2kio ∆i . ∆z mzo3 mzo2  Using the equilibrium condition which relates zo to io , we determine that ¨ = 2g ∆z − g ∆i . ∆z zo io Taking the Laplace transform yields the transfer function (valid around the equilibrium point) ∆Z(s) −g/io = 2 . ∆I(s) s − 2g/zo  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  57  Problems  R(s )  +  K1 s (s+1)  Y (s)  1 +K 2s  1 0.9 0.8 0.7 <---- time to 90% = 0.39 sec  y(t)  0.6 0.5 0.4 0.3 0.2 0.1 0 0  0.2  0.4  0.6  0.8  1  1.2  1.4  1.6  1.8  2  time(sec)  FIGURE P2.35 The equivalent block diagram and the system step response.  P2.36  (a) Given R(s) = 1/s2 , the partial fraction expansion is Y (s) =  s2 (s  24 3 8/3 3/4 1 13/12 = − + + 2− . + 2)(s + 3)(s + 4) s+2 s+3 s+4 s s  Therefore, using the Laplace transform table, we determine that the ramp response is 8 3 13 y(t) = 3e−2t − e−3t + e−4t + t − , 3 4 12  t≥0.  (b) For the ramp input, y(t) ≈ 0.21 at t = 1. second (see Figure P2.36a).  (c) Given R(s) = 1, the partial fraction expansion is Y (s) =  24 12 24 12 = − + . (s + 2)(s + 3)(s + 4) s+2 s+3 s+4  Therefore, using the Laplace transform table, we determine that the impulse response is y(t) = 12e−2t − 24e−3t + 412e−4t ,  t≥0.  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  53  Problems  P2.28  The signal flow graph is shown in Figure P2.28. -d G  B +b  P  +c  D  +a -m +e  -k  M  +g +f  +h  S  C  FIGURE P2.28 Signal flow graph.  (a) The PGBDP loop gain is equal to -abcd. This is a negative transmission since the population produces garbage which increases bacteria and leads to diseases, thus reducing the population. (b) The PMCP loop gain is equal to +efg. This is a positive transmission since the population leads to modernization which encourages immigration, thus increasing the population. (c) The PMSDP loop gain is equal to +ehkd. This is a positive transmission since the population leads to modernization and an increase in sanitation facilities which reduces diseases, thus reducing the rate of decreasing population. (d) The PMSBDP loop gain is equal to +ehmcd. This is a positive transmission by similar argument as in (3). P2.29  Assume the motor torque is proportional to the input current Tm = ki . Then, the equation of motion of the beam is J φ̈ = ki , where J is the moment of inertia of the beam and shaft (neglecting the inertia of the ball). We assume that forces acting on the ball are due to gravity and friction. Hence, the motion of the ball is described by mẍ = mgφ − bẋ  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  54  CHAPTER 2  Mathematical Models of Systems  where m is the mass of the ball, b is the coefficient of friction, and we have assumed small angles, so that sin φ ≈ φ. Taking the Laplace transfor of both equations of motion and solving for X(s) yields X(s)/I(s) = P2.30  gk/J . + b/m)  s2 (s2  Given H(s) =  k τs + 1  where τ = 4µs = 4 × 10−6 seconds and 0.999 ≤ k < 1.001. The step response is Y (s) =  k 1 k k · = − . τs + 1 s s s + 1/τ  Taking the inverse Laplace transform yields y(t) = k − ke−t/τ = k(1 − e−t/τ ) . The final value is k. The time it takes to reach 98% of the final value is t = 15.6µs independent of k. P2.31  From the block diagram we have Y1 (s) = G2 (s)[G1 (s)E1 (s) + G3 (s)E2 (s)] = G2 (s)G1 (s)[R1 (s) − H1 (s)Y1 (s)] + G2 (s)G3 (s)E2 (s) . Therefore, Y1 (s) =  G1 (s)G2 (s) G2 (s)G3 (s) R1 (s) + E2 (s) . 1 + G1 (s)G2 (s)H1 (s) 1 + G1 (s)G2 (s)H1 (s)  And, computing E2 (s) (with R2 (s) = 0) we find G4 (s) E2 (s) = H2 (s)Y2 (s) = H2 (s)G6 (s) Y1 (s) + G5 (s)E2 (s) G2 (s)   or E2 (s) =  G4 (s)G6 (s)H2 (s) Y1 (s) . G2 (s)(1 − G5 (s)G6 (s)H2 (s))  Substituting E2 (s) into equation for Y1 (s) yields Y1 (s) =  G1 (s)G2 (s) R1 (s) 1 + G1 (s)G2 (s)H1 (s)    © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  55  Problems  +  G3 (s)G4 (s)G6 (s)H2 (s) Y1 (s) . (1 + G1 (s)G2 (s)H1 (s))(1 − G5 (s)G6 (s)H2 (s))  Finally, solving for Y1 (s) yields Y1 (s) = T1 (s)R1 (s) where T1 (s) =   G1 (s)G2 (s)(1 − G5 (s)G6 (s)H2 (s)) (1 + G1 (s)G2 (s)H1 (s))(1 − G5 (s)G6 (s)H2 (s)) − G3 (s)G4 (s)G6 (s)H2 (s)    .    .  Similarly, for Y2 (s) we obtain Y2 (s) = T2 (s)R1 (s) . where T2 (s) =   P2.32  G1 (s)G4 (s)G6 (s) (1 + G1 (s)G2 (s)H1 (s))(1 − G5 (s)G6 (s)H2 (s)) − G3 (s)G4 (s)G6 (s)H2 (s)  The signal flow graph shows three loops: L1 = −G1 G3 G4 H2 L2 = −G2 G5 G6 H1 L3 = −H1 G8 G6 G2 G7 G4 H2 G1 . The transfer function Y2 /R1 is found to be Y2 (s) G1 G8 G6 ∆1 − G2 G5 G6 ∆2 = , R1 (s) 1 − (L1 + L2 + L3 ) + (L1 L2 ) where for path 1 ∆1 = 1 and for path 2 ∆ 2 = 1 − L1 . Since we want Y2 to be independent of R1 , we need Y2 /R1 = 0. Therefore, we require G1 G8 G6 − G2 G5 G6 (1 + G1 G3 G4 H2 ) = 0 .  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  56  CHAPTER 2  P2.33  Mathematical Models of Systems  The closed-loop transfer function is G3 (s)G1 (s)(G2 (s) + K5 K6 ) Y (s) = . R(s) 1 − G3 (s)(H1 (s) + K6 ) + G3 (s)G1 (s)(G2 (s) + K5 K6 )(H2 (s) + K4 )  P2.34  The equations of motion are m1 ÿ1 + b(ẏ1 − ẏ2 ) + k1 (y1 − y2 ) = 0 m2 ÿ2 + b(ẏ2 − ẏ1 ) + k1 (y2 − y1 ) + k2 y2 = k2 x Taking the Laplace transform yields (m1 s2 + bs + k1 )Y1 (s) − (bs + k1 )Y2 (s) = 0 (m2 s2 + bs + k1 + k2 )Y2 (s) − (bs + k1 )Y1 (s) = k2 X(s) Therefore, after solving for Y1 (s)/X(s), we have Y2 (s) k2 (bs + k1 ) = . 2 X(s) (m1 s + bs + k1 )(m2 s2 + bs + k1 + k2 ) − (bs + k1 )2  P2.35  (a) We can redraw the block diagram as shown in Figure P2.35. Then, T (s) =  K1 /s(s + 1) K1 = 2 . 1 + K1 (1 + K2 s)/s(s + 1) s + (1 + K2 K1 )s + K2  (b) The signal flow graph reveals two loops (both touching): L1 =  −K1 s(s + 1)  and  L2 =  −K1 K2 . s+1  Therefore, T (s) =  K1 /s(s + 1) K1 = 2 . 1 + K1 /s(s + 1) + K1 K2 /(s + 1) s + (1 + K2 K1 )s + K1  (c) We want to choose K1 and K2 such that s2 + (1 + K2 K1 )s + K1 = s2 + 20s + 100 = (s + 10)2 . Therefore, K1 = 100 and 1 + K2 K1 = 20 or K2 = 0.19. (d) The step response is shown in Figure P2.35.  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  58  CHAPTER 2  Mathematical Models of Systems  (d) For the impulse input, y(t) ≈ 0.65 at t = 1 seconds (see Figure P2.36b). (a) Ramp input  (b) Impulse input 0.8  2  1.8  0.7  1.6 0.6 1.4 0.5  y(t)  y(t)  1.2  1  0.8  0.4  0.3  0.6 0.2 0.4 0.1  0.2  0  0  1  2  0  3  0  1  Time (sec)  2  3  Time (sec)  FIGURE P2.36 (a) Ramp input response. (b) Impulse input response.  P2.37  The equations of motion are m1  d2 x = −(k1 + k2 )x + k2 y dt2  and  m2  d2 y = k2 (x − y) + u . dt2  When m1 = m2 = 1 and k1 = k2 = 1, we have d2 x = −2x + y dt2 P2.38  and  d2 y =x−y+u . dt2  The equation of motion for the system is J  d2 θ dθ + b + kθ = 0 , dt2 dt  where k is the rotational spring constant and b is the viscous friction coefficient. The initial conditions are θ(0) = θo and θ̇(0) = 0. Taking the  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  59  Problems  Laplace transform yields J(s2 θ(s) − sθo ) + b(sθ(s) − θo ) + kθ(s) = 0 . Therefore, θ(s) =  (s + Jb θo ) (s2  +  b Js  +  K J)  =  (s + 2ζωn )θo . + 2ζωn s + ωn2  s2  Neglecting the mass of the rod, the moment of inertia is detemined to be J = 2M r 2 = 0.5 kg · m2 . Also, s  ωn =  k = 0.02 rad/s J  and ζ =  b = 0.01 . 2Jωn  Solving for θ(t), we find that q θo θ(t) = p e−ζωn t sin(ωn 1 − ζ 2 t + φ) , 1 − ζ2  where tan φ =  p  1 − ζ 2 /ζ). Therefore, the envelope decay is θo θe = p e−ζωn t . 1 − ζ2  So, with ζωn = 2 × 10−4 , θo = 4000o and θf = 10o , the elapsed time is computed as t= P2.39  1 θo ln p = 8.32 hours . ζωn 1 − ζ 2 θf  When t < 0, we have the steady-state conditions i1 (0) = 1A ,  va (0) = 2V  and  vc (0) = 5V ,  where vc (0) is associated with the 1F capacitor. After t ≥ 0, we have 2  di1 + 2i1 + 4(i1 − i2 ) = 10e−2t dt  and Z  i2 dt + 10i2 + 4(i2 − i1 ) − i1 = 0 .  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  60  CHAPTER 2  Mathematical Models of Systems  Taking the Laplace transform (using the initial conditions) yields 2(sI1 − i1 (0)) + 2I1 + 4I1 − 4I2 =  10 s+2  or  (s + 3)I1 (s) − 2I2 (s) =  s+7 s+2  and 1 [ I2 −vc (0)]+10I2 +4(I2 −I1 ) = I1 (s) or s  −5sI1 (s)+(14s+1)I2 (s) = 5s .  Solving for I2 (s) yields I2 =  5s(s2 + 6s + 13) , 14(s + 2)∆(s)  where ∆(s) =  s+3  −2  −5s  14s + 1  = 14s2 + 33s + 3 .  Then, Vo (s) = 10I2 (s) . P2.40  The equations of motion are J1 θ̈1 = K(θ2 − θ1 ) − b(θ̇1 − θ̇2 ) + T  and J2 θ̈2 = b(θ̇1 − θ̇2 ) .  Taking the Laplace transform yields (J1 s2 + bs + K)θ1 (s) − bsθ2 (s) = Kθ2 (s) + T (s) and (J2 s2 + bs)θ2 (s) − bsθ1 (s) = 0 . Solving for θ1 (s) and θ2 (s), we find that θ1 (s) =  (Kθ2 (s) + T (s))(J2 s + b) ∆(s)  and  θ2 (s) =  b(Kθ2 (s) + T (s)) , ∆(s)  where ∆(s) = J1 J2 s3 + b(J1 + J2 )s2 + J2 Ks + bK . P2.41  Assume that the only external torques acting on the rocket are control torques, Tc and disturbance torques, Td , and assume small angles, θ(t). Using the small angle approximation, we have ḣ = V θ  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  61  Problems  J θ̈ = Tc + Td , where J is the moment of inertia of the rocket and V is the rocket velocity (assumed constant). Now, suppose that the control torque is proportional to the lateral displacement, as Tc (s) = −KH(s) , where the negative sign denotes a negative feedback system. The corresponding block diagram is shown in Figure P2.41. Td  H desired=0  K +  Tc  + +  -  1 Js 2  V s  H( s)  FIGURE P2.41 Block diagram.  P2.42  (a) The equation of motion of the motor is J  dω = Tm − bω , dt  where J = 0.1, b = 0.06, and Tm is the motor input torque. (b) Given Tm (s) = 1/s, and ω(0) = 0.7, we take the Laplace transform of the equation of motion yielding sω(s) − ω(0) + 0.6ω(s) = 10Tm or ω(s) =  0.7s + 10 . s(s + 0.6)  Then, computing the partial fraction expansion, we find that ω(s) =  A B 16.67 15.97 + = − . s s + 0.6 s s + 0.6  The step response, determined by taking the inverse Laplace transform, is ω(t) = 16.67 − 15.97e−0.6t ,  t≥0.  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  62  CHAPTER 2  P2.43  Mathematical Models of Systems  The work done by each gear is equal to that of the other, therefore Tm θm = TL θL . Also, the travel distance is the same for each gear, so r1 θ m = r2 θ L . The number of teeth on each gear is proportional to the radius, or r1 N 2 = r2 N 1 . So, θm r2 N2 = = , θL r1 N1 and N1 θ m = N2 θ L N1 θL = θm = nθm , N2 where n = N1 /N2 . Finally, Tm θL N1 = = =n. TL θm N2  P2.44  The inertia of the load is JL =  πρLr 4 . 2  Also, from the dynamics we have T2 = JL ω̇2 + bL ω2 and T1 = nT2 = n(JL ω̇2 + bL ω2 ) . So, T1 = n2 (JL ω̇1 + bL ω1 ) ,  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  63  Problems  since ω2 = nω1 . Therefore, the torque at the motor shaft is T = T1 + Tm = n2 (JL ω̇1 + bL ω1 ) + Jm ω̇1 + bm ω1 . P2.45  Let U (s) denote the human input and F (s) the load input. The transfer function is P (s) =  G(s) + KG1 (s) Gc (s) + KG1 (s) U (s) + F (s) , ∆(s) ∆(s)  where ∆ = 1 + GH(s) + G1 KBH(s) + Gc E(s) + G1 KE(s) . P2.46  Consider the application of Newton's law ( mv we obtain  P  F = mẍ). From the mass  mv ẍ1 = F − k1 (x1 − x2 ) − b1 (ẋ1 − ẋ2 ). Taking the Laplace transform, and solving for X1 (s) yields X1 (s) =  b1 s + k1 1 F (s) + X2 (s), ∆1 (s) ∆1 (s)  where ∆1 := mv s2 + b1 s + k1 . From the mass mt we obtain mt ẍ2 = −k2 x2 − b2 ẋ2 + k1 (x1 − x2 ) + b1 (ẋ1 − ẋ2 ). Taking the Laplace transform, and solving for X2 (s) yields X2 (s) =  b1 s + k1 X1 (s), ∆2 (s)  where ∆2 := mt s2 + (b1 + b2 )s + k1 + k2 . Substituting X2 (s) above into the relationship fpr X1 (s) yields the transfer function ∆2 (s) X1 (s) = . F (s) ∆1 (s)∆2 (s) − (b1 s + k1 )2  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  64  CHAPTER 2  P2.47  Mathematical Models of Systems  Using the following relationships h(t) =  Z  (1.6θ(t) − h(t))dt  ω(t) = θ̇(t) J ω̇(t) = Km ia (t) va (t) = 50vi (t) = 10ia (t) + vb (t) θ̇ = Kvb we find the differential equation is d3 h Km + 1+ 3 dt 10JK   P2.48    Km dh 8Km d2 h + = vi . 2 dt 10JK dt J  (a) The transfer function is V2 (s) (1 + sR1 C1 )(1 + sR2 C2 ) = . V1 (s) R1 C 2 s (b) When R1 = 100 kΩ, R2 = 200 kΩ, C1 = 1 µF and C2 = 0.1 µF , we have V2 (s) 0.2(s + 10)(s + 50) = . V1 (s) s  P2.49  (a) The closed-loop transfer function is T (s) =  G(s) 6205 = 3 . 1 + G(s) s + 13s2 + 1281s + 6205  (b) The poles of T (s) are s1 = −5 and s2,3 = −4 ± j35. (c) The partial fraction expansion (with a step input) is Y (s) = 1 −  1.0122 0.0061 + 0.0716j 0.0061 − 0.0716j + + . s+5 s + 4 + j35 s + 4 − j35  (d) The step response is shown in Figure P2.49. The real and complex roots are close together and by looking at the poles in the s-plane we have difficulty deciding which is dominant. However, the residue at the real pole is much larger and thus dominates the response.  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  65  Problems  1 0.9 0.8  Amplitude  0.7 0.6 0.5 0.4 0.3 0.2 0.1 0 0  0.2  0.4  0.6  0.8  1  1.2  1.4  1.6  1.8  2  Time (secs)  FIGURE P2.49 Step response.  P2.50  (a) The closed-loop transfer function is T (s) =  s3  +  45s2  14000 . + 3100s + 14500  (b) The poles of T (s) are s1 = −5 and  s2,3 = −20 ± j50.  (c) The partial fraction expansion (with a step input) is Y (s) =  0.9655 1.0275 0.0310 − 0.0390j 0.0310 + 0.0390j − + + . s s+5 s + 20 + j50 s + 20 − j50  (d) The step response is shown in Figure P2.50. The real root dominates the response. (e) The final value of y(t) is yss = lim sY (s) = 0.9655 . s→0  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  CHAPTER 2  Mathematical Models of Systems  1 0.9 0.8 0.7  Amplitude  66  0.6 0.5 0.4 0.3 0.2 0.1 0 0  0.2  0.4  0.6  0.8  1  1.2  1.4  1.6  1.8  2  Time (secs)  FIGURE P2.50 Step response.  P2.51  Consider the free body diagram in Figure P2.51. Using Newton's Law and summing the forces on the two masses yields M1 ẍ(t) + b1 ẋ(t) + k1 x(t) = b1 ẏ(t) M2 ÿ(t) + b1 ẏ(t) + k2 y(t) = b1 ẋ(t) + u(t)  k1x  M1  k1  x . . b1(x - y)  k2  M1 x  . . b1(y - x) k2 y  b1 M2  M2  y  y u(t)  FIGURE P2.51 Free body diagram.  u(t)  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  67  Advanced Problems  Advanced Problems AP2.1  The transfer function from V (s) to ω(s) has the form ω(s) Km = . V (s) τm s + 1 In the steady-state, ωss = lim s s→0    Km 5 = 5Km . τm s + 1 s   So, Km = 70/5 = 14 . Also, ω(t) = Vm Km (1 − e−t/τm ) where V (s) = Vm /s. Solving for τm yields τm =  −t . ln(1 − ω(t)/ωss )  When t = 2, we have τm =  −2 = 3.57 . ln(1 − 30/70)  Therefore, the transfer function is ω(s) 14 = . V (s) 3.57s + 1 AP2.2  The closed-loop transfer function form R1 (s) to Y2 (s) is Y2 (s) G1 G4 G5 (s) + G1 G2 G3 G4 G6 (s) = R1 (s) ∆ where ∆ = [1 + G3 G4 H2 (s)][1 + G1 G2 H3 (s)] . If we select G5 (s) = −G2 G3 G6 (s) then the numerator is zero, and Y2 (s)/R1 (s) = 0. The system is now decoupled.  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  68  CHAPTER 2  AP2.3  Mathematical Models of Systems  (a) Computing the closed-loop transfer function: G(s)Gc (s) R(s) . Y (s) = 1 + Gc (s)G(s)H(s)     Then, with E(s) = R(s) − Y (s) we obtain E(s) =    1 + Gc (s)G(s)(H(s) − 1) R(s) . 1 + Gc (s)G(s)H(s)   If we require that E(s) ≡ 0 for any input, we need 1 + Gc (s)G(s)(H(s) − 1) = 0 or H(s) =  Gc (s)G(s) − 1 n(s) = . Gc (s)G(s) d(s)  Since we require H(s) to be a causal system, the order of the numerator polynomial, n(s), must be less than or equal to the order of the denominator polynomial, d(s). This will be true, in general, only if both Gc (s) and G(s) are proper rational functions (that is, the numerator and denominator polynomials have the same order). Therefore, making E ≡ 0 for any input R(s) is possible only in certain circumstances. (b) The transfer function from Td (s) to Y (s) is Gd (s)G(s) Y (s) = Td (s) . 1 + Gc (s)G(s)H(s)     With H(s) as in part (a) we have Gd (s) Y (s) = Td (s) . Gc (s)     (c) No. Since Y (s) =    Gd (s)G(s) Td (s) = T (s)Td (s) , 1 + Gc (s)G(s)H(s)   the only way to have Y (s) ≡ 0 for any Td (s) is for the transfer function T (s) ≡ 0 which is not possible in general (since G(s) 6= 0). AP2.4  (a) With q(s) = 1/s we obtain τ (s) =  1/Ct s+  QS+1/R Ct  ·  1 . s  Define α :=  QS + 1/R Ct  and  β := 1/Ct .  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  69  Advanced Problems  Then, it follows that τ (s) =  β 1 −β/α β/α · = + . s+α s s+α s  Taking the inverse Laplace transform yields τ (t) =  β −β −αt β e + = [1 − e−αt ] . α α α  1 (b) As t → ∞, τ (t) → αβ = Qs+1/R . (c) To increase the speed of response, you want to choose Ct , Q, S and R such that  α :=  Qs + 1/R Ct  is "large." AP2.5  Considering the motion of each mass, we have M3 ẍ3 + b3 ẋ3 + k3 x3 = u3 + b3 ẋ2 + k3 x2 M2 ẍ2 + (b2 + b3 )ẋ2 + (k2 + k3 )x2 = u2 + b3 ẋ3 + k3 x3 + b2 ẋ1 + k2 x1 M1 ẍ1 + (b1 + b2 )ẋ1 + (k1 + k2 )x1 = u1 + b2 ẋ2 + k2 x2 In matrix form the three equations can be written as   0  M1   0 M2    AP2.6  0  0            0    0  0   ẍ1   b1 + b2 −b2 0       0  b2 + b3 −b3   ẍ2  +  −b2  M3    ẍ3    −b3  b3  −k2 0  k1 + k2  + k2 + k3 −k3  −k2 −k3  k3        ẋ1      ẋ   2    ẋ3                  x 1   u1       x  =  u  .  2   2    x3  u3  Considering the cart mass and using Newton's Law we obtain M ẍ = u − bẋ − F sin ϕ where F is the reaction force between the cart and the pendulum. Considering the pendulum we obtain m  d2 (x + L sin ϕ) = F sin ϕ dt2    © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  70  CHAPTER 2  Mathematical Models of Systems  m  d2 (L cos ϕ) = F cos ϕ + mg dt2  Eliminating the reaction force F yields the two equations (m + M )ẍ + bẋ + mLϕ̈ cos ϕ − mLϕ̇2 sin ϕ = u mL2 ϕ̈ + mgL sin ϕ + mLẍ cos ϕ = 0 If we assume that the angle ϕ ≈ 0, then we have the linear model (m + M )ẍ + bẋ + mLϕ̈ = u mL2 ϕ̈ + mgLϕ = −mLẍ AP2.7  The transfer function from the disturbance input to the output is Y (s) =  1 Td (s) . s + 20 + K  When Td (s) = 1, we obtain y(t) = e−(20+K)t . Solving for t when y(t) < 0.1 yields t>  2.3 . 20 + K  When t = 0.05 and y(0.05) = 0.1, we find K = 26.05. AP2.8  The closed-loop transfer function is T (s) =  200K(0.25s + 1) (0.25s + 1)(s + 1)(s + 8) + 200K  The final value due to a step input of R(s) = A/s is v(t) → A  200K . 200K + 8  We need to select K so that v(t) → 50. However, to keep the percent overshoot to less than 10%, we need to limit the magnitude of K. Figure AP2.8a shows the percent overshoot as a function of K. Let K = 0.06 and select the magnitude of the input to be A = 83.3. The inverse Laplace transform of the closed-loop response with R(s) = 83.3/s is v(t) = 50 + 9.85e−9.15t − e−1.93t (59.85 cos(2.24t) + 11.27 sin(2.24t))  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  71  Advanced Problems  The result is P.O. = 9.74% and the steady-state value of the output is approximately 50 m/s, as shown in Figure AP2.8b.  25  Percent Overshoot (%)  20  15  10  5  0  0  0.01  0.02  0.03  0.04  0.05 K  0.06  0.07  0.08  0.09  0.1  Step Response 60  System: untitled1 Peak amplitude: 54.9 Overshoot (%): 9.74 At time (sec): 1.15  50  Amplitude  40  30  20  10  0  0  0.5  1  1.5 Time (sec)  FIGURE AP2.8 (a) Percent overshoot versus the gain K. (b) Step response.  AP2.9  The transfer function is Vo (s) Z2 (s) =− , Vi (s) Z1 (s)  2  2.5  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  72  CHAPTER 2  Mathematical Models of Systems  where Z1 (s) =  R1 R1 C 1 s + 1  and Z2 (s) =  R2 C 2 s + 1 . C2 s  Then we can write KI Vo (s) = Kp + + KD s Vi (s) s where KP = −    R1 C 1 +1 , R2 C 2   KI = −  1 , R1 C 2  KD = −R2 C1 .  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  73  Design Problems  Design Problems CDP2.1  The model of the traction drive, capstan roller, and linear slide follows closely the armature-controlled dc motor model depicted in Figure 2.18 in Dorf and Bishop. The transfer function is T (s) =  rKm , s [(Lm s + Rm )(JT s + bm ) + Kb Km ]  where JT = Jm + r 2 (Ms + Mb ) .  Va(s) -  1 JTs+bm  Km Lms+Rm  1 s  q  r  X(s)  Kb  Back EMF  DP2.1  w  The closed-loop transfer function is Y (s) G1 (s)G2 (s) = . R(s) 1 + G1 (s)H1 (s) − G2 (s)H2 (s) When G1 H1 = G2 H2 and G1 G2 = 1, then Y (s)/R(s) = 1. Therefore, select G1 (s) =  DP2.2  1 G2 (s)  and H1 (s) =  G2 (s)H2 (s) = G22 (s)H2 (s) . G1 (s)  At the lower node we have 1 1 v + + G + 2i2 − 20 = 0 . 4 3     Also, we have v = 24 and i2 = Gv . So 1 1 v + + G + 2Gv − 20 = 0 4 3     © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  74  CHAPTER 2  Mathematical Models of Systems  and G= DP2.3  20 − v    1 4  +  1 3  3v    =  1 S. 12  Taking the Laplace transform of 3 1 1 y(t) = e−t − e−2t − + t 4 4 2 yields Y (s) =  1 1 3 1 − − + 2 . s + 1 4(s + 2) 4s 2s  Similarly, taking the Laplace transform of the ramp input yields R(s) =  1 . s2  Therefore G(s) = DP2.4  Y (s) 1 = . R(s) (s + 1)(s + 2)  For an ideal op-amp, at node a we have vin − va vo − va + =0, R1 R1 and at node b vin − vb = C v̇b , R2 from it follows that   1 1 + Cs Vb = Vin . R2 R2   Also, for an ideal op-amp, Vb − Va = 0. Then solving for Vb in the above equation and substituting the result into the node a equation for Va yields Vo = Vin  1 R2  "  2 1 − R + Cs 2  1 R2  + Cs 2  or Vo (s) R2 Cs − 1 =− . Vin (s) R2 Cs + 1  #  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  75  Design Problems  For vin (t) = At, we have Vin (s) = A/s2 , therefore 2 2 vo (t) = A e−βt + t − β β     where β = 1/R2 C. DP2.5  The equation of motion describing the motion of the inverted pendulum (assuming small angles) is ϕ̈ +  g ϕ=0. L  Assuming a solution of the form ϕ = k cos ϕ, taking the appropriate derivatives and substituting the result into the equation of motion yields the relationship ϕ̇ =  r  g . L  If the period is T = 2 seconds, we compute ϕ̇ = 2π/T . Then solving for L yields L = 0.99 meters when g = 9.81 m/s2 . So, to fit the pendulum into the grandfather clock, the dimensions are generally about 1.5 meters or more.  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  76  CHAPTER 2  Mathematical Models of Systems  Computer Problems CP2.1  The m-file script is shown in Figure CP2.1. pq = 1 9 P= -5 -2 Z= -2 value = 4  p=[1 7 10]; q=[1 2]; % Part (a) pq=conv(p,q) % Part (b) P=roots(p), Z=roots(q) % Part (c) value=polyval(p,-1)  24  20  FIGURE CP2.1 Script for various polynomial evaluations.  The m-file script and step response is shown in Figure CP2.2. numc = [1]; denc = [1 1]; sysc = tf(numc,denc) numg = [1 2]; deng = [1 3]; sysg = tf(numg,deng) % part (a) sys_s = series(sysc,sysg); sys_cl = feedback(sys_s,[1]) % part (b) step(sys_cl); grid on  Transfer function: s+2 ------------s^2 + 5 s + 5  Step Response From: U(1) 0.4  0.35  0.3  To: Y(1)  0.25  Amplitude  CP2.2  0.2  0.15  0.1  0.05  0  0  0.5  1  1.5  2  Time (sec.)  FIGURE CP2.2 Step response.  2.5  3  3.5  4  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  77  Computer Problems  Given ÿ + 4ẏ + 3y = u with y(0) = ẏ = 0 and U (s) = 1/s, we obtain (via Laplace transform) Y (s) =  s(s2  1 1 = . + 4s + 3) s(s + 3)(s + 1)  Expanding in a partial fraction expansion yields Y (s) =  1 1 1 − − . 3s 6(s + 3) 2(s + 1)  Taking the inverse Laplace transform we obtain the solution y(t) = 0.3333 + 0.1667e−3t − 0.5e−t . The m-file script and step response is shown in Figure CP2.3.  Step Response 0.35  0.3  0.25  Amplitude  CP2.3  n=[1]; d=[1 4 3]; sys = tf(n,d); t=[0:0.1:5]; y = step(sys,t); ya=0.3333+0.1667*exp(-3*t)-0.5*exp(-t); plot(t,y,t,ya); grid; title('Step Response'); xlabel('Time (sec)'); ylabel('Amplitude');  0.2  0.15  0.1  0.05  0  0  FIGURE CP2.3 Step response.  0.5  1  1.5  2  2.5 3 Time (sec)  3.5  4  4.5  5  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  78  CHAPTER 2  CP2.4  Mathematical Models of Systems  The mass-spring-damper system is represented by mẍ + bẋ + kx = f . Taking the Laplace transform (with zero initial conditions) yields the transfer function X(s)/F (s) =  s2  1/m . + bs/m + k/m  The m-file script and step response is shown in Figure CP2.4. m=10; k=1; b=0.5; num=[1/m]; den=[1 b/m k/m]; sys = tf(num,den); t=[0:0.1:150]; step(sys,t) Step Response From: U(1) 1.8  1.6  1.4  1 To: Y(1)  Amplitude  1.2  0.8  0.6  0.4  0.2  0  0  50  100  150  Time (sec.)  FIGURE CP2.4 Step response.  CP2.5  The spacecraft simulations are shown in Figure CP2.5. We see that as J is decreased, the time to settle down decreases. Also, the overhoot from 10o decreases as J decreases. Thus, the performance seems to get better (in some sense) as J decreases.  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  79  Computer Problems Nominal (solid); Off-nominal 80% (dashed); Off-nominal 50% (dotted) 18 16  Spacecraft attitude (deg)  14 12 10 8 6 4 2 0 0  10  20  30  40  50  60  70  80  90  100  Time (sec)  %Part (a) a=1; b=8; k=10.8e+08; J=10.8e+08; num=k*[1 a]; den=J*[1 b 0 0]; sys=tf(num,den); sys_cl=feedback(sys,[1]); % % Part (b) and (c) t=[0:0.1:100]; % % Nominal case f=10*pi/180; sysf=sys_cl*f ; y=step(sysf,t); % % Off-nominal case 80% J=10.8e+08*0.8; den=J*[1 b 0 0]; sys=tf(num,den); sys_cl=feedback(sys,[1]); sysf=sys_cl*f ; y1=step(sysf,t); % % Off-nominal case 50% J=10.8e+08*0.5; den=J*[1 b 0 0]; sys=tf(num,den); sys_cl=feedback(sys,[1]); sysf=sys_cl*f ; y2=step(sysf,t); % plot(t,y*180/pi,t,y1*180/pi,'--',t,y2*180/pi,':'),grid xlabel('Time (sec)') ylabel('Spacecraft attitude (deg)') title('Nominal (solid); Off-nominal 80% (dashed); Off-nominal 50% (dotted)')  FIGURE CP2.5 Step responses for the nominal and off-nominal spacecraft parameters.  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  80  CHAPTER 2  CP2.6  Mathematical Models of Systems  The closed-loop transfer function is T (s) =  4s6 + 8s5 + 4s4 + 56s3 + 112s2 + 56s , ∆(s) p= 7.0709 -7.0713 1.2051 + 2.0863i 1.2051 - 2.0863i 0.1219 + 1.8374i 0.1219 - 1.8374i -2.3933 -2.3333 -0.4635 + 0.1997i -0.4635 - 0.1997i  num1=[4]; den1=[1]; sys1 = tf(num1,den1); num2=[1]; den2=[1 1]; sys2 = tf(num2,den2); num3=[1 0]; den3=[1 0 2]; sys3 = tf(num3,den3); num4=[1]; den4=[1 0 0]; sys4 = tf(num4,den4); num5=[4 2]; den5=[1 2 1]; sys5 = tf(num5,den5); num6=[50]; den6=[1]; sys6 = tf(num6,den6); num7=[1 0 2]; den7=[1 0 0 14]; sys7 = tf(num7,den7); sysa = feedback(sys4,sys6,+1); sysb = series(sys2,sys3); sysc = feedback(sysb,sys5); sysd = series(sysc,sysa); syse = feedback(sysd,sys7); sys = series(sys1,syse) poles % pzmap(sys) % p=pole(sys) z=zero(sys)  z= 0 1.2051 + 2.0872i 1.2051 - 2.0872i -2.4101 -1.0000 + 0.0000i -1.0000 - 0.0000i  Polezero map 2.5  2  1.5  1  Imag Axis  0.5  0  -0.5  -1  -1.5  -2  -2.5 -8  -6  -4  -2  0  Real Axis  FIGURE CP2.6 Pole-zero map.  2  4  6  8  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  81  Computer Problems  where ∆(s) = s10 + 3s9 − 45s8 − 125s7 − 200s6 − 1177s5 − 2344s4 − 3485s3 − 7668s2 − 5598s − 1400 . CP2.7  The m-file script and plot of the pendulum angle is shown in Figure CP2.7. With the initial conditions, the Laplace transform of the linear system is θ(s) =  s2  θ0 s . + g/L  To use the step function with the m-file, we can multiply the transfer function as follows: θ(s) =  s2 θ0 , 2 s + g/L s  which is equivalent to the original transfer function except that we can use the step function input with magnitude θ0 . The nonlinear response is shown as the solid line and the linear response is shown as the dashed line. The difference between the two responses is not great since the initial condition of θ0 = 30◦ is not that large.  30  L=0.5; m=1; g=9.8; theta0=30; % Linear simulation sys=tf([1 0 0],[1 0 g/L]); [y,t]=step(theta0*sys,[0:0.01:10]); % Nonlinear simulation [t,ynl]=ode45(@pend,t,[theta0*pi/180 0]); plot(t,ynl(:,1)*180/pi,t,y,'--'); xlabel('Time (s)') ylabel('\theta (deg)')  20  θ (deg)  10  0  -10  function [yd]=pend(t,y) L=0.5; g=9.8; yd(1)=y(2); yd(2)=-(g/L)*sin(y(1)); yd=yd';  -20  -30  0  2  4  6 Time (s)  FIGURE CP2.7 Plot of θ versus xt when θ0 = 30◦ .  8  10  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  82  CHAPTER 2  CP2.8  Mathematical Models of Systems  The system step responses for z = 5, 10, and 15 are shown in Figure CP2.8.  z=5 (solid), z=10 (dashed), z=15 dotted) 1.5  x(t)  1  0.5  0  0  0.5  1  1.5  2  2.5 3 Time (sec)  3.5  4  4.5  5  FIGURE CP2.8 The system response.  CP2.9  (a,b) Computing the closed-loop transfer function yields  T (s) =  G(s) s2 + 2s + 1 = 2 . 1 + G(s)H(s) s + 4s + 3  The poles are s = −3, −1 and the zeros are s = −1, −1. (c) Yes, there is one pole-zero cancellation. The transfer function (after pole-zero cancellation) is  T (s) =  s+1 . s+3  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  83  Computer Problems  Pole?Zero Map 1  0.8  0.6  0.4  Imaginary Axi s  0.2  0  ?-0.2  ?-0.4  ?-0.6  ?-0.8  ?-1 ?-3  ?-2.5  ?-2  ?-1.5  ?-1  ?-0.5  0  Real Axi s  ng=[1 1]; dg=[1 2]; sysg = tf(ng,dg); nh=[1]; dh=[1 1]; sysh = tf(nh,dh); sys=feedback(sysg,sysh) % pzmap(sys) % pole(sys) zero(sys)  >> Transfer function: s^2 + 2 s + 1 ------------s^2 + 4 s + 3  poles  p= -3 -1  zeros  z= -1 -1  FIGURE CP2.9 Pole-zero map.  CP2.10  Figure CP2.10 shows the steady-state response to a unit step input and a unit step disturbance. We see that K = 1 leads to the same steady-state response.  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  CHAPTER 2  Mathematical Models of Systems  0.35  K=[0.1:0.1:10]; sysg=tf([1],[1 20 20]); for i=1:length(K) nc=K(i); dc=[1];sysc=tf(nc,dc); syscl=feedback(sysc*sysg,1); systd=feedback(sysg,sysc); y1=step(syscl); Tf1(i)=y1(end); y2=step(systd); Tf2(i)=y2(end); end plot(K,Tf1,K,Tf2,'--') xlabel('K') ylabel('Steady-state response')  0.3  0.25 Steady−state response  84  0.2  0.15  0.1  Disturbance Response Steady-State 0.05  K=1 0  FIGURE CP2.10 Gain K versus steady-state value.  Input Response Steady-State  0  1  2  3  4  5 K  6  7  8  9  10  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  C H A P T E R  3  State Variable Models  Exercises E3.1  One possible set of state variables is (a) the current iL2 through L2 , (b) the voltage vC2 across C2 , and (c) the current iL1 through L1 . We can also choose vC1 , the voltage across C1 as the third state variable, in place of the current through L1 .  E3.2  We know that the velocity is the derivative of the position, therefore we have dy =v , dt and from the problem statement dv = −k1 v(t) − k2 y(t) + k3 i(t) . dt This can be written in matrix form as               0 1 y 0 d  y    + i . = dt v −k2 −k1 v k3  Define u = i, and let k1 = k2 = 1. Then,  ẋ = Ax + Bu where   A=  0  1  −1 −1     ,    B=  0 k3     , and    x=  y v     .  85  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  86  CHAPTER 3  E3.3  State Variable Models  The charactersitic roots, denoted by λ, are the solutions of det(λI − A) = 0. For this problem we have   λ  det(λI − A) = det   −1  1 λ+2     = λ(λ + 2) + 1 = λ2 + 2λ + 1 = 0 .  Therefore, the characteristic roots are λ1 = −1 and λ2 = −1 . E3.4  The system in phase variable form is ẋ = Ax + Bu y = Cx where   E3.5    A=    0 0    1  0   1   ,  0  −8 −6 −4         0     B=  0  ,    C=  20  h  1 0 0  i  .  From the block diagram we determine that the state equations are ẋ2 = −(f k + d)x2 + ax1 + f u ẋ1 = −kx2 + u and the output equation is y = bx2 . Therefore, ẋ = Ax + Bu y = Cx + Du , where   A= E3.6  0  −k  a −(f k + d)     ,    B=  1 f     ,  C=  h  0 b  i  (a) The state transition matrix is Φ(t) = eAt = I + At +  1 2 2 A t + ··· 2!  and D = [0] .  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  87  Exercises  But A2 = 0, thus A3 = A4 = · · · = 0. So,   Φ(t) = eAt = I + At =   1 0 0 1      +  0 1 0 0      t =   1 t 0 1     .  (b) The state at any time t ≥ 0 is given by x(t) = Φ(t)x(0) and since x1 (0) = x2 (0) = 1, we determine that x1 (t) = x1 (0) + tx2 (0) = 1 + t x2 (t) = x2 (0) = 1 . E3.7  The state equations are ẋ1 = x2 x˙2 = −100x1 − 20x2 + u or, in matrix form   ẋ =   0  1  −100 −20      x +   0 1    u .  So, the characteristic equation is determined to be   det(λI − A) = det   λ  −1  100 λ + 20     = λ2 + 20λ + 100 = (λ + 10)2 = 0 .  Thus, the roots of the characteristic equation are λ1 = λ2 = −10 . E3.8  The characteristic equation is    λ −1  det(λI − A) = det   0 λ   0  6  0 −1 λ+3       = λ(λ2 + 3λ + 6) = 0 .    Thus, the roots of the characteristic equation are λ1 = 0 ,  λ2 = −1.5 + j1.9365 and λ3 = −1.5 − j1.9365 .  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  88  CHAPTER 3  E3.9  State Variable Models  Analyzing the block diagram yields 1 ẋ1 = −x1 + x2 + r 2 3 ẋ2 = x1 − x2 − r 2 3 y = x1 − x2 − r. 2 In state-variable form we have   ẋ =   −1 1    1 2 − 32    x+    1  r ,  −1  The characteristic equation is   h i 3 x + −1 r . y= 1 − 2   5 1 s2 + s + 1 = (s + 2)(s + ) = 0 . 2 2 E3.10  (a) The characteristic equation is   det[λI − A] = det   λ    −6   = λ(λ+ 5)+ 6 = (λ+ 2)(λ+ 3) = 0 .  1 (λ + 5)  So, the roots are λ1 = −2 and λ2 = −3. (b) We note that −1  Φ(s) = [sI − A]    =  s  −6  1 s+5  −1       s+5 6 1   . = (s + 2)(s + 3) −1 s  Taking the inverse Laplace transform yields the transition matrix   E3.11  Φ(t) =   3e−2t − 2e−3t −e−2t  +  6e−2t − 6e−3t  e−3t  −2e−2t  +  3e−3t  A state variable representation is     .  ẋ = Ax + Br y = Cx where   A=  0  1  −12 −8     ,    B=  0 1     ,  C=  h  12 4  i  .  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  89  Exercises  The equation of motion is L  di + Ri + vc = vin dt  where vc =  1 C  Z  i dt .  Unit step response 1.8 1.6 1.4 1.2 State response  E3.12  x1: capacitor voltage 1 0.8 0.6 0.4 0.2 x2: inductor current  0 −0.2  0  0.05  0.1  0.15 0.2 Time(sec)  0.25  0.3  FIGURE E3.12 State variable time history for a unit step input.  Selecting the state variables x1 = vc and x2 = i, we have 1 x2 C R 1 1 ẋ2 = − x2 − x1 + vin . L L L ẋ1 =  This can be written in matrix form as   ẋ =   0  1/C  −1/L −R/L      x +   0 1/L     vin .  0.35  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  90  CHAPTER 3  State Variable Models  When C = 0.001F , R = 4Ω, and L = 0.1H, we have   ẋ =   0  1000  −10  −40      x+  0 10     vin .  The step response is shown in Figure E3.12. E3.13  (a) Select the state variables as x1 = y and x2 = ω. (b) The corresponding state equation is ẋ1 = −x1 − ax2 + 2u ẋ2 = bx1 − 4u or, in matrix form   ẋ =   −1 −a b  0      x +   2 −4      u  and x =   x1 x2     .  (c) The characteristic equation is   det[λI − A] = det   λ+1 a −b  λ     = λ2 + λ + ab = 0 .  So, the roots are 1 1√ 1 − 4ab . λ=− ± 2 2 E3.14  Assume that the mass decay is proportional to the mass present, so that Ṁ = −qM + Ku where q is the constant of proportionality. Select the state variable, x, to be the mass, M . Then, the state equation is ẋ = −qx + Ku .  E3.15  The equations of motion are mẍ + kx + k1 (x − q) + bẋ = 0 mq̈ + kq + bq̇ + k1 (q − x) = 0 .  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  91  Exercises  In state variable form we have   0    (k+k1 )  −  m ẋ =   0   k1 m  1  0  0  b −m  k1 m  0  0  0  1  0  1) − (k+k m  b −m  where x1 = x, x2 = ẋ, x3 = q and x4 = q̇. E3.16        x     The governing equations of motion are m1 ẍ + k1 (x − q) + b1 (ẋ − q̇) = u(t) m2 q̈ + k2 q + b2 q̇ + b1 (q̇ − ẋ) + k1 (q − x) = 0 . Let x1 = x, x2 = ẋ, x3 = q and x4 = q̇. Then,   0     − k1  ẋ =  m1  0   k1 m2    1  0  0  b1 −m 1  k1 m1  b1 m1  0  0  1  b1 m2  2) − (k1m+k 2  − (b1m+b2 2 )  E3.17  h  0 0 1 0  i  x.  At node 1 we have C1 v̇1 =  va − v1 v2 − v1 + R1 R2  C2 v˙2 =  vb − v2 v1 − v2 + . R3 R2  and at node 2 we have  Let x 1 = v1 and x 2 = v2 .     0     1       m    x +  1  u(t) .  0          Since the output is y(t) = q(t), then y=    0  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  92  CHAPTER 3  State Variable Models  Then, in matrix form we have    −  E3.18  ẋ =     1 R1 C1  +  1 R2 C1  − R21C2    1 R2 C1  −    1 R3 C2  +  1 R2 C2  The governing equations of motion are         x+  1 R1 C1  0  0  1 R3 C2     va vb     .  di1 + v = va dt di2 L2 + v = vb dt dv . iL = i1 + i2 = C dt  Ri1 + L1  Let x1 = i1 , x2 = i2 , x3 = v, u1 = va and u2 = vb . Then,     ẋ =     y=  E3.19  h  − LR1  0  0  0  1 C  1 C  0 0 1  First, compute the matrix    − L11 − L12  1 L1      x+ 0      0 i    0   1 L2  0  0  x + [0] u .    sI − A =   s  −1  3 s+4     u       .  Then, Φ(s) is Φ(s) = (sI − A)−1 where ∆(s) = s2 + 4s + 3, and G(s) = E3.20  h  10 0  i     s+4 ∆(s) 3 − ∆(s)      1  s+4 1  = ∆(s) −3 s 1 ∆(s) s ∆(s)     0 1    =  s2  10 . + 4s + 3  The linearized equation can be derived from the observation that sin θ ≈ θ when θ ≈ 0. In this case, the linearized equations are θ̈ +  g k θ + θ̇ = 0 . L m  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  93  Exercises  Let x1 = θ and x2 = θ̇. Then in state variable form we have ẋ = Ax y = Cx where   A=  E3.21  0  1  −g/L −k/m  The transfer function is     ,  C=  h  1 0  i  ,  G(s) = C [sI − A]−1 B + D =  and  s2    x(0) =   θ(0) θ̇(0)    .  −1 . + 2s + 1  The unit step response is y(t) = −1 + e−t + te−t . E3.22  The transfer function is G(s) =  s2  s−6 . − 7s + 6  The poles are at s1 = 1 and s2 = 6. The zero is at s = 6. So, we see that there is a pole-zero cancellation. We can write the system in state variable form as √ ẋ = x − 2u √ 2 y=− x 2 and the transfer function is G(s) = E3.23  1 . s−1  The system in state variable form can be represented by ẋ = Ax + Bu y = Cx + Du where     A=    0  1  0  0    0   1   ,  −1 −3 −3         0     B=  0  ,    1  C=  h  0 1 −1  i  ,  D=  h  1  i  .  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  94  CHAPTER 3  State Variable Models  + -  U(s)  + - --  1 s  x3  3 3  FIGURE E3.23 Block diagram.  1 s  x2  1 s  x1  +  X(s)  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  95  Problems  Problems P3.1  The loop equation, derived from Kirchoff's voltage law, is di 1 R 1 = v− i − vc dt L L L where vc =  1 C  Z  i dt .  (a) Select the state variables as x1 = i and x2 = vc . (b) The corresponding state equations are R 1 1 v− x1 − x2 L L L 1 x1 . ẋ2 = C  ẋ1 =  (c) Let the input u = v. Then, in matrix form, we have   ẋ =   −R/L −1/L 1/C  0      x+  1/L 0    u .  -R/L 1/C  v 1/s  1/L  x1  1/s x2  -1/L  FIGURE P3.1 Signal flow graph.  P3.2  Let −2 −2R1 R2 , a22 = , (R1 + R2 )C (R1 + R2 )L 1 R2 = b12 = , b21 = −b22 = . (R1 + R2 )C (R1 + R2 )L  a11 = b11  The corresponding block diagram is shown in Figure P3.2.  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  96  CHAPTER 3  State Variable Models  2/(R1+R2)C  v1  1/s  1/(R1+R2)C  x1  R2  (a) 1/(R1+R2)C  v2  1/s  x2  -  2R1R2/(R1+R2)C  a 11 v1  b11 x1  1/s  b12 (b) b21 v2  1/s x2  b22 a 22  FIGURE P3.2 (a) Block diagram. (b) Signal flow graph.  P3.3  Using Kirchoff's voltage law around the outer loop, we have L  diL − vc + v2 − v1 = 0 . dt  Then, using Kirchoff's current law at the node, we determine that C  dvc = −iL + iR , dt  where iR is the current through the resistor R. Considering the right loop we have iR R − v2 + vc = 0  or  iR = −  vc v2 + . R R  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  97  Problems  Thus, dvc vc iL v2 =− − + dt RC C RC  vc v1 v2 diL = + − . dt dt L L  and  In matrix form, the state equations are    ẋ1 ẋ2      =  0  1/L  −1/C −1/RC     x1 x2      +  1/L −1/L 0  1/RC     v1 v2     ,  where x1 = iL and x2 = vc . The signal flow graph is shown in Figure P3.3.  v1 1/L -1/L -1/C  1/s  v2 1/L  x2  1/s  x1 -1/RC  1/RC  FIGURE P3.3 Signal flow graph.  P3.4  (a) The block diagram model for phase variable form is shown in Figure P3.4a. The phase variable form is given by     ẋ =     y=  h  0  1  0  0  −10 −6 −4 10 2 1  i            0   0       1  x +  0 r  x.  1    (b) The block diagram in input feedforward form is shown in Figure P3.4b. The input feedforward form is given by     ẋ =     y=  h            −4 1 0   1       −6 0 1   x +  2  r(t)  −10 0 0 1 0 0  i  x.  10    © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  98  CHAPTER 3  State Variable Models  1 2 x2  x3 R(s)  1 s  + -  --  1 s  x1  1 s  +  10  + +  Y(s)  4 6 10 (a) 1 2  R(s)  10  +  . . x3 1 + + x2 s -  1 + s -  +  . x1  1 s  Y(s)  4 6 10 (b)  FIGURE P3.4 (a)Block diagram model for phase variable form. (b) Block diagram model for input feedforward form.  P3.5  (a) The closed-loop transfer function is T (s) =  s3  s+1 . + 4s2 − 11s + 1  (b) A matrix differential equation is ẋ = Ax + Bu y = Cx where     A=    0  1  0  0    0   1   ,  −1 11 −4             0     B=  0  ,  1  The block diagram is shown in Figure P3.5.  C=  h  1 1 0  i  .  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  99  Problems  1 R(s)  -  x3  1 s  + --  x2  1 s  x1  1 s  +  1  +  Y(s)  4 -11 1  FIGURE P3.5 Block diagram model.  P3.6  The node equations are dv1 vi − v1 + iL − =0 dt 4000 dv2 v2 0.0005 − iL + − i3 = 0 dt 1000 diL 0.002 + v2 − v1 = 0 . dt 0.00025  Define the state variables x 1 = v1  x 2 = v2  x3 = iL .  Then, ẋ = Ax + Bu where    −1  A= 0   P3.7      0 −4000   −2 2000   ,  500 −500  Given K = 1, we have KG(s) ·  0        0   1    B=  0 2000    0  0    (s + 1)2 1 = . s s(s2 + 1)  We then compute the closed-loop transfer function as T (s) =  s2 + 2s + 1 s−1 + 2s−2 + s−3 = . 3s3 + 5s2 + 5s + 1 3 + 5s−1 + 5s−2 + s−3  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  100  CHAPTER 3  State Variable Models  A state variable model is     ẋ =     y=  P3.8  h  0  1  0  0  −1/3 −5/3 −5/3 1 2 1  The state-space equations are  i            0   0       1  x +  0 r  x.  1/3    ẋ1 = x2 ku ẋ2 = −g x3 ẋ3 = u . This is a set of nonlinear equations. P3.9  (a) The closed-loop transfer function is T (s) =  10s−3 10 = , Js3 + (b + 10J)s2 + 10bs + 10K1 1 + 10.1s−1 + s−2 + 5s−3  where K1 = 0.5, J = 1, and b = 0.1. (b) A state-space model is     ẋ =     ω=  h  0  1  0  0  1 0 0          0   0       1  x +  0 r  −5 −1 −10.1 i    x.  10    (c) The characteristic equation is    s −1  det[sI − A] = det   0 s   5  1  0 −1 s + 10.1       = s3 + 10.1s2 + s + 5 = 0 .    The roots of the characteristic equation are s1 = −10.05  and s2,3 = −0.0250 ± 0.7049j .  All roots lie in the left hand-plane, therefore, the system is stable.  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  101  Problems  P3.10  (a) From the signal flow diagram, we determine that a state-space model is given by   −K1      ẋ =  y=    K2  x +   −K1 −K2 y1 y2    =      K1 −K2 K1    K2    1 0  r1 r2     x .  0 1  (b) The characteristic equation is det[sI − A] = s2 + (K2 + K1 )s + 2K1 K2 = 0 . (c) When K1 = K2 = 1, then   A=  −1  1  −1 −1     .  The state transition matrix associated with A is   o  n  Φ = L−1 [sI − A]−1 = e−t   P3.11  cos t  sin t  − sin t cos t  The state transition matrix is   Φ(t) =   (2t − 1)e−t 2te−t  −2te−t  (−2t + 1)e−t     .  So, when x1 (0) = x2 (0) = 10, we have x(t) = Φ(t)x(0) or x1 (t) = 10e−t x2 (t) = 10e−t P3.12  (a) A state variable representation is given by     ẋ =     0  1  0  0            0   0       1  x +  0 r  −48 −44 −12  1       .  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  102  CHAPTER 3  State Variable Models  y = [40 8 0]x . (b) The state transition matrix is Φ(t) =    . . Φ1 (t)..Φ2 (t)..Φ3 (t)    ,  where     e−6t − 3e−4t + 3e−2t    −6t + 12e−4t − 6e−2t Φ1 (t) =   −6e   36e−6t − 48e−4t + 12e−2t   P3.13    Φ3 (t) =               Φ2 (t) =     1 −6t − 41 e−4t + 18 e−2t 8e − 43 e−6t + e−4t − 14 e−2t 9 −6t − 4e−4t + 12 e−2t 2e  (a) The RLC circuit state variable representation is   ẋ =   −10 −4 6  0      x+  4 0  3 −6t − 2e−4t + 54 e−2t 4e − 29 e−6t + 8e−4t − 25 e−2t  27e−6t − 32e−4t + 5e−2t       .      u .  The characteristic equation is s2 + 10s + 24 = 0 . All roots of the characteristic equation (that is, s1 = −4 and s2 = −6) are in the left half-plane; therefore the system is stable. (b) The state transition matrix is   Φ(t) =   3e−6t − 2e−4t  −3e−6t  +  3e−4t  2e−6t + 2e−4t −2e−6t  +  3e−4t     .  (c) Given x1 (0) = 0.1 ,  x2 (0) = 0 and e(t) = 0 ,  we have i(t) = x1 (t) = 0.3e−6t − 0.2e−4t vc (t) = x2 (t) = −0.3e−6t + 0.3e−4t .         © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  103  Problems  (d) When x(0) = 0 and u(t) = E, we have x(t) =  Z  t  Φ(t − τ )Bu(τ )dτ ,  0  where   Bu(t) =   Integrating yields  4E 0     .  x1 (t) = (−2e−6t + 2e−4t )E x2 (t) = (1 + 2e−6t − 3e−4t )E . P3.14  A state space representation is ẋ = Ax + Br ,  y = Cx  where   P3.15      A=     0  1  0  0  0  1  0  0  0      0   0     ,  1    −50 −34 −10 −12     ,  0       B=  A state variable representation is     ẋ =     0  1  0  0     0     0     C = [50 1 0 0] .  1            0   0       1  x+ 0 r  −16 −31 −10  y = [56 14 0]x . The block diagram is shown in Figure P3.15.  1    © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  104  CHAPTER 3  State Variable Models  14 R(s)  1 s  + -  --  x3  x2  1 s  1 s  x1  56  +  +  Y(s)  10 31 16  FIGURE P3.15 Block diagram model.  (a) The characteristic equation is  x1 - solid; x2 - dotted; x3 - dashed 0.5 0 -0.5  Step response)  P3.16  -1 -1.5 -2 -2.5 -3 -3.5  0  20  40  60  80  100  Time (s)  FIGURE P3.16 Step response of magnitude 0.285◦ .    s   det(sI − A) = det   0.0071  0  −1 s + 0.111 −0.07  0      −0.12     s + 0.3  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  105  Problems  = s3 + 0.411s2 + 0.032s + 0.00213 = 0 . The roots are s1 = −0.3343  and  s2,3 = −0.0383 ± 0.0700j .  All the poles lie in the left half-plane, therefore, the system is stable. (b) The solution of the system to a step of magnitude 0.285◦ is given by x1 (t) = −2.66 − 0.11e−0.33t + e−0.038t (2.77 cos 0.07t + 0.99 sin 0.07t) x2 (t) = 0.037e−0.33t − e−0.038t (0.037 cos 0.07t + 0.23 sin 0.07t) x3 (t) = 0.069 − 0.075e−0.33t + e−0.038t (0.006 cos 0.07t − 0.06 sin 0.07t) P3.17  The transfer function is G(s) = C(sI − A)−1 B =  P3.18  −4s + 12 . s3 − 14s2 + 37s + 20  Define the state variables as x1 = φ1 − φ2 ω1 x2 = ωo ω2 x3 = . ωo Then, the state equations of the robot are ẋ1 = ωo x2 − ωo x3 −J2 ωo x1 − ẋ2 = J1 + J2 J1 ωo ẋ3 = x2 + J1 + J2  b x2 + J1 b x2 − J2  b Km x3 + i J1 J1 ωo b x3 J2  or, in matrix form   0  1    ẋ = ωo   a−1   a  where a=  J1 , (J1 + J2 )  b1 =  −b1  b2 −b2  b , J1 ωo            −1   0       b1  x +  d i  b2 =  0  b and J2 ωo    d=  Km . J1 ωo  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  106  CHAPTER 3  P3.19  State Variable Models  The state equation is given by   0  ẋ =   1  −2 −3    x  where x1 (0) = 1 and x2 (0) = −1. The state transition matrix is   Φ(t) =   −e−2t + 2e−t −e−2t + e−t 2e−2t  −  2e−t  2e−2t  − e−t     .  The system response is         x1 (t) = −e−2t + 2e−t x1 (0) + −e−2t + e−t x2 (0)         x2 (t) = 2e−2t − 2e−t x1 (0) + 2e−2t − e−t x2 (0) . The state response is shown in Figure P3.19.  1 0.8  x1  0.6  System response  0.4 0.2 0 -0.2 -0.4  x2  -0.6 -0.8 -1  0  1  2  3 Time (s)  4  5  6  FIGURE P3.19 Response with x1 (0) = 1 and x2 (0) = −1.  P3.20  The state equation is given by   ẋ =   − 0.693 6.7 −1  0 − 0.693 9.2    x    where x(0) =   0.3 × 1016 7×  1016     .  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  107  Problems  The state transition matrix is   Φ(t) =   e−0.103433t 35.5786(e−0.103433t  −  0  e−0.0753261t )  The system response is  e−0.075326t     .  x1 (t) = e−0.103433t x1 (0) h  i  x2 (t) = 35.5786 e−0.103433t − e−0.0753261t x1 (0) + e−0.075326t x2 (0) . The state response is shown in Figure P3.20. 7  Nucleide densities in atoms per unit volume  6  X=Xenon 135 I=Iodine 135  5  4  3  2  1  0  -1  0  10  20 30 Time (hours)  40  FIGURE P3.20 Nuclear reactor state response to initial conditions.  50  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  108  CHAPTER 3  P3.21  State Variable Models  Referring to Figure P3.21 we have 1 1 1 Y (s) = W (s) = h1 U (s) + Q(s) s s s 1 h1 U (s) + 2 [h0 U (s) − a0 Y (s) − a1 sY (s) + a1 h1 U (s)] . = s s     Gathering like terms and re-arranging yields   1+  a1 a0 + 2 Y (s) = s s     h1 h0 a1 h1 U (s) + 2 + 2 s s s   or Y (s) =    h1 s + h0 + a1 h1 U (s) . s 2 + a1 s + a0   Computing the transfer function from the state variable representation yields G(s) = C (sI − A)−1 B =  h  1 0  i     s+a1 s2 +a1 s+a0 −a0 s2 +a1 s+a0  1 s2 +a1 s+a0 s s2 +a1 s+a0     h1 h0    =  h1 s + h0 + a1 h1 . s 2 + a1 s + a0  h1  U(s)  Q(s)  h0 + --  1 s  +  +  1 s  W(s)  a1 a0 FIGURE P3.21 Block diagram with labeled signals.  P3.22  The governing equations are L C1  di = v2 dt  dv1 1 1 + (v1 − v) + (v1 − v2 ) = 0 dt R1 R2  Y(s)  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  109  Problems  C2  dv2 1 v2 + (v2 − v1 ) + i + =0. dt R2 R3  Let u = v, x1 = i, x2 = v1 and x3 = v2 . Then,     ẋ =     0 0 − C12  − a1    1 R1  +  1 R2  1 R2 C2      1 L 1 C1 R2  0  −    1 R2 C2  +  1 R3 C2        x+       0 1 R1 C1  0      u    y = [0 0 1]x . P3.23  A state variable representation is given by     ẋ =     0  1  0  0            0   0       1  x +  0 r  −30 −31 −10  y = [1 0 0]x .  1    Other representations include the input feedforward representation              −10 1 0   0         ẋ =   −31 0 1  x +  0  r   −30 0 0  y = [1 0 0]x ,  1    the physical variable representation   1  −3   ẋ =  0 −2  0                  0   0       1 x+  0 r  0 −5  y = [1 0 0]x ,    1    and the decoupled representation   0  −3  ẋ =   0 −2  0    0   1       0  x +  1 r  0 −5  1    © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  110  CHAPTER 3  State Variable Models  y= P3.24    1 1 1 x. − 6 3 2   The matrix representation of the state equations is   ẋ =   3 0 0 2      x +     1 1    0 1  u1 u2      +  0 1    d .  When u1 = 0 and u2 = d = 1, we have ẋ1 = 3x1 + u2 ẋ2 = 2x2 + 2u2 So we see that we have two independent equations for x1 and x2 . With U2 (s) = 1/s and zero initial conditions, the solution for x1 is found to be x1 (t) = L  −1  1 {X1 (s)} = L s(s − 3)    1 1 1 1 −1 =L − + = − 1 − e3t 3s 3 s − 3 3 −1      and the solution for x2 is x2 (t) = L−1 {X2 (s)} = L−1 P3.25    2 s(s − 2)    1 1 = L−1 − + s s−2   = −1+e2t .  Since Φ(s) = (sI − A)−1 , we have   Φ(s) =   s+1 −2  0 s+3  −1     =  s+3 2  0 s+1     1 ∆(s)  where ∆(s) = (s + 1)(s + 3). The state transition matrix is   Φ(t) = L−1 {Φ(s)} =  P3.26    e−t  0  e−t − e−3t e−3t  The state variable differential equation is   ẋ =   0  1  −25 −6  y = [1 0]x .      x +   0 25    r     .  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  111  Problems  and   s+6 1  Φ(s) = (sI − A)−1 =   −25  s    1 ∆(s)    where ∆(s) = s2 + 6s + 25. P3.27  Equating the change in angular momentum to the sum of the external torques yields J θ̈ − Hω cos θ = −bθ̇ − kθ where b is the damping coefficient, k is the spring constant, and J is the wheel moment of inertia. Defining the state variables x1 = θ and x2 = ẋ and the input u = ω, we can write the equations of motion as ẋ1 = x2 k b H ẋ2 = − x1 − x2 + u cos x1 J J J With a small angle assumption (that is, cos x1 ≈ 1) we have   ẋ =   0  1  −k/J  −b/J  y=θ=  P3.28  h  1 0  i  x.      x +     0  u  H/J  The governing equations of motion are m1 y¨1 + k(y1 − y2 ) + by˙1 = u m2 y¨2 + k(y2 − y1 ) + by˙2 = 0 y = y2 . Let x1 = y1 , x2 = ẏ1 , x3 = y2 and x4 = ẏ2 . Then   0     − k  ẋ =  m1  0   k m2  y=  h  1  0  − mb1  k m1  0  0  0  1  − mk2  − mb2  0  0 0 1 0  i  x.  0         0      1       m  x +  1 u   0         0  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  112  CHAPTER 3  P3.29  State Variable Models  The equations of motion are I q¨1 + M gL sin q1 + k(q1 − q2 ) = 0 J q¨2 − k(q1 − q2 ) = u . Let x1 = q1 , x2 = q̇1 , x3 = q2 , and x4 = q̇2 and linearize the equations using small angle assumptions (i.e. sin q1 ≈ q1 ). Then, we have x˙1 = x2 M gL k x˙2 = − x1 − (x1 − x3 ) I I x˙3 = x4 k 1 x˙4 = (x1 − x3 ) + u . J J  P3.30  Using Kirchoff's current law, we find that C  dv c = i2 + i3 dt  where i3 = current in R3 . Let i1 = current in R1 . Using Kirchoff's voltage law, we have L  diL = v1 − R1 i1 dt  and R1 i1 + R2 i2 + vc = v1 . But i2 = i1 − iL , so (R1 + R2 )i1 = v1 − vc + R2 iL . Using Kirchoff's voltage law once again, we calculate i3 as i3 =  v2 − vc . R3  Utilizing the above equations, we can solve for diL /dt and dvc /dt, as follows: diL R2 R1 R1 R2 = v1 + vc − iL dt L(R1 + R2 ) L(R1 + R2 ) L(R1 + R2 ) vc v1 vc vc R1 iL v2 = − − − + dt C(R1 + R2 ) C(R1 + R2 ) CR3 C(R1 + R2 ) CR3  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  113  Problems  Define the state variables x1 = vc and x2 = iL . Then, in matrix form we have   ẋ =   1 +R2 +R3 ) − (R CR3 (R1 +R2 )  R1 L(R1 +R2 )  y = i2 = P3.31  h  1 − (R1 +R 2)    1 − C(RR1 +R 2)  x+  R1 R2 − L(R 1 +R2 ) 1 − (R1R+R 2)    i  x+  h  1 CR3  1 C(R1 +R2 ) R2 L(R1 +R2 )  1 (R1 +R2 )  0 i  0        v1 v2  v1 v2         A state variable representation is   ẋ =   0    1    x +   −3 −4  0 30    u .  The state transition matrix can be computed as follows: n  Φ = L−1 [sI − A]−1  o     s      1 s+4  = L−1  ∆(s) −3   =  3 −t − 21 e−3t 2e − 32 e−t + 23 e−3t  1  1 −t − 21 e−3t 2e − 12 e−t + 32 e−3t     where ∆(s) = s2 + 4s + 3 = (s + 1)(s + 3) . P3.32  A state variable representation is ṁ1 = −k1 m1 + r ṁ2 = k1 m1 − k2 m2 where k1 and k2 are constants of proportionality. In matrix form, we have   ẋ = Ax + Br =   −k1  0  k1 −k2      x+  1 0    r  where x1 = m1 and x2 = m2 . Let k1 = k2 = 1 and assume that r(t) = 0 and x1 = 1 and x2 = 0. Then   x(t) = Φ(t)x(0) =   e−t  0  te−t  e−t       x(0) =   e−t te−t     .  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  114  CHAPTER 3  State Variable Models  The simulation is shown in Figure P3.32.  0.4  1 0.9  0.35  0.8 0.3 0.25  x1  0.6  x2  state history, x(t)  0.7  0.5 0.4  0.2 0.15  0.3 0.1 0.2 0.05  x2  0.1 0 0  5  t=0  0 0  10  0.5  time (sec)  1  x1  FIGURE P3.32 Actual versus approximate state response.  P3.33  The system (including the feedback) is described by   ẋ = Ax =   0  1  −1/2 −1    x .  The charactersitic equation is   det[λI − A] = det   λ  −1  1/2 λ + 1     = λ2 + λ +  The roots of the characteristic equation are 1 1 λ1,2 = − ± j . 2 2  1 =0. 2  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  115  Problems  The system response is   x(t) = eAt x(0) =   e−t/2 cos 2t + e−t/2 sin 2t  2e−t/2 sin 2t  −e−t/2 sin 2t  e−t/2 cos 2t − e−t/2 sin 2t    = e−t/2   2 sin 2t cos 2t − sin 2t       x(0)    where x1 (0) = 0 and x2 (0) = 1. P3.34  (a) The state space representation is   1    −6 −11 −6   0  ẋ =   0            0   0       1  x +  0 r  0  y = [6 0 0] x .  1    (b) The element φ11 (t) of the state transition matrix is φ11 (t) = e−3t − 3e−2t + 3e−t . P3.35  The state equations are 1 8 [80θ − 50h] = −x1 + x2 50 5 θ̇ = ẋ2 = ω = x3 Km Km Kb Km Ka 353 25000 ω̇ = ẋ3 = ia = − ω+ vi = − x3 + vi . J JRa JRa 30 3 ḣ = ẋ1 =  In state variable form, we have   8  −1 5  ẋ =   0 0  P3.36    0  0 1  0 − 353 30          x+      0 0 25000 3       vi .    Using Newton's Law and summing the forces on the two masses yields M1 ẍ(t) + b1 ẋ(t) + k1 x(t) = b1 ẏ(t) M2 ÿ(t) + b1 ẏ(t) + k2 y(t) = b1 ẋ(t) + u(t) Let z1 = x, z2 = ẋ, z3 = y, and z4 = ẏ .  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  116  CHAPTER 3  State Variable Models  Then we write the system in state variable form as   0     − k1  ż =  M1  0    1  0  0  b1 −M 1  0  b1 M1  0  0  1  b1 M2  k2 −M 2  b1 −M 2  0  y= P3.37  h  1 0 0 0  i  z.         0     0         z +  u   0        1 M2  From the block diagram in Figure P3.37, we obtain ẋ1 ẋ2 ẋ3 y  = x2 = x3 = −10x1 − 4x2 − 3x3 + u = x1 + 12x2 + 5x3  or     ẋ =     0  1  0  0  −10 −4 −3  y = [1 12 5] x .            0   0       1  x +  0 u 1    The third-order differential equation model is ... y +3ÿ + 4ẏ + 10y = 5ü + 12u̇ + u .  5 12  U(s)  + - --  ∫  x3  ∫  x2  3 4 10  FIGURE P3.37 Block diagram with states labeled.  ∫  x1  ++ +  Y(s)  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  117  Advanced Problems  Advanced Problems AP3.1  With the state variables are defined as      x     z=  ẋ  ,    i  the nonlinear equations of motion are            ż1       ż  =  g −  2     ż3  z2 K (Io +z3 )2 m (Xo +z1 )2  1 L (v  − Rz3 )       ,    where the control is the voltage v. We assume that z1 = x is measurable. The linearized equations of motion are ż = Az + Bv y = Cz  where     A=    0  1  0  2K Io2 m Xo3  0  0  0  Io − 2K m Xo2 −R L  The transfer function is       ,         0  h i    B= .  0  , and C = 1 0 0   1 L    G(s) = C(sI − A)−1 B . With the constants R = 23.2 L = 0.508 m = 1.75 K = 2.9 × 10−4 Io = 1.06 Xo = 4.36 × 10−3  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  118  CHAPTER 3  State Variable Models  the transfer function is G(s) = AP3.2  s3  +  −36.38 . + 4493s + 205195  45.67s2  The differential equation describing the motion of y is mÿ + bẏ + ky = bu̇ + ku . Taking Laplace tranforms (with zero initial conditions) yields the transfer function Y (s) (b/m)s + (k/m) = 2 . U (s) s + (b/m)s + (k/m) In state space form, we have   ẋ =  y=  AP3.3  h  0  1  −k/m −b/m k/m b/m  i      x +   0 1  x.    u  The transfer function is Y (s) 2s2 + 6s + 5 = 3 . R(s) s + 4s2 + 5s + 2 In (nearly) diagonal form, we have   1  −1  A=  0 −1  0    0   0   ,  0 −2         0     B=  1  , and    1  C=  h  1 1 1  i  .  The matrix A is not exactly diagonal due to the repeated roots in the denominator of the transfer function. AP3.4  The differential equations describing the motion of y and q are mÿ + k2 ẏ + k1 (y − q) = f −bq̇ + k1 (y − q) = f where k1 = 2 and k2 = 1. Assume the mass m = 1. Then with the state  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  119  Advanced Problems  variables defined as z =   h  y ẏ q 0 1    ż =   −3 0   iT  y=  1 0 0  i      0      −1/b  0      2 z +    2/b 0 −2/b  h  , we have the state variable model  z    1    f    If we model a large bump at high speeds as an impulse and a small bump at low speeds as a step, then b = 0.8 provides good performance. In both cases, the ride settles out completely in about 10 seconds. AP3.5  The differential equations describing the motion of x and θ are (M + m)ẍ + M L cos θ θ̈ − M L sin θ θ̇ 2 = −kx g sin θ + cos θẍ + Lθ̈ = 0 Assuming θ and θ̇ are small, it follows that (M + m)ẍ + M Lθ̈ = −kx ẍ + Lθ̈ = −gθ Define the state variables as z = able model is   0    −k/m  ż =   0    h  x ẋ θ θ̇  1  0  0  gM/m  0  0  iT  . Then, the state vari  0   0    k/(Lm) 0 −g(M + m)/(Lm) 0  AP3.6 AP3.7  z  1     Computing the closed-loop system yields   A − BK =   −1  1  −K1 −K2     ,  The characteristic polynomial is    B=  0 1     , and  C=  h  2 1  i  .  |sI − (A − BK)| = s2 + (K2 + 1)s + K1 + K2 = 0. The roots are in the left-half plane whenever K2 +1 > 0 and K1 +K2 > 0.  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  120  CHAPTER 3  AP3.8  State Variable Models  (a) A state variable representation is given by x˙1 x˙2 x˙3 y  = x2 = x3 = −Kx1 − 12x2 − 6x3 + Kr = x1  or, in matrix form     ẋ =     y=  h  0  1  0  0  1 0 0  x          0   0       1  x +  0 r  −K −12 −6 i    K    (b) The characteristic roots are found by solving det [λI − A] = 0 or λ3 + 6λ2 + 12λ + K = 0 When K = 8, we have characteristic roots at λ1 = −2, λ2 = −2, and λ3 = −2, as desired. (c) The unit step response is given by y(t) = 1 − e−2t − 2te−2t − 2t2 e−2t .  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  121  Design Problems  Design Problems CDP3.1  The transfer model of the traction drive, capstan roller, and linear slide was given in CDP2.1 as rKm X(s) = , Va (s) s [(Lm s + Rm )(JT s + bm ) + Kb Km ] where JT = Jm + r 2 (Ms + Mb ) . Define x1 = x, x2 = ẋ, and x3 = ẍ. Then, a state variable representation is ẋ = Ax + Bva y = Cx where    0  A=  0   0  C=  DP3.1  h  1  0  0  1  b Km − Rm bLmm+K JT  m JT − Lm bLmm+R JT  1 0 0  i  .       ,        B=    (a) The equation of motion of the spring-mass-damper is mÿ + bẏ + ky = u or ÿ = −  b k 1 ẏ − y + u . m m m  Select the state variables x1 = y  and x2 = ẏ .  Then, we have ẋ = Ax + Bu y = Cx  0 0 rKm Lm J T         © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  CHAPTER 3  State Variable Models  where   A=  0  1  −20 −9       ,  0  B=  1     ,  C=  h  1 0  i  .  A is the system matrix. The characteristic equation is   det[λI − A] = det   s  −1  20 s + 9     = s2 + 9s + 20 = 0 .  The roots of the characteristic equation are s1 = −4 and and the transistion matrix is   Φ(t) =   5e−4t − 4e−5t  e−4t − e−5t  −20e−4t + 20e−5t −4e−4t + 5e−5t  s2 = −5 ,     .  (b) Assume the initial conditions are x1 (0) = 1 and x2 (0) = 2. The zeroinput response is shown in Figure DP3.1. (c) Suppose we redesign the system by choosing b and k to quickly damp out x2 and x1 . We can select b and k to achieve critical damping. Critical damping: b/m=20, k/m=100  b/m=9, k/m=20 2  2  1.5 1 x1  1 0  x1 State response, x  0.5 State response, x  122  0  x2  -1  -0.5 -2 x2  -1  -3 -1.5  -2  0  0.5  1 Time(sec)  FIGURE DP3.1 Zero input state response.  1.5  2  -4  0  0.5  1 Time(sec)  1.5  2  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  123  Design Problems  If we desire the characteristic polynomial to be pd (s) = (s + 10)2 = s2 + 20s + 100, then we need b = 20 and k = 100. DP3.2  The desired transfer function is 6 Y (s) = 2 . U (s) s + 7s + 10 The transfer function derived from the phase variable representation is Y (s) d = 2 . U (s) s + bs + a Therefore, we select d = 6, a = 10 and b = 7. Assume the aircraft lands precisely on the centerline. The linearized equations of motion are m3 ẍ3 + KD ẋ3 + K2 (x3 − x2 ) = 0 m2 ẍ2 + K2 (x2 − x3 ) + K1 (x2 − x1 ) = 0 2 m1 ẍ1 = − √ K2 (x1 − x2 ) 2 where x1 (0) = x2 (0) = ẋ2 (0) = ẋ3 = 0 and ẋ1 (0) = 60. The system response is shown in Figure DP3.3 where KD = 215. The aircraft settles out at 30 m, although initially it overshoots by about 10 m at 1 second.  45 40 35 30  Amplitude  DP3.3  25 20 15 10 5 0 0  1  2  3  4  5 Time (secs)  FIGURE DP3.3 Aircraft arresting gear response.  6  7  8  9  10  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  124  CHAPTER 3  DP3.4  State Variable Models  We can model the bungi cord system as a mass-spring-damper. This is actually an over-simplification because the bungi cord cannot "push" the jumper down as a spring would—it can only exert a restoring force when the cord is stretched (that is, when the jumper exceeds the length, L, of the cord). The problem is nonlinear! When the distance of the jumper from the platform is less than L we should model the cord spring constant and damping as K = 0 and b = 0, respectively. Only gravity acts on the jumper. Also, when ẋ (the jumper velocity) is negative (where we define positive towards the ground), then we should model b = 0. A reasonable set of equations of motion are ẋ1 = x2 K b ẋ2 = − x1 − x2 + g m m where x1 is the distance measured from the top of the platform and x2 is the jumper velocity. For the initial conditions we have x1 (0) = 10 and x2 (0) = 0. A reasonable set of parameters for the bungi cord are L = 40 m, K = 40 N/m and b = 20 kg/m. The system response is shown in Figure DP3.4 for a person with m = 100 kg. The accelerations experienced by the jumper never exceed 1.5 g.  global MASS GRAVITY LENGTH K b MASS=100; HEIGHT=100; GRAVITY=9.806; LENGTH=40; SPRINGCONSTANT=40; SPRINGDAMPING=20; x0=[10;0]; t=0; dt=0.1; n=round(120/dt); for i=1:n; if x0(1)            1. For example, we can take K = 8. (b) The transfer function from Td (s) to ω(s) is given by −10s ω(s) = 2 . Td (s) s + 10s + 100 The error plot is shown in Figure AP4.4, where e(s) = −ω(s) (V (s) = 0.)  0.2 0.18 0.16 0.14  e(t)  0.12 0.1 0.08 0.06 0.04 0.02 0 0  1  2  3  4  5  6  7  8  9  10  Time (sec)  FIGURE AP4.4 Error plot with a ramp disturbance input.  AP4.5  (a) The transfer function from the disturbance Td (s) to the output Y (s) is Y (s) −s = 3 . 2 Td (s) s + 4s + 4s + K The steady-state error (when Td (s) = 1/s) is ess = lim s s→0  s3  +  4s2  1 s =0. + 4s + K s  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  158  CHAPTER 4  Feedback Control System Characteristics  (b) The closed-loop transfer function is Y (s) K = 3 . R(s) s + 4s2 + 4s + K The steady-state error (when R(s) = 1/s2 ) is ess = lim s(1 − T (s)) s→0  s3 + 4s2 + 4s 4 1 = lim = . 2 3 2 s→0 s s(s + 4s + 4s + K) K  (c) Let K = 8. Then, Y (s) −s = 3 . Td (s) s + 4s2 + 4s + 8 The error plot is shown in Figure AP4.5, for r(t) = 0.  0.15  0.1  e(t)  0.05  0  -0.05  -0.1  0  2  4  6  8  10  12  Time (sec)  FIGURE AP4.5 Error plot with a step disturbance input and K =8.  AP4.6  (a) The transfer function is Vo (s) 1 + RCs = . V (s) 2 + RCs  14  16  18  20  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  159  Advanced Problems  (b) The system sensitivity is defined as SCG =  ∂G/G . ∂C/C  Therefore, the sensitivity is determined to be SCG =  RCs = (2 + RCs)(1 + RCs) 1+  1 2 RCs  (c) Let V (s) = 1/s. Then  Vo (s) =    1+  1 RCs   .  1 + RCs 1 0.5 0.5RC = + . 2 + RCs s s RCs + 2  Taking the inverse Laplace transform yields vo (t) = 0.5(1 + e−2t/RC )u(t) where u(t) is the unit step function. A plot of vo (t) versus t/RC is shown in Figure AP4.6.  1 0.95 0.9 0.85  Vo  0.8 0.75 0.7 0.65 0.6 0.55 0.5  0  0.5  1  1.5  2  2.5 t / RC  FIGURE AP4.6 Step response.  3  3.5  4  4.5  5  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  160  CHAPTER 4  AP4.7  Feedback Control System Characteristics  (a) The transfer function from Td (s) to Y (s) is s Y (s) =− . Td (s) s(s + 1) + K (b) The transfer function from N (s) to Y (s) is Y (s) K = . N (s) s(s + 1) + K (c) Let Td (s) = A/s and N (s) = B/s. Then, ess = −yss = lim s s→0  A K B s − lim s = −B . s(s + 1) + K s s→0 s(s + 1) + K s  So, K has no effect on the steady-state errors. However, choosing K = 100 will minimize the effects of the disturbance Td (s) during the transient period. AP4.8  (a) The closed-loop transfer function is T (s) =  Kb . s + Kb + 2  (b) The sensitivity is determined to be SbT =  ∂T /T s+2 = . ∂b/b s + Kb + 2  (c) The transfer function from Td (s) to Y (s) is Y (s) b = . Td (s) s + Kb + 2 So, choose K as large as possible, to make Y (s)/Td (s) as "small" as possible. Thus, select K = 50 . This also minimizes SbT at low frequencies.  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  161  Design Problems  Design Problems The model of the traction drive, capstan roller, and linear slide was developed in CDP2.1: θ(s) Km = . Va (s) s [(Lm s + Rm )(JT s + bm ) + Kb Km ] The step response for the closed-loop system (with the tachometer not in the loop) and various values of the controller gain Ka is shown below.  % System parameters Ms=5.693; Mb=6.96; Jm=10.91e-03; r=31.75e-03; bm=0.268; Km=0.8379; Kb=0.838; Rm=1.36; Lm=3.6e-03; Lm=0; % Controller gain Ka=100; % Motor and slide model Jt=Jm+r^2*(Ms+Mb); num=[Km]; den=[Lm*Jt Rm*Jt+Lm*bm Kb*Km+Rm*bm 0]; sys=tf(num,den); %Closed-loop tf and step response sys_cl=feedback(Ka*sys,[1]); step(sys_cl) 1.5  Ka=2 Ka=5 Ka=10 Ka=100  1 Theta step response  CDP4.1  0.5  0  0  0.2  0.4  0.6 Time (sec)  0.8  1  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  162  CHAPTER 4  (a) The transfer function from the load disturbance to the output speed is −G(s) −s ω(s) = = 2 . Td (s) 1 + Gc G(s) s + 4s + K Thus, the effect on ω(s) (of a unit step disturbance) at steady-state is lim ω(t) = lim s  t→∞  s→0    −s 2 s + 4s + K    1 =0. s  We see that the load disturbance has no effect on the output at steadystate. (b) The system response for 10 ≤ K ≤ 25 is shown in Figure DP4.1. K=10,12,16,18,20,23,25 100.04 100.02 100  K=25  99.98 99.96  w(t)  DP4.1  Feedback Control System Characteristics  99.94  K=10  99.92 99.9 99.88 99.86 99.84 0  0.5  1  1.5  2  2.5  3  Time(sec)  FIGURE DP4.1 Speed control system response.  For example , if we select K = 16, then ωn = 4, ζ = response due to a unit step disturbance is −s ω(s) = 2 s + 4s + 16  1 s     =  1 2,  −1 . (s + 2)2 + 12  Hence, if we are originally at ω(t) = 100 for t < τ , we have √ 1 ω(t) = 100 − √ e−2t sin 12t 12  t≥τ .  and the  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  163  Design Problems  DP4.2  With θd = 0, we have θ(s) =  s G(s) T (s) = 3 Td (s) . KK1 d 2 s + 4s + 9s + KK1 1 + G(s) s  For Td = A/s, we have θ(s) =  A . s3 + 4s2 + 9s + KK1  The system response to a unit step disturbance for various values of KK1 are shown in Figure DP4.2. From the plot we see that when KK1 is small the response is slow but not oscillatory. On the other hand, when KK1 is large the response is fast but highly oscillatory. In fact, if KK1 > 35, the system is unstable. Thus, we might select KK1 = 10 as a reasonable trade-off between fast performance and stability. Unit step response for KK1=1,5,10,15,20,25  0.12 0.1  KK1=1  0.08 0.06  KK1=5  q  0.04 0.02 0 -0.02 KK1=25  -0.04 -0.06  0  1  2  3  4  5  6  7  8  time(sec)  FIGURE DP4.2 Aircraft roll angle control system response to a disturbance.  DP4.3  (a) The closed-loop transfer function is T (s) =  ω(s) K = 2 . ωd (s) s + 5s + KK1  9  10  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  164  CHAPTER 4  Feedback Control System Characteristics  Then, E(s) = (1 − T (s))ωd (s) =  s2 + 5s + K(K1 − 1) 1 . s2 + 5s + KK1 s  So, if 0.99 < K1 < 1.01 , then |ess | < 0.01 . (b) The transfer function from Td (s) to ω(s) is ω(s) =  s2  −s Td (s) . + 5s + KK1  So, with E(s) = −ω(s) and Td (s) = 2/s2 , we have lim sE(s) =  s→0  2 . KK1  Therefore, we select KK1 > 20 to obtain ess < 0.1. DP4.4  The steady-state error for a step input command is zero for any K1 . The transfer function from Td (s) to Y (s) is Y (s) G(s) 2 = = 3 . 2 Td (s) 1 + KG(s) s + 5s + 4s + 2K Thus, the output at steady-state due to a step disturbance Td (s) = A/s is lim sY (s) =  s→0  A . K  We want to maximize K to reduce the effect of the disturbance. As we will see in Chapter 6, we cannot select K too high or the system will become unstable. That is why the problem statement suggests a maximum gain of K = 10. For the design we choose K = 10 . DP4.5  The transfer function from V (s) to Vo (s) is Vo (s)/V (s) =  ks s+a  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  165  Design Problems  where k=  R2 + R3 R2  and a =  1 . R1 C  Computing the step response, we find that vo (t) = ke−at = 5e−100t . Solving for R1 , R2 , R3 and C yields R1 C = 0.01 DP4.6  and  R2 =4. R3  (a) The closed-loop transfer function is θ(s) =  s2  K/J θd (s) . + K/J  Since J > 0, the system is unstable when K < 0 and marginally stable when K > 0. (b) Since the system is marginally stable, the system response does not have a steady-state value—it oscillates indefinitely. (c) The closed-loop transfer function is θ(s) =  KD s + KP θd (s) . Js2 + KD s + KP  The system is stable for all KD > 0 and KP > 0, given that J > 0. (d) The tracking error E(s) = θd (s) − θ(s) is E(s) =  Js2 . Js2 + KD s + KP  Therefore, using the final value theorem we obtain the steady-state value lim sE(s) = lim s  s→0  DP4.7  s→0  Js2 1 · =0. 2 Js + KD s + KP s  (a) The closed-loop transfer function is Y (s) =  s2  s 1 Td (s) = 2 + Ks + 2K s + Ks + 2K  where the disturbance is a unit step Td (s) = 1/s. Considering the poles of the closed-loop system, we find that when K > 8 the system  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  166  CHAPTER 4  Feedback Control System Characteristics  has two real poles. In that case the disturbance step response is   1 y(t) = √ e−αt − e−βt , K 2 − 8K  where α=  K−  √  K 2 − 8K 2  and β =  K+  √  K 2 − 8K 2  Bounding the maximum y(t) yields the inequality |y(t)| = √  K2  1 e−αt − e−βt ≤ 0.05. − 8K  We know that e−αt − e−βt ≤ 1, for any α and β computed as shown above where K > 8. So, if we choose K such that √  1 ≤ 0.05. K 2 − 8K  we will guarantee that the maximum bound of 0.05 is not exceeded. Solving for K yields K > 24.4. For any K > 24.4 we know that the maximum value of the disturbance step response will be less than 0.05. When K = 24.4 the maximum unit step disturbance response is 0.035. Solving explicitly for K so that the maximum is 0.05 we find that K = 16.3 (this was found numerically since it is very difficult to obtain analytically). (b) Since the system is type 2, we know that the steady-state value of the disturbance step response is zero for a unit step disturbance. DP4.8  (a) The sensitivities are "  #  "  #  SτT1  ∂T /T −s2 (τ2 s + 1) τ1 = = T (s) ∂τ1 /τ1 K  SτT2  ∂T /T −s2 (τ1 s + 1) τ2 = = T (s) ∂τ2 /τ2 K  and  where we assume that K 6= 0.  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  167  Design Problems  (b) Computing the closed-loop transfer function yields Y (s) =  s (τ1 s + 1) Td (s) s (τ1 s + 1) (τ2 s + 1) + K  When Td (s) = 1/s, using the final value theorem we find that s (τ1 s + 1) =0 s→0 s (τ1 s + 1) (τ2 s + 1) + K  lim sY (s) = lim  s→0  as long as K 6= 0. We assume here that final value theorem applies (i.e., the system is stable, more on this in Chapter 6).  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  168  CHAPTER 4  Feedback Control System Characteristics  Computer Problems CP4.1  The step response and an m-file script which generates the step response is shown in Figure CP4.1. The closed-loop transfer function is T (s) =  s2  12 . + 2s + 22  The percent overshoot is P.O. = 50.2% and the steady-state error is ess = 0.45.  Step Response 0.9  0.8  System: sys_cl Peak amplitude: 0.82 Overshoot (%): 50.2 At time (sec): 0.67  0.7  Amplitude  0.6  num = [12]; den = [1 2 10]; sys = tf(num,den); sys_cl = feedback(sys,[1]) step(sys_cl)  0.5  0.4  0.3  0.2  0.1  0  0  1  2  3 Time (sec)  4  5  6  FIGURE CP4.1 Step response.  CP4.2  The transfer function is G(s) =  s2  4 . + 2s + 20  An m-file script which generates the step response is shown in Figure CP4.2. The step response is also shown in Figure CP4.2. The step response is generated using the step function. In the script, the transfer function numerator is represented by num and the denominator is represented by den. The steady-state value is yss = 0.2 and the desired value is 1.0. Therefore, the steady-state error is ess = 0.8 .  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  169  Computer Problems  Step Response 0.35  0.3  0.25  Amplitude  num=[4]; den=[1 2 20]; sys = tf(num,den); axis([0 6 0 1]); t=[0:0.01:6]; step(sys,t) y = step(sys,t); yss = y(length(t))  0.2  0.15  0.1  yss = 0.20  0.05  0  0  1  2  3 Time (sec)  4  5  6  FIGURE CP4.2 Step response.  CP4.3  The step responses and the m-file script which generates the step responses is shown in Figure CP4.3.  7  K=10 K=200 K=500  6  K=[10,200,500]; t=[0:0.01:7]; for i=1:3 num=5*K(i); den=[1 15 K(i)]; sys = tf(num,den) y(:,i)= step(sys,t); end plot(t,y(:,1),t,y(:,2),'--',t,y(:,3),':') legend('K=10','K=200','K=500',-1)  5  4  3  2  1  0  FIGURE CP4.3 Step responses for K = 10, 100, 500.  0  1  2  3  4  5  6  7  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  170  CHAPTER 4  CP4.4  Feedback Control System Characteristics  (a,b) The m-file and plots are shown in Figure CP4.4.  Step response  1.5  y(t) 1  ng=1;dg=[1 1.91 0];sysg=tf(ng,dg); K=10; syscl=feedback(K*sysg,1); figure(1) subplot(211) step(syscl) subplot(212) syst=feedback(sysg,K) step(syst)  0.5  0  0  1  2  3 Time (sec)  4  5  6  Disturbance response  0.2 0.15  y(t) 0.1 0.05 0  0  1  2  3 Time (sec)  4  5  6  FIGURE CP4.4 Step response and disturbance response.  (c) The estimated steady-state tracking error due to a unit step input is zero, and the estimated steady-state tracking error to a unit disturbance is 0.1. (d) The estimated maximum tracking error due to a unit step input is 0.4, and the estimated maximum tracking error to a unit disturbance is 0.14. The maximum occurs at approximately t = 1 s. CP4.5  The step response and the m-file script which generates the step response is shown in Figure CP4.5. The closed-loop transfer function is determined to be T (s) =  s2  10 . + 3.7s + 10  Using the m-file script, a trial-and-error search on k yields k = 3.7 . The percent overshoot P.O. = 10% and the steady-state value is 1, as expected.  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  171  Computer Problems 1.4  1.2  1  0.8 y(t)  k = 3.7; % Final value of k=3.7 numcg = [10]; dencg = [1 k 0]; sys_o = tf(numcg,dencg); sys_cl = feedback(sys_o,[1]) t = [0:0.1:5]; [y,t] = step(sys_cl,t); plot(t,y,[0 5],[1.1 1.1],'--'); grid xlabel('Time (sec)'); ylabel('y(t)');  0.6  0.4  Transfer function: 10 ---------------s^2 + 3.7 s + 10  0.2  0  0  0.5  1  1.5  2  2.5 Time (sec)  3  3.5  4  4.5  5  FIGURE CP4.5 Step response.  CP4.6  The closed-loop transfer function is T (s) =  K s−a+K  where K = 2. When a = 1 and R(s) = 1/s, the final value is lim sT (s)R(s) = lim T (s) =  s→0  s→0  K =2. K −a  The output is within 2% of the final value at around t = 4.6 seconds. The plot of the step responses for a = 1, 0.5, 2, 5 is shown in Figure CP4.6. The output is unstable for a > 2.  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  172  CHAPTER 4  Feedback Control System Characteristics  K=2; t=[0:0.1:5]; num=K*[1]; a=[1 0.5 2 5]; for i=1:4 den=[1 -a(i)]; sys = tf(num,den); sys_cl = feedback(sys,[1]); y(:,i)=step(sys_cl,t); end plot(t,y(:,1),t,y(:,2),':',t,y(:,3),'--',t,y(:,4),'-.') axis([0 5 0 5]); xlabel('Time (sec)'), ylabel('y(t)') title('a=1 (solid); a=0.5 (dotted); a=2 (dashed); a=5 (dashdot)') a=1 (solid); a=0.5 (dotted); a=2 (dashed); a=5 (dashdot) 5  4.5  4  3.5  y(t)  3  2.5  2  1.5  1  0.5  0  0  0.5  1  1.5  2  2.5 Time (sec)  3  3.5  4  4.5  5  FIGURE CP4.6 Step response for a=1, 0.5, 2, and 5.  CP4.7  The transfer function from the disturbance to the output is T (s) =  G(s) 1 = . 2 1 + K0 G(s) Js + bs + k + K0  The disturbance response is shown in Figure CP4.7. The compensated system response is significantly reduced from the uncompensated system response. The compensated system output is about 11 times less than the uncompensated system output. So, closed-loop feedback has the advantage of reducing the effect of unwanted disturbances on the output.  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  173  Computer Problems  J=1; k=5; c=0.9; num=[1/J]; den=[1 c/J k/J]; sys = tf(num,den); t=[0:0.1:10]; % yu=step(sys,t); % Part (a) K0=50; numk=[K0]; denk=[1]; sysk = tf(numk,denk); sys_cl = feedback(sys,sysk); yc=step(sys_cl,t); % Part (b) plot(t,yu,t,yc,'--') xlabel('Time (sec)'), ylabel('\theta') title('Uncompensated response (solid) & Compensated response (dashed)') Uncompensated response (solid) & Compensated response (dashed) 0.35  0.3  0.25  q  0.2  0.15  0.1  0.05  0  0  1  2  3  4  5 Time (sec)  6  7  8  9  10  FIGURE CP4.7 Disturbance responses for both the uncompensated and compensated systems.  CP4.8  The step responses for the proportional and PI controller are shown in Figure CP4.8. The steady-state tracking error for the proportional controller is ess = 0.33 . Increasing the complexity of the controller from a proportional controller to a proportional plus integral (PI) controller allows the closed-loop system to track the unit step response with zero steady-state error. The cost is controller complexity, which translates into higher costs ($).  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  174  CHAPTER 4  Feedback Control System Characteristics  numg=[10]; deng=[1 10]; sysg = tf(numg,deng); t=[0:0.001:0.5]; % Part (a) numc=[2]; denc=[1]; sysc = tf(numc,denc); sys_o = series(sysc,sysg); sys_cl = feedback(sys_o,[1]); yk=step(sys_cl,t); % Part (b) numc=[2 20]; denc=[1 0]; sysc = tf(numc,denc); sys_o = series(sysc,sysg); sys_cl = feedback(sys_o,[1]); yp=step(sys_cl,t); % plot(t,yk,t,yp,'--') xlabel('Time (sec)'),ylabel('y(t)') title('Proportional controller (solid) & PI controller (dashed)') Proportional controller (solid) & PI controller (dashed) 1  0.9  0.8  0.7  y(t)  0.6  0.5  0.4  0.3  0.2  0.1  0  0  0.05  0.1  0.15  0.2  0.25 Time (sec)  0.3  0.35  0.4  0.45  0.5  FIGURE CP4.8 Step response for proportional controller and PI controller.  CP4.9  (a) The closed-loop transfer function is T (s) =  G(s) 10s2 + 500s R(s) = 2 R(s) . 1 + G(s)H(s) s + 200s + 5000  The step response is shown in Figure CP4.9a.  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  175  Computer Problems  (b) The response of the system to the sinusoidal disturbance N (s) =  s2  100 + 100  is shown in Figure CP4.9b. (c) In the steady-state, the magnitude of the peak response is 0.095 and the frequency is 10 rad/sec (see Figure CP4.9b). % Part (a) ng=10*[1 0]; dg=[1 100]; sysg=tf(ng,dg); nh=[5]; dh=[1 50]; sysh=tf(nh,dh); sys=feedback(sysg,sysh) figure(1) step(sys)  >> Transfer function: 10 s^2 + 500 s -----------------s^2 + 200 s + 5000 Step Response  % Part (b) sysn=-feedback(sysg*sysh,1) syss=tf([100],[1 0 100]); % This is the sinusoidal input figure(2) t=[0:0.001:7]; step(syss*sysn,t)  10  Amplitude  8  6  4  2 Step Response 0.1  0  0.08  0  0.02  0.04  0.06  0.08  0.1  Time (sec) 0.06  (a)  Amplitude  0.04 0.02 0 −0.02 −0.04 −0.06 −0.08 −0.1  0  1  2  3 4 Time (sec)  5  6  7  (b)  FIGURE CP4.9 (a) Unit step response. (b) Response to sinusoidal noise input at ω = 10 rad/sec.  CP4.10  (a) The closed-loop transfer function is T (s) =  Gc (s)G(s) K(s + 1) R(s) = R(s) . 2 1 + G(s)Gc (s) (s + 15)(s + s + 6.5) + K(s + 1)  (b) The step responses are shown in Figure CP4.10a.  0.12  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  176  CHAPTER 4  Feedback Control System Characteristics  (c) The unit disturbance response of the system is shown in Figure CP4.10b. The steady-state value is 0.14. Step Response  0.7  K=5 K=10 K=50  0.6  0.2 0.18 0.16  System: syscl Final Value: 0.14  0.5  0.4  Amplitude  Step response  0.14  0.3  0.12 0.1 0.08  0.2  0.06 0.04  0.1 0.02  0  0  0  1  2  3 4 Time (s)  5  6  7  0  1  2  3 Time (sec)  (a)  4  5  6  (b)  FIGURE CP4.10 (a) Unit step responses for K = [5, 10, 50]. (b) Disturbance unit step response.  The m-file is shown in Figure CP4.11a and the step responses in Figure CP4.11b.  CP4.11  5  K=10 K=12 K=15  4 3  Step response  K=[10, 12, 15]; t=[0:0.1:20]; ng=[20]; dg=[1 4.5 64]; sysg=tf(ng,dg); nh=[1]; dh=[1 1]; sysh=tf(nh,dh); for i=1:length(K) sys=K(i)*sysg; syscl=feedback(sys,sysh) y(:,i)= step(syscl,t); end plot(t,y(:,1),t,y(:,2),'--',t,y(:,3),':') xlabel('Time (s)') ylabel('Step response') legend('K=10','K=12','K=15',-1)  2 1 0 −1 −2 −3  0  5  (a)  FIGURE CP4.11 (a) M-file script. (b) Unit step responses for K = [10, 12, 15].  10 Time (s)  (b)  15  20  7  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  C H A P T E R  5  The Performance of Feedback Control Systems  Exercises E5.1  For a zero steady-state error, when the input is a step we need one integration, or a type 1 system. A type 2 system is required for ess = 0 for a ramp input.  E5.2  (a) The closed-loop transfer function is T (s) =  Y (s) G(s) 240 240 = = = 2 . R(s) 1 + G(s) (s + 4)(s + 6) + 240 s + 2ζωn s + ωn2  The steady-state error is given by ess =  A , 1 + Kp  where R(s) = A/s and Kp = lim G(s) = s→0  240 = 10 . 24  Therefore, ess =  A . 11  (b) The closed-loop system is a second-order system with natural frequency √ ωn = 264 , 177  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  178  CHAPTER 5  The Performance of Feedback Control Systems  and damping ratio 10 ζ= √ = 0.31 . 2 264 The percent overshoot is thus computed to be √ 2 P.O. = 100e−πζ/ 1−ζ = 36% . E5.3  The closed-loop transfer function is G(s) K K Y (s) = = = 2 . I(s) 1 + G(s) s(s + 14) + K s + 14s + K Utilizing Table 5.6 in Dorf & Bishop, we find that the optimum coefficients are given by s2 + 1.4ωn s + ωn2 . We have s2 + 14s + K , so equating coefficients yields ωn = 10 and K = ωn2 = 100 . We can also compute the damping ratio as ζ=  14 = 0.7 . 2ωn  Then, using Figure 5.8 in Dorf & Bishop, we find that P.O. ≈ 5%. E5.4  (a) The closed-loop transfer function is T (s) =  G(s) 2(s + 8) = 2 . 1 + G(s) s + 6s + 16  (b) We can expand Y (s) in a partial fraction expansion as 2(s + 8) A 1 s+4 Y (s) = 2 =A − 2 (s + 6s + 16) s s s + 6s + 16     .  Taking the inverse Laplace transform (using the Laplace transform tables), we find √ y(t) = A[1 − 1.07e−3t sin( 7t + 1.21)] . (c) Using the closed-loop transfer function, we compute ζ = 0.75 and  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  179  Exercises  ωn = 4. Thus, 8 a = = 2.67 , ζωn 3 where a = 8. From Figure 5.13(a) in Dorf & Bishop, we find (approximately) that P.O. = 4% . (d) This is a type 1 system, thus the steady-state error is zero and y(t) → A as t → ∞. E5.5  The closed-loop transfer function is Y (s) K = 2 . R(s) s + 4s + K Utilizing Table 5.6 in Dorf & Bishop, we find that the optimum coefficients are given by s2 + 1.4ωn s + ωn2 . We have s2 + 4s + K , so equating coefficients yields ωn = 2.86 and K = ωn2 = 8.16 . We can also compute the damping ratio as ζ=  4 = 0.7 . 2ωn  Then, using Figure 5.8 in Dorf & Bishop, we find that P.O. ≈ 5%. E5.6  (a) The closed-loop transfer function is T (s) =  Y (s) G(s) 100 = = 2 , R(s) 1 + GH(s) s + 100Ks + 100  where H(s) = 1 + Ks and G(s) = 100/s2 . The steady-state error is computed as follows: ess = lim s[R(s) − Y (s)] = lim s[1 − T (s)] s→0  s→0  "  = lim 1 − s→0  1+  100 s2 100 (1 + s2  A s2  #  A = KA . Ks) s  (b) From the closed-loop transfer function, T (s), we determine that ωn =  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  180  CHAPTER 5  The Performance of Feedback Control Systems  10 and ζ=  100K = 5K . 2(10)  We want to choose K so that the system is critically damped, or ζ = 1.0. Thus, K=  1 = 0.20 . 5  The closed-loop system has no zeros and the poles are at p  s1,2 = −50K ± 10 25K 2 − 1 . The percent overshoot to a step input is P.O. = 100e  √−5πK  1−25K 2  for  0 < K < 0.2  and P.O. = 0 for K > 0.2. E5.7  The closed-loop transfer function is Y (s) KG(s) K(s + 2) K(s + 2) = = = 2 . R(s) 1 + KG(s) s(s + 1) + K(s + 2) s + s(K + 1) + 2K √ √ Therefore, ωn = 2K and ζ = 2K+1 . So, 2K T (s) =  a 4 = . ζωn K +1 From Figure 5.13a in Dorf & Bishop, we determine that a ≈ 1.5 ζωn when ζ = 0.707. Thus, solving for K yields 4 = 1.5 K +1 or K = 1.67. E5.8  The pole-zero map is shown in Figure E5.8. Since the dominant poles are at s = −2 ± 2.45i we have a damping ratio ζ = 0.63. We estimate the percent overshoot to be √ 2 P.O. = 100e−πζ/ 1−ζ = 7.69% The step response is shown in Figure E5.8b. The actual overshoot is 8%.  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  181  Exercises  Pole−Zero Map 2.5 2 1.5  Imaginary Axis  1 0.5 0 −0.5 −1 −1.5 −2 −2.5 −25  −20  −15  −10  −5  0  Real Axis  Step Response 1.4  System: sys Time (sec): 1.28 Amplitude: 1.08  1.2  Amplitude  1  0.8  0.6  0.4  0.2  0  0  0.5  1  1.5 Time (sec)  2  FIGURE E5.8 (a) Pole-zero map. (b) Unit step response.  E5.9  (a) The closed-loop transfer function is T (s) =  s2 +  √  K . 2Ks + K  2.5  3  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  182  CHAPTER 5  The Performance of Feedback Control Systems  √ The √ damping ratio is ζ = 2/2 and the natural frequency is ωn = K. Therefore, we compute the percent overshoot to be √ 2 P.O. = 100e−πζ/ 1−ζ = 4.3% for ζ = 0.707. The settling time is estimated via Ts =  8 4 =√ . ζωn 2K  (b) The settling time is less than 1 second whenever K > 32. E5.10  The second-order closed-loop transfer function is given by T (s) =  ωn2 . s2 + 2ζωn s + ωn2  From the percent overshoot specification, we determine that P.O. ≤ 5% implies ζ ≥ 0.69. From the settling time specification, we find that Ts < 4 implies ωn ζ > 1. p And finally, from the peak time specification we have Tp < 1 implies ωn 1 − ζ 2 > π. The constraints imposed on ζ and ωn by the performance specifications define the permissible area for the poles of T (s), as shown in Figure E5.10. Im(s) wn  z = 0.69  1-z 2 = P  desired region  46o for poles  wn  Re(s)  1-z 2 = - P  z wn = -1  FIGURE E5.10 Permissible area for poles of T (s).  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  183  Exercises  E5.11  The system is a type 1. The error constants are Kp = ∞  and Kv = 1.0 .  Therefore, the steady-state error to a step input is 0; the steady-state error to a ramp input is 1.0A0 , where A0 is the magnitude (slope) of the ramp input. (a) The tracking error is given by E(s) =  (s + 9)(s + 2)(s + 4) R(s) = R(s) . 1 + Gc G(s) (s + 9)(s + 2)(s + 4) + K(s + 6)  The steady-state tracking error (with R(s) = 1/s) is lim sE(s) =  s→0  72 . 72 + 6K  We require ess < 0.05, so solving for K yields K > 228. (b) The tracking error due to the disturbance is E(s) =  −G(s) −(s + 9)(s + 6) Td (s) = Td (s) . 1 + Gc G(s) (s + 9)(s + 2)(s + 4) + K(s + 6)  The tracking error is shown in Figure E5.12.  0 -0.01 -0.02 -0.03  Amplitude  E5.12  -0.04 -0.05 -0.06 -0.07 -0.08  0  0.1  0.2  0.3  0.4  0.5 Time (secs)  FIGURE E5.12 Tracking error due a step disturbance.  0.6  0.7  0.8  0.9  1  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  184  CHAPTER 5  E5.13  The Performance of Feedback Control Systems  The system is a type 0. The error constants are Kp = 0.4 and Kv = 0. The steady-state error to a ramp input is ∞. The steady-state error to a step input is ess =  E5.14  1 = 0.71. 1 + Kp  (a) The tracking error is given by E(s) = [1 − T (s)] R(s) . The steady-state tracking error (with R(s) = 1/s) is ess = lim s [1 − T (s)] R(s) = lim [1 − T (s)] = 1 − T (0) . s→0  s→0  The closed-loop transfer function is T (s) =  K(s + 0.1) , s(s + 0.1)(s + 2) + K(s + 3)  and T (0) = 0.033. Therefore, ess = 1 − T (0) = 0.967.  (b) Use Gp (s) = 30. Then,  ess = lim s [1 − T (s)Gp (s)] R(s) = 1−lim T (s)Gp (s) = 1−30 T (0) = 0 . s→0  The plot of y(t) is shown in Figure E5.15. 1.4  1.2  Response using only dominate poles 1  0.8 y(t)  E5.15  s→0  0.6  Actual response  0.4  0.2  0  0  0.2  0.4  0.6  0.8  1 Time (sec)  1.2  1.4  1.6  1.8  FIGURE E5.15 Plot of y(t) with T (s) (solid line) and approximate Ta (s) (dashed line).  2  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  185  Exercises  Using the dominant poles, the approximate closed-loop transfer function is Ta (s) =  s2  50 . + 10s + 50  The actual transfer function is T (s) = E5.16  500 . (s + 10)(s2 + 10s + 50)  The partial fraction expansion is y(t) = −  10(z − 1) −t 10(z − 8) −8t e + e + 1.25 . 7z 56z  The plot of y(t) for z = 2, 4, 6 is shown in Figure E5.16. z=2 (solid) & z=4 (dashed) & z=6 (dotted) 1.4  1.2  1  y(t)  0.8  0.6  0.4  0.2  0  0  1  2  3 Time (sec)  4  5  6  FIGURE E5.16 Plot of y(t) for z=2, 4, 6.  E5.17  The desired pole locations for the 5 different cases are shown in Figure E5.17.  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  186  CHAPTER 5  The Performance of Feedback Control Systems  Im(s) desired region for poles  wn = 10  37  o  o 53  Re(s)  (a) 0.6 < z < 0.8 and wn <10 desired region for poles  Im(s)  wn = 10  45  o  o  Re(s)  60  (b) 0.5 < z < 0.707 and  wn > 1 0 Im(s)  desired region for poles  wn = 10  o  Re(s)  60  wn = 5  (c) 0.5 < z and 5 < wn <10  FIGURE E5.17 Desired pole locations.  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  187  Exercises Im(s)  desired region for poles  wn = 10 o  Re(s)  45  wn = 5  (d) 0.707 >  z and 5 < wn <10 Im(s)  o  wn = 6  Re(s)  53 desired region for poles  (e) 0.6 < z and wn < 6  FIGURE E5.17 CONTINUED: Desired pole locations.  E5.18  The output is given by Y (s) = T (s)R(s) = K  G(s) R(s) . 1 + G(s)  When K = 1, the steady-state error is ess = 0.2 which implies that lim sY (s) = 0.8 .  s→0  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  188  CHAPTER 5  The Performance of Feedback Control Systems  Since we want ess = 0, it follows that lim sY (s) = 1 ,  s→0  or 0.8K = 1 . Therefore, K = 1.25. (a) The characteristic equation is s2 = 2ζωn s + ωn2 = s2 + 3.17s + 7 = 0 , from which it follows that √ ωn = 7 = 2.65,  ζ=  3.17 = 0.6 . 2ωn  Therefore, we compute the percent overshoot and the estimated settling time to be √ 4 2 P.O. = 100e−πζ/ 1−ζ = 9.53% and Ts = = 2.5 s . ζωn (b) The unit step response is shown in Figure E5.19.  Step Response 1.4 System: sys Peak amplitude: 1.1 Overshoot (%): 9.53 At time (sec): 1.47  1.2  System: sys Settling Time (sec): 2.25  1  Amplitude  E5.19  0.8  0.6  0.4  0.2  0  0  FIGURE E5.19 Unit step response.  0.5  1  1.5 Time (sec)  2  2.5  3  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  189  Exercises  E5.20  (a) The closed-loop transfer function is T (s) =  s2  +  √  K . 2Ks + K  The damping ratio is √  2 2 √ and the natural frequency is ωn = K. Therefore, we compute the percent overshoot to be √ 2 P.O. = 100e−πζ/ 1−ζ = 4.3% ζ=  for ζ = 0.707. The settling time is estimated via Ts =  4 8 =√ . ζωn 2K  (b) The settling time is less than 1 second whenever K > 32.  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  190  CHAPTER 5  The Performance of Feedback Control Systems  Problems P5.1  (a) The system error is E(s) =  1 1+  Ka Km sτm +1  R(s)  where R(s) =  25o /sec . s  So, lim e(t) = lim sE(s) =  t→0  s→0  25 . 1 + Ka Km  (b) If we desire ess ≤ 1o /sec, then 25o /s ≤ 1o /sec , 1 + Ka Km and solving for Ka Km yields Ka Km ≥ 24 . (c) The closed-loop transfer function is T (s) =  Vb (s) Ka Km = . Vc (s) sτm + 1 + Ka Km  The step response of the system (i.e. vc (t) = A) is −(Ka Km +1) AKa Km t τm vb (t) = 1−e 1 + Ka Km      .  So, at settling time, we have 1−e  −(1+Ka Km ) t τm  ≥ 0.98 ,  where τm = 0.4. Setting t = 0.03 and solving for Ka Km yields Ka Km ≥ 52 . P5.2  (a) The settling time specification Ts =  4 < 0.6 ζωn  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  191  Problems  is used to determine that ζωn > 6.67. The P.O. < 20% requirement is used to determine ζ < 0.45  which implies  θ < 63o  and the P.O. > 10% requirement is used to determine ζ > 0.60  which implies  θ > 53o ,  since cos θ = ζ. The desired region for the poles is shown in Figure P5.2.  Im(s) desired region for poles  53  o 63  o  Re(s)  s = -6.67  FIGURE P5.2 Desired region for pole placement.  (b) The third root should be at least 10 times farther in the left halfplane, so |r3 | ≥ 10|ζωn | = 66.7 . (c) We select the third pole such that r3 = −66.7. Then, with ζ = 0.45 and ζωn = 6.67, we determine that ωn = 14.8. So, the closed-loop transfer function is T (s) =  66.7(219.7) , (s + 66.7)(s2 + 13.3s + 219.7)  where the gain K = (66.7)(219.7) is chosen so that the steady-state  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  192  CHAPTER 5  The Performance of Feedback Control Systems  tracking error due to a step input is zero. Then, T (s) =  G(s) , 1 + G(s)  G(s) =  T (s) . 1 − T (s)  or  P5.3  Given the input R(s) =  1 , s3  we compute the steady-state error as ess  1 = lim s s→0 1 + G(s)     1 1 = lim 2 3 s→0 s s G(s)       = lim  s→0  Since we require that ess ≤ 0.5 cm, we determine  1 s2    K s2     =  1 . K  K≥2. P5.4  (a) The closed-loop transfer function is T (s) =  G(s) K ωn2 = 2 = 2 . 1 + G(s) s + 2s + K s + 2ζωn s + ωn2  Thus, ωn =  √  K  and  √ ζ = 1/ωn = 1/ K .  √ Our percent overshoot requirement√of 5% implies that ζ = 1/ 2 , which in turn implies that ωn = 2. However, the corresponding time to peak would be 4.4 Tp = √ = 3.15 . 2 Our desired Tp = 1.1—we cannot meet both specification simultaneously. (b) Let Tp = 1.1∆ and P.O. = 0.05∆, where ∆ is the relaxation factor to be determined. We have that K = ωn2 and ζωn = 1, so 1 ζ=√ . K  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  193  Problems  Thus, √ √ 2 P.O. = e−πζ/ 1−ζ = e−π/ K−1 . Also, Tp = √  π = 1.1∆ . K −1  Therefore, from the proceeding two equations we determine that P.O. = 0.05∆ = e−1.1∆ . Solving for ∆ yields f (∆) = ln ∆ + ln(0.05) + 1.1∆ = 0 . The plot of f (∆) versus ∆ is shown in Figure P5.4. From the plot we  2 1 0  *  D=2.07  -1  f(D)  -2 -3 -4 -5 -6 -7 -8  0  0.5  1  1.5  2  D FIGURE P5.4 Solving for the zeros of f.  see that ∆ = 2.07 results in f (∆) = 0. Thus, P.O. = 0.05∆ = 10% Tp = 1.1∆ = 2.3 sec.  2.5  3  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  194  CHAPTER 5  The Performance of Feedback Control Systems  So, we can meet the specifications if they are relaxed by a factor of about 2 (i.e. ∆ = 2.07). P5.5  (a) The closed-loop transfer function is T (s) =  s2  K1 K2 (s + 1) . + K1 K2 s + K1 K2  A percent overshoot less than 5% implies ζ ≥ 0.69. So, choose ζ = 0.69. Then set 2ζωn = K1 K2 and ωn2 = K1 K2 . Then 2(0.69)ωn = ωn2 ; and solving for ωn yields ωn = 1.38 . Therefore K1 K2 = ωn2 = 1.9. When K1 K2 ≥ 1.9 it follows that ζ ≥ 0.69. (b) We have a type 2 system, so the steady-state tracking error to both a step and ramp input is zero. (c) For a step input, the optimum ITAE characteristic equation is s2 + 1.4ωn s + ωn2 = 0 . For a ramp input, the optimum ITAE characteristic equation is s2 + 3.2ωn s + ωn2 = 0 . Thus, K1 K2 = ωn2 = 3.2ωn . So, ωn = 3.2 and K1 K2 = 10.24. P5.6  We have a ramp input, r(t) = t. So Kv = lim sG(s) = lim s s→0  s→0    75 75(s + 1) = = 0.6 , s(s + 5)(s + 25) 125   and ess = P5.7  |R| 1 = = 1.67 . Kv 0.6  (a) The closed-loop transfer function is T (s) =  Is2  K1 K2 . + K1 K2 K3 s + K1 K2  The steady-state tracking error for a ramp input is ess = lim sE(s) = lim s(1 − T (s))R(s) = lim s(1 − T (s)) s→0  s→0  s→0  1 s2  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  195  Problems  = lim  s→0  Is + K1 K2 K3 = K3 . Is2 + K1 K2 K3 s + K1 K2  But we desire ess = 0.01 m, so K3 = 0.01. (b) For P.O. = 10%, we have ζ = 0.6. Also, 2ζωn =  0.01K1 K2 25  and ωn2 =  K1 K2 . 25  Thus, solving for K1 K2 yields K1 K2 = 36 × 104 . P5.8  (a) The closed-loop transfer function is T (s) =  P (s) G(s)/s 20 = = . R(s) 1 + G(s)H(s)/s s(s + 40)  Therefore, the closed-loop system time constant is τ = 1/40 sec. (b) The transfer function from Td (s) to the output P (s) is P (s) −G(s) −20 = = . Td (s) 1 + G(s)H(s)/s s + 40 The response to a unit step disturbance is 1 p(t) = − (1 − e−40t ) . 2 At settling time, p(t) = 0.98pss = −0.49. Thus, solving for t(= Ts ) we determine that Ts = 0.098 sec. P5.9  We need to track at the rate ω=  v 16000 = = 1.78 × 10−3 radians/sec . r 2500  The desired steady-state tracking error is ess ≤  1 degree = 0.1754 × 10−2 rad . 10  Therefore, with ess =  |ω| , Kv  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  196  CHAPTER 5  The Performance of Feedback Control Systems  we compute Kv as Kv =  1.78 × 10−3 = 1.02 . 0.1754 × 10−2  This assumes that the system is type 1. (a) The armature controlled DC motorblock diagram is shown in Figure P5.10.  P5.10  amplifier  R(s )  +  Km  K  1 J s+b  R a+ L as  -  w(s)  Kb back emf  FIGURE P5.10 Armature controlled DC motor block diagram.  (b) The closed-loop transfer function is T (s) =  ω(s) KG(s) = , R(s) 1 + KKb G(s)  where G(s) =  Km . (Ra + La s)(Js + b)  Thus, T (s) =  K , s2 + 2s + 1 + K  where Ra = La = J = b = Kb = Km = 1. The steady-state tracking error is ess  A = lim s(R(s) − Y (s)) = lim s (1 − T (s)) s→0 s→0 s   K A = A(1 − T (0)) = 1 − = . 1+K 1+K     (c) For a percent overshoot of 15%, we determine that ζ = 0.5. √ From our characteristic polynomial we have 2ζωn = 2 and ωn = 1 + K. Solving for ωn yields ωn = 2, thus K = 3.  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  197  Problems  P5.11  (a) The closed-loop transfer function is T (s) =  K Y (s) = . R(s) s+K  To include the initial condition, we obtain the differential equation: ẏ(t) + Ky(t) = Kr(t) . Taking the Laplace transform yields: sY (s) − y(to ) + KY (s) = K    A s    ,  where y(to ) = Q. Computing the inverse Laplace transform, L−1 {Y (s)} yields y(t) = A(1 − e−Kt ) + Qe−Kt . Also, the tracking error is given by e(t) = A − y(t) = e−Kt (A − Q) . Thus, the performance index, I is determined to be (for K > 0) I=  Z  0  ∞  2 −2Kt  (A − Q) e  2  dt = (A − Q) =  Q)2  (A − 2K    1 −2K    e−2Kt  ∞ 0  .  (b) The minimum I is obtained when K = ∞, which is not practical. (c) Set K at the maximum value allowable such that the process does not saturate. For example, if K = 50, then I= P5.12  (A − Q)2 . 100  The optimum ITAE transfer function for a ramp input is T (s) =  3.25ωn2 s + ωn3 . s3 + 1.75ωn s + 3.25ωn2 s + ωn3  The steady-state tracking error, ess = 0, for a ramp input. The step response is shown in Figure P5.12 for ωn = 10. The percent overshoot is P.O. = 39%, and the settling time is Ts = 0.72 s .  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  198  CHAPTER 5  The Performance of Feedback Control Systems  1.4 PO=39% 1.2 Ts=0.72s 1  y(t)  0.8  0.6  0.4  0.2  0  0  0.2  0.4  0.6  0.8  1 time (sec)  1.2  1.4  1.6  1.8  2  FIGURE P5.12 Step input system response.  The step responses for the actual system and the approximate system are shown in Figure P5.13. It can be seen that the responses are nearly identical. 1  0.9  0.8  0.7  0.6 y(t)  P5.13  0.5  0.4  0.3  0.2  0.1  0  0  0.5  1  1.5  2  2.5 Time (sec)  3  3.5  4  4.5  5  FIGURE P5.13 Closed-loop system step response: Actual T(s) (solid line) and second-order approximation (dashed line).  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  199  Problems  P5.14  Consider L(s) =  2(c1 s + 1) . (s + 1)(s + 2)  After cancellation of like factors, we compute H(s)/L(s), H(s) s3 + 7s2 + 24s + 24 = . L(s) (s + 3)(s + 4)2(c1 s + 1) Therefore, M (s) = s3 + 7s2 + 24s + 24 , and ∆(s) = 2[c1 s3 + (7c1 + 1)s2 + (12c1 + 7)s + 12] . Then, following the procedure outlined in Section 5.10, we have M o (0) = 24,  M 1 (0) = 24,  M 2 (0) = 14,  M 3 (0) = 6,  and  ∆0 (0) = 24, ∆1 (0) = (12c1 + 7)2, ∆2 (0) = 2(2 · (7c1 + 1)), ∆3 (0) = 12c1 . For q = 1: M2 = 240, and ∆2 = 4[144c21 + 25] . Then, equating ∆2 and M2 , we find c1 , c1 = 0.493 . So, L(s) = P5.15  0.986s + 2 0.986(s + 2.028) 2(0.493s + 1) = 2 = . (s + 1)(s + 2) s + 3s + 2 (s + 1)(s + 2)  The closed-loop transfer function is T (s) =  (s +  4)(s2  K(s + 1) . + s + 10) + K(s + 1)  The percent overshoot as function of the gain, K, is shown in Figure P5.15. It can be seen that the percent overshoot decreases as the gain increases approaching a minimum around 85%. The larger the gain, the smaller the percent overshoot. For a gain K ≈ 250, we have essentially minimized the percent overshoot.  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  200  CHAPTER 5  The Performance of Feedback Control Systems  160  150  Percent Overshoot (%)  140  130  120  110  100  90  80  0  50  100  150  200  250  300  350  K  FIGURE P5.15 Percent overshoot versus the gain, K.  P5.16  The open-loop transfer function is G(s) =  10 . (s + 1)(50Cs + 1)  Define τ = 50C. Then, the closed-loop transfer function is Vo (s) 10 10/τ   = = τ +1 2 Vin (s) (s + 1)(τ s + 1) + 10 s + τ s+  11 τ  .  With ωn2 =  11 τ  τ +1 1 and ζ = √ = , 2τ ωn 2  we can solve for τ , yielding τ 2 − 20τ + 1 = 0 . Therefore, τ = 19.95 and 0.05. For each value of τ we determine C as follows: τ = 19.95 = 50C, implies C = 0.399F , and τ = 0.05 = 50C, implies C = 1mF .  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  201  Problems  P5.17  (a) The closed-loop transfer function is T (s) =  Y (s) 12K = 2 . R(s) s + 12s + 12K  The percent overshoot specification P.O. ≤ 10% implies ζ ≥ 0.59. From the characteristic equation we find that ωn2 = 12K Solving for K yields √ 2(0.59) 12K = 12  and ζωn = 6 .  which implies that K = 8.6 .  So, any gain in the interval 0 < K < 8.6 is valid. The settling time is Ts = 4/ζωn = 4/6 seconds and satisfies the requirement. Notice that Ts is not a function of K. (b) The sensitivity is T SK (s) =  1 s(s + 12) = 2 1 + G(s) s + 12s + 120  when K = 10. (c) The sensitivity at DC (s = 0) is T SK (0) = 0 .  (d) In this case, s = j2π · 1 beat/sec = j2π. So, the sensitivity at s = 2πj is T |SK (j2π)| =  P5.18  85.1084 = 0.77 . 110.31  We select L(s) as L(s) =  1 , αs + 1  then H(s) 6αs + 6 = 3 . L(s) s + 6s2 + 11s + 6 Therefore, M (s) = 6αs + 6 ,  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  202  CHAPTER 5  The Performance of Feedback Control Systems  and M o (0) = 6, M 1 (0) = 6α, M 2 (0) = 0. Also, ∆(s) = s3 + 6s2 + 11s + 6 , and ∆o (0) = 6 , ∆1 (0) = 11 , ∆2 (0) = 12. So, computing M2 and ∆2 yields M2 = 36α2 ,  and  ∆2 = 49 . Finally, equating M2 = ∆2 yields 36α2 = 49 , or α = 1.167 . Thus, L(s) = P5.19  1 0.857 = . 1.167s + 1 s + 0.857  (a) The closed-loop transfer function is T (s) =  s3  + 6s2  8 . + 12s + 8  (b) The second-order approximation is L(s) =  d2  s2  1 , + d1 s + 1  where d1 and d2 are to be determined using the methods of Section 5.10 in Dorf & Bishop. Given M (s) = 8d2 s2 + 8d1 s + 8 ∆(s) = s3 + 6s2 + 12s + 8 we determine that M2 M4 ∆2 ∆4  = −128d2 + 64d21 = 64d22 = 48 = 18 .  Equating M2 = ∆2 and M4 = ∆4 , and solving for d1 and d2 yields d1 = 1.35  and d2 = 0.53 .  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  203  Problems 1  0.9  0.8  0.7  y(t)  0.6  0.5  0.4  0.3  0.2  0.1  0  0  0.5  1  1.5  2  2.5 Time (sec)  3  3.5  4  4.5  5  FIGURE P5.19 Closed-loop system step response: Actual T (s) (solid line) and second-order approximation (dashed line).  Thus, the second-order approximation is L(s) =  0.53s2  1 . + 1.35s + 1  (c) The plot of the step response for the actual system and the approximate system is shown in Figure P5.19. P5.20  The steady-state error is (s + 5)(s + 11) + K(1 − K1 ) 55 + K(1 − K1 ) = . s→0 (s + 5)(s + 11) + K 55 + K  ess = lim  To achieve a zero steady-state tracking error, select K1 as follows: K1 = 1 + P5.21  55 . K  The closed-loop transfer function is T (s) =  s+a . s2 + (2k + a)s + 2ak + 1  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  204  CHAPTER 5  The Performance of Feedback Control Systems  (a) If R(s) = 1/s, we have the tracking error E(s) = R(s) − Y (s) = [1 − T (s)]R(s) or E(s) =  s2 + (2k + a − 1)s + 2ak + 1 − a 1 · . s2 + (2k + a)s + 2ak + 1 s  From the final value theorem we obtain ess = lim sE(s) = s→0  2ak + 1 − a . 2ak + 1  Selecting k = (a − 1)/(2a) leads to a zero steady-state error due to a unit step input. (b) To meet the percent overshoot specification we desire ζ ≥ 0.69. From T (s) we find ωn2 = 2ak + 1 and 2ζωn = 2k + a. Therefore, solving for a and k yields a = 1.5978  and k = 0.1871  when we select ζ = 0.78. We select ζ > 0.69 to account for the zero in the closed-loop transfer function which will impact the percent overshoot. With a and k, as chosen, we have T (s) =  s2  s + 1.598 + 1.972s + 1.598  and the step response yields P.O. ≈ 4%. P5.22  The closed-loop transfer function is T (s) =  2(2s + τ ) . (s + 0.2K)(2s + τ ) + 4  (a) If R(s) = 1/s, we have the unit step response Y (s) =  2(2s + τ ) 1 . (s + 0.2K)(2s + τ ) + 4 s  From the final value theorem we obtain yss = lim sY (s) = s→0  2τ . 0.2Kτ + 4  Selecting K = 10− 20/τ leads to yss = 1 and a zero steady-state error due to a unit step input.  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  205  Problems  (b) The characteristic equation is (s + 0.2K)(2s + τ ) + 4 = 2s2 + (0.4K + τ )s + 0.2Kτ + 4 = 0 . So, with K = 10 − 20/τ , the natural frequency and damping ratio are: ωn =  √  τ  and ζ =  τ 2 + 4τ − 8 . 4τ 3/2  The settling time and percent overshoot are found using the standard design formulas √ π −ζπ 1−ζ 2 p Tp = and P.O. = 100e ωn 1 − ζ 2  with ωn and ζ given above (as a function of τ ). Since the closed-loop system has a zero at s = −τ /2, the formulas for Tp and P.O. will only be approximate. Also, note that for the closed-loop system poles to be in the left half-plane (that √ is, all the poles have negative real parts), we require that τ > 2 3 − 2 ≈ 1.4642. As seen √ in the next chapter, this is the condition for stability. Having τ > 2 3 − 2 insures that the damping ratio ζ is positive.  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  206  CHAPTER 5  The Performance of Feedback Control Systems  Advanced Problems (a) The steady-state error is ess = lim s(1 − T (s))R(s) = 1 − T (0) = 1 − s→0  108(3) =0. 9(36)  (b) Assume the complex poles are dominant. Then, we compute a = 0.75 , ζωn since a = 3, ζ = 0.67 and ωn = 6. Using Figure 5.13 in Dorf & Bishop, we estimate the settling time and percent overshoot to be P.O. = 45%  and Ts =  4 = 1 second . ζωn  (c) The step response is shown in Figure AP5.1. The actual settling time and percent overshoot are P.O. = 34.4%  and Ts = 1.18 second .  Step Response 1.4  1.2  1  Amplitude  AP5.1  0.8  0.6  0.4  0.2  0  0  0.2  0.4  0.6  FIGURE AP5.1 Closed-loop system step response.  0.8  1 Time (sec)  1.2  1.4  1.6  1.8  2  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  207  Advanced Problems  The closed-loop transfer function is T (s) =  s3  +  28s2  5440(τz s + 1) . + (432 + 5440τz )s + 5440  The closed-loop step responses are shown in Figure AP5.2. The performance results are summarized in Table AP5.2. tau=0 (solid) & tau=0.05 (dashed) & tau=0.1 (dotted) & tau=0.5 (dot-dash) 1.4  1.2  1  0.8 y(t)  AP5.2  0.6  0.4  0.2  0  0  0.2  0.4  0.6  0.8 Time (sec)  1  1.2  1.4  1.6  FIGURE AP5.2 Closed-loop system step responses.  τz  Tr  Ts  P.O.  0  0.16  0.89  32.7%  p = −20, −4 ± 16j  0.05  0.14  0.39  4.5%  p = −10.4, −8.77 ± 21.06j  0.1  0.10  0.49  0%  p = −6.5, −10.74 ± 26.84j  0.5  0.04  1.05  29.2%  p = −1.75, −13.12 ± 54.16j  TABLE AP5.2  closed-loop poles  Performance summary.  As τz increases from 0 to 0.1, the P.O. decreases and the response is faster  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  208  CHAPTER 5  The Performance of Feedback Control Systems  and more stable. However, as τz is increased beyond 0.1, the P.O. and Ts increase, although Tr continues to decrease. The closed-loop transfer function is T (s) =  τp  s3  1 . + (1 + 2τp )s2 + 2s + 1  The closed-loop step responses for τp = 0, 0.5, 2, 5 are shown in Figure AP5.3. The performance results are summarized in Table AP5.3. tau=5 (solid) & tau=2 (dotted) & tau=0.5 (dashed) & tau=0 (dot-dash) 1.5  1  y(t)  AP5.3  0.5  0  0  5  10  15  20  25 Time (sec)  30  35  40  45  FIGURE AP5.3 Closed-loop system step responses.  τp  Tr  Ts  P.O.  0  4  5.8  0%  0.5  3.6  7.4  4.75%  p = −2.84, −0.58 ± 0.6j  2  4.6  22.4  27.7%  p = −2.14, −0.18 ± 0.45j  5  6  45.8  46%  p = −2.05, −0.07 ± 0.3j  TABLE AP5.3  Performance summary.  closed-loop poles p = −1, −1  50  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  209  Advanced Problems  As τp increases, the P.O., Tr and Ts also increase; adding the pole makes the system less stable with more overshoot. The system transfer function is Y (s) =  15 15K R(s) + Td (s) . (s + 5)(s + 7) + 15K (s + 5)(s + 7) + 15K  When considering the input response, we set Td (s) = 0, and similarly, when considering the disturbance response, we set R(s) = 0. The closedloop step input and disturbance responses for K = 1, 10, 100 are shown in Figure AP5.4. The performance results are summarized in Table AP5.4. Unit step input response  Unit step distrubance response  1.6  0.35  1.4  0.3  1.2 0.25 1  y(t)  0.2 y(t)  AP5.4  0.8  0.15 0.6 0.1 0.4  0.05  0.2  0  0  0.2  0.4 0.6 Time (sec)  0.8  1  0  0  0.2  0.4 0.6 Time (sec)  0.8  1  FIGURE AP5.4 Closed-loop system input and disturbance responses (K =1: solid line, K =10: dotted line, and K =100:dashed line).  TABLE AP5.4  K  ess  Ts  P.O.  |y/d|max  1  0.7  0.45  0%  0.3  10  0.19  0.6  17.3%  0.1  100  0.023  0.59  60.0%  0.01  Performance summary.  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  210  CHAPTER 5  The Performance of Feedback Control Systems  The best value of the gain is K = 10, which is compromise between (i) percent overshoot, and (ii) disturbance rejection and tracking error. The system transfer function is 50(s + α)(s + 2) R(s) s(s + 3)(s + 4) + 50(s + α)(s + 2) 50s(s + 2) Td (s) . + s(s + 3)(s + 4) + 50(s + α)(s + 2)  Y (s) =  Disturbance response: alpha=0 (solid) & alpha=10 (dashed) & alpha=100 (dotted) 10 9 8 7 6 5 y(t)  AP5.5  4 3 2 1 0 -1  0  0.05  0.1  0.15  0.2  0.25  Time (sec)  FIGURE AP5.5 Closed-loop system disturbance response.  When considering the input response, we set Td (s) = 0, and similarly, when considering the disturbance response, we set R(s) = 0. The steadystate tracking error is ess = lim s(1 − T (s))R(s) = lim 1 − s→0  s→0  50(s + α)(s + 2) . s(s + 3)(s + 4) + 50(s + α)(s + 2)  When α = 0, we have ess = 1 −  100 = 0.11 , 100 + 12  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  211  Advanced Problems  and, for α 6= 0 ess = 0 . The closed-loop step input and disturbance responses for α = 0, 10, 100 are shown in Figure AP5.5. For disturbance rejection and steady-state tracking error the best value of the parameter is α = 100 . However, when considering both the disturbance and input response we would select the parameter α = 10 , since it offers a good compromise between input response overshoot (about 5% for α = 10) and disturbance rejection/tracking error. AP5.6  (a) The closed-loop transfer function is T (s) =  KKm . KKm + s(s + Km Kb + 0.01)  The steady-state tracking error for a ramp input R(s) = 1/s2 is ess = lim s(1 − T (s))R(s) s→0  s + Km Kb + 0.01 s→0 KKm + s(s + Km Kb + 0.01) Km Kb + 0.01 = . KKm  = lim  (b) With Km = 10 and Kb = 0.05 , we have Km Kb + 0.01 10(0.05) + 0.01 = =1. KKm 10K Solving for K yields K = 0.051 .  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  212  CHAPTER 5  The Performance of Feedback Control Systems  (c) The plot of the step and ramp responses are shown in Figure AP5.6. The responses are acceptable. Step input response 1.4 1.2  y(t)  1 0.8 0.6 0.4 0.2 0  0  2  4  6  8  10 Time (sec)  12  14  16  18  20  14  16  18  20  Ramp input response 20  y(t)  15  10  5  0  0  2  4  6  8  10 Time (sec)  12  FIGURE AP5.6 Closed-loop system step and ramp responses.  AP5.7  The performance is summarized in Table AP5.7 and shown in graphical form in Fig. AP5.7.  K  Estimated Percent Overshoot  Actual Percent Overshoot  1000  8.8 %  8.5 %  2000  32.1 %  30.2 %  3000  50.0 %  46.6 %  4000  64.4 %  59.4 %  5000  76.4 %  69.9 %  TABLE AP5.7  Performance summary.  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  213  Advanced Problems  80  Percent Overshoot (% )  70  Actual P.O . Estimated P.O.  60 50 40 30 20 10 0 1000  2000  3000 K  4000  5000  FIGURE AP5.7 Percent overshoot versus K.  The closed-loop transfer function is T (s) =  100K . s(s + 50)(s + 100) + 100K  The impact of the third pole is more evident as K gets larger as the estimated and actual percent overshoot deviate in the range 0.3% at K = 1000 to 6.5% at K = 5000. AP5.8  The closed-loop transfer function is T (s) =  K(s + 2) . s2 + ( 23 + K)s + 13 + 2K  Comparing T (s) to a second-order system we have ωn =  q  1/3 + 2K  2/3 + K ζ= p 2 1/3 + 2K  For the closed-loop transfer function to have complex roots, we require K 2 − (20/3)K − (8/9) < 0. This occurs when −0.13 ≤ K ≤ 6.8. When K = 1/3, we have the minimum ζ = 0.5, as shown in Figure AP5.8.  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  214  CHAPTER 5  The Performance of Feedback Control Systems  0.58  0.57  0.56  ζ  0.55  0.54  0.53  0.52  0.51  0.5  0  0.05  0.1  0.15  0.2  0.25 K  0.3  0.35  0.4  0.45  0.5  FIGURE AP5.8 Damping ratio, ζ, versus K.  AP5.9  The closed-loop characteristic equation is s4 + 40s3 + 375s2 + KP s + KI = 0. The desired characteristic equation is √ (s + a)(s + b)(s2 + 2ωn s + ωn2 ) = 0. Expanding the desired characteristic equation and equating terms to the actual characteristic equation yields √ abωn2 = KI , ωn2 (a + b) + 2abωn = KP √ 2(a + b)ωn + ab = 375,  √  2ωn + a + b = 40  This represents 4 equations with 5 unknowns (a, b, KP , KI , and ωn ). We can choose one variable as part of the controller design. Let KI = 0.1KP . Then, solving the 4 equations for the remaining 4 variables yields a = 29.15, b = 0.1, KP = 1720, KI = 172, and ωn = 7.6. The resulting Ts = 1.1s and P.O. = 6.4%, as shown in Figure AP5.9.  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  215  Advanced Problems  Step Response 1.4 System: sysa Peak amplitude: 1.06 Overshoot (%): 6.39 At time (sec): 0.618  1.2  Amplitude  1  System: sysa Settling Time (sec): 1.09  0.8  0.6  0.4  0.2  0  0  1  2  3  4 5 Time (sec)  FIGURE AP5.9 Step response for KP = 1720 and KI = 172.  6  7  8  9  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  216  CHAPTER 5  The Performance of Feedback Control Systems  Design Problems The plant model with parameters given in Table CDP2.1 in Dorf and Bishop is given by: 26.035 θ(s) = , Va (s) s(s + 33.142) where we neglect the motor inductance Lm . The closed-loop transfer function from the disturbance to the output is θ(s) 26.035 = 2 . Td (s) s + 33.142s + 26.035Ka For a unit step disturbance input the steady-state response is θss =  1 . Ka  Therefore, we want to use the maximum Ka while keeping the percent overshoot less than 5%. The step response for the closed-loop system (with the tachometer not in the loop) and Ka = 22 is shown below. Values of Ka greater than 22 lead to overshoots greater than 5%.  Step response 1.4 1.2  q(t)/A  1 0.8 0.6 0.4 0.2 0  0  0.1  0.2  0.3  0.4  0.5 Time (sec)  0.6  0.7  0.8  0.9  1  0.7  0.8  0.9  1  Unit disturbance response 0.05 0.04 0.03 q(t)  CDP5.1  0.02 0.01 0  0  0.1  0.2  0.3  0.4  0.5 Time (sec)  0.6  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  217  Design Problems  DP5.1  (a) The closed-loop transfer function is 12.2K 12.2K φ(s) = = 3 . 2 φd (s) s(s + 2.2)(s + 7) + 12.2K s + 9.2s + 15.4s + 12.2K (b) For K = 0.7, we have the characteristic equation s3 + 9.2s2 + 15.4s + 8.54 = 0 , with roots s1 = −7.23 and s2,3 = −0.98 ± 0.46j. For K = 3, we have the characteristic equation s3 + 9.2s2 + 15.4s + 36.6 = 0 , with roots s1 = −7.83 and s2,3 = −0.68 ± 2.05j. And for K = 6, we have the characteristic equation s3 + 9.2s2 + 15.4s + 73.2 = 0 , with roots s1 = −8.4 and s2,3 = −0.4 ± 2.9j.  (c) Assuming the complex conjugate pair are the dominant roots, we expect the following: (i) for K = 0.7: P.O.=0.13% and Tp = 6.8 sec (ii) for K = 3: P.O.=35.0% and Tp = 1.5 sec (iii) for K = 6: P.O.=65.2% and Tp = 1.1 sec (d),(e) We select K = 1.71 to have a P.O. = 16% and Tp = 2.18sec. All four cases (K = 0.7, 3, 6, 1.71) are shown in Figure DP5.1. In each case, the approximate transfer function is derived by neglecting the non-dominant real pole and adjusting the gain for zero steady-state error. The approximate transfer functions are 1.18 0.7908 = + 1.965s + 1.18 (s + 0.98 + 0.46j)(s + 0.98 − 0.46j) 4.67 3.299 TK=3 (s) = 2 = s + 1.37s + 4.67 (s + 0.68 + 2.05j)(s + 0.68 − 2.05j) 8.71 6.399 TK=6 (s) = 2 = s + 0.796s + 8.71 (s + 0.4 + 2.9j)(s + 0.4 − 2.9j) 2.77 1.458 TK=1.71 (s) = 2 = s + 1.679s + 2.77 (s + 0.83 + 1.43j)(s + 0.83 − 1.43j) TK=0.7 (s) =  s2  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  218  CHAPTER 5  The Performance of Feedback Control Systems  K=0.7  K=3  1  1 phi  1.5  phi  1.5  0.5  0  0.5  0  5 time (sec) K=6  0  10  1.5  10  0  5 time (sec)  10  1 phi  phi  5 time (sec) K=1.71  1.5  2  1  0.5  0.5 0  0  0  5 time (sec)  10  0  FIGURE DP5.1 Step responses (actual response:solid lines; approximate response: dotted lines).  DP5.2  The closed-loop transfer function is T (s) =  Kωn2 , s3 + 2ζωn s2 + ωn2 s + Kωn2  where ζ = 0.6. From the second-order system approximation, we have Tp =  ωn  π . 1 − ζ2  p  So, with ζ = 0.6 given, we should select ωn "large" to make Tp "small." Also, from the problem hint, let 0.2 < K/ωn < 0.4 . As a first attempt, we can select ωn = 10. See Figure DP5.8 for various values of K/ωn . Our final selection is K = 3.33  and ωn = 10 .  This results in P.O. = 3.6% and Tp = 0.66 second.  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  219  Design Problems  1.4  1.2 K/ωn=0.4 1 K/ωn=0.33  y(t)  0.8 K/ω =0.2 n  0.6  0.4  0.2  0  0  0.5  1  1.5 Time (sec)  2  2.5  3  FIGURE DP5.2 Closed-loop system response.  DP5.3  The closed-loop transfer function is T (s) =  s2  K . + qs + K  From the ITAE specification, we desire T (s) =  ωn2 . s2 + 1.4ωn s + ωn2  But 2ζωn = 1.4ωn  which implies  ζ = 0.7 .  Since we want Ts ≤ 0.5, we require ζωn ≥ 8. So, ωn ≥  8 = 11.4 . 0.7  We can select ωn = 12. Then, T (s) =  144 . s2 + 16.8s + 144  Therefore, K = 144 and q = 16.8. The predicted percent overshoot is  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  220  CHAPTER 5  The Performance of Feedback Control Systems  P.O. = 4.5%. DP5.4  The loop transfer function is Gc (s)G(s) =  10K 10K/70 = . (s + 70)(s + 3)(s + 7) (s/70 + 1)(s + 3)(s + 7)  The second-order approximation is obtained by neglecting the fastest firstorder pole. Thus, K/7 . (s + 3)(s + 7)  Gc (s)G(s) ≈ The closed-loop transfer function is T (s) =  s2  K/7 . + 10s + 21 + K/7  When ζ ≥ 0.52, we have less than 15% overshoot. So, we have 2ζωn = 10 and ωn =  q  21 + K/7.  Eliminating ωn and solving for K (with P.O. ≤ 15%) yields K ≤ 500.19 . Also, Kp = lim GGc (s) = s→0  K 7(21)  and ess =  1 1 = < 0.12 K 1 + Kp 1 + 147  implies K ≥ 1078 . Therefore, we have an inconsistency. We require 1078 ≤ K to meet the steady-state requirement and K ≤ 500.18 to meet the percent overshoot requirement. It is not possible to meet both specifications.  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  221  Design Problems  DP5.5  The closed-loop characteristic equation is 1 + K1 G1 (s) + K2 G1 G2 (s) = 1 +  2K2 K1 − =0 s(s + 1) s(s + 1)(s + 2)  or s3 + 3s2 + (2 + K1 )s + 2(K1 − K2 ) = 0 . Assuming that K1 > 0 and K2 > 0, the range of the gains for stability is 0 < K2 < K1 . DP5.6  The closed-loop transfer function is T (s) =  s2  K1 . + (K1 K2 + 1)s + K1  The percent overshoot specification P.O. ≤ 2% is satisfied when ζ > 0.78. The peak time specification Tp ≤ 0.5 s is satisfied when ωn = 10 and ζ = 0.78. So, given K1 = ωn2  and  K1 K2 + 1 = 2ζωn ,  we determine that the specifications are satisfied when K1 = 100 and K2 = 0.15 . DP5.7  The plant is G(s) =  2 s(s + 1)(s + 4)  and the PD controller is Gc (s) = KD s + KP . The characteristic equation is s3 + 6s2 + (8 + 2KD )s + 2KP = 0. The desired characteristic equation is (s + a)(s2 + 2ζωn s + ωn2 ) = s3 + (2ζωn + a)s2 + (ωn2 + 2ζωn a)s + aωn2 = 0. Equating the desired characteristic equation to the actual characteristic  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  222  CHAPTER 5  The Performance of Feedback Control Systems  equation yields 2ζωn + a = 6,  ωn2 + 2ζωn a = 8 + 2KD ,  aωn2 = 2KP ,  where ζ = 0.69 and ωn = 3 to meet the design specifications. This represents 3 equations in 3 unknowns (a, KD , and KP ). Solving yields a = 1.86, KD = 4.35 and KP = 8.37. The step response is shown in Figure DP5.7.  Step Response 1.4  System: sys_cl Peak amplitude: 1.04 Overshoot (%): 4.28 At time (sec): 1.48  1.2  Amplitude  1  System: sys_cl Settling Time (sec): 1.96  0.8  0.6  0.4  0.2  0  0  0.5  1  1.5 Time (sec)  2  2.5  3  FIGURE DP5.7 Step response withKD = 4.35 and KP = 8.37.  DP5.8  The closed-loop transfer function is T (s) =  s2  K + 6s + 5 + K  The damping ratio and natural frequency is ζ=√  3 K +5  and  ωn =  √  K +5  Using the design formulas for second-order systems we have √ 4 2 P O = 100e−ζπ/ 1−ζ and Ts = . ζωn We know that the formula for Ts is approximate and that the formulas apply only to systems with ζ < 1. For K = 1 the closed-loop poles are  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  223  Design Problems  both real, so there is no overshoot and the design formula for settling time does not apply. Thus we obtain the results shown in Table DP5.8. We can choose K = 10 as a good trade-off between percent overshoot, settling time, and steady-state tracking error. The disturbance response is shown in Figure DP5.8. TABLE DP5.8  Step response for K=1, 10, and 20.  K  % P.O.  Ts , sec  Ts , sec  Estimated  Actual  ess  1  0  -  3.24  0.83  10  2.13  1.33  1.38  0.33  20  9.48  1.33  1.19  0.20  1 0.9 0.8 0.7  y(t)  0.6 0.5 0.4 0.3 0.2 0.1 0  0  0.2  0.4  0.6  0.8  1 Time (sec)  1.2  FIGURE DP5.8 Closed-loop system disturbance response for K = 10.  1.4  1.6  1.8  2  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  224  CHAPTER 5  The Performance of Feedback Control Systems  Computer Problems With the impulse input we have R(s) = 1. The transfer function is Y (s) =  15 15 R(s) = . (s + 3)(s + 5) (s + 3)(s + 5)  Therefore, taking the inverse Laplace transforms yields the output response: y(t) =  15 −3t 15 −5t e − e . 2 2  The impulse response and the analytic response is shown in Figure CP5.1.  n=15; d=[1 8 15]; t=[0:0.1:6]; ya=(15/2)*exp(-3.*t)-(15/2)*exp(-5.*t); sys = tf(n,d) y=impulse(sys,t); plot(t,y,t,ya,'o') xlabel('Time (sec)'), ylabel('y(t)'), legend('Computer','Analytic',-1) 1.4  Computer Analytic  1.2  1  0.8 y(t)  CP5.1  0.6  0.4  0.2  0  0  FIGURE CP5.1 Impulse responses.  1  2  3 Time (sec)  4  5  6  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  225  Computer Problems  CP5.2  The ramp response is shown in Figure CP5.2. The unity feedback system is type 2, so that the steady-state tracking error is lim ess → 0. t→∞  60  50  Amplitude  40  n=[1 10]; d=[1 15 0 0]; t=[0:0.1:50]; sys= tf(n,d); sys_cl = feedback(sys,[1]); u=t; lsim(sys_cl,u,t);  30  20  10  0  0  5  10  15  20  25 Time (sec)  30  35  40  45  50  FIGURE CP5.2 Ramp responses.  CP5.3  The m-file script and the four plots are shown in Figure CP5.3. The plots can be compared to Figure 5.17 in Dorf & Bishop. wn=2, zeta=0  2 1  1  0  0  -1  -1  -2  0  5  10  15  20  wn=1, zeta=0  1  wn=2, zeta=0.1  2  -2  0  5  15  20  wn=1, zeta=0.2  1  0.5  10  0.5  0 0  -0.5 -1  0  FIGURE CP5.3 Impulse responses.  5  10  15  20  -0.5  0  5  10  15  20  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  226  CHAPTER 5  The Performance of Feedback Control Systems  w1=2; z1=0; w2=2; z2=0.1; w3=1; z3=0; w4=1; z4=0.2; t=[0:0.1:20]; % num1=[w1^2]; den1=[1 2*z1*w1 w1^2]; sys1 = tf(num1,den1); [y1,x1]=impulse(sys1,t); % num2=[w2^2]; den2=[1 2*z2*w2 w2^2]; sys2 = tf(num2,den2); [y2,x2]=impulse(sys2,t); % num3=[w3^2]; den3=[1 2*z3*w3 w3^2]; sys3 = tf(num3,den3); [y3,x3]=impulse(sys3,t); % num4=[w4^2]; den4=[1 2*z4*w4 w4^2]; sys4 = tf(num4,den4); [y4,x4]=impulse(sys4,t); % clf subplot(221),plot(t,y1),title('wn=2, zeta=0') subplot(222),plot(t,y2),title('wn=2, zeta=0.1') subplot(223),plot(t,y3),title('wn=1, zeta=0') subplot(224),plot(t,y4),title('wn=1, zeta=0.2')  FIGURE CP5.3 CONTINUED: Impulse response m-file script.  CP5.4  The closed-loop system is T (s) =  s2  21 . + 2s + 21  Therefore, the natural frequency is √ ωm = 21 = 4.58 and the damping ratio is computed as 2ζωn = 2 , which implies ζ = 0.218 . The percent overshoot is estimated to be √ 2 P.O. = 100e−ζπ/ 1−ζ = 50% , since ζ = 0.218. The actual overshoot is shown in Figure CP5.4.  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  227  Computer Problems Step Response From: U(1) 1.5  To: Y(1)  1  Amplitude  numc=[21]; denc=[1 0]; sysc = tf(numc,denc); numg=[1]; deng=[1 2]; sysg = tf(numg,deng); sys_o = series(sysc,sysg); sys_cl = feedback(sys_o,[1]) step(sys_cl)  0.5  0  0  1  2  3  4  5  6  Time (sec.)  FIGURE CP5.4 Impulse responses.  CP5.5  For a step input, the optimum ITAE characteristic equation is s3 + 1.75ωs2 + 2.15ω 2 s + ω 3 = 0 . Examining Figure 5.30 for n=3 in Dorf & Bishop, we estimate that ωTs = 8. So, once we decide on the desired Ts we can estimate ω. For this design we let Ts =8 so that ω = 1. Computing the desired characteristic equation and the actual characteristic equation and comparing the coefficients leads to the following relationships: p = 1.75ω − 2ζωn K = (2.15ω 2 − ωn2 − 2ζωn p)/ωn2 z = (ω 3 − pωn2 )/(Kωn2 ) where ζ = 0.59 and ωn = 0.45. The controller and prefilter are Gc (s) = 6.42  s + 0.58 s + 1.22  and  Gp (s) =  The unit step response is shown in Figure CP5.5.  1 . 1.3s + 0.75  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  228  CHAPTER 5  The Performance of Feedback Control Systems  Step Response 1.4  wn=0.45; zeta=0.59 ng=wn^2; dg=[1 2*zeta*wn wn^2]; sysg=tf(ng,dg); Ts=8; w=8/Ts; p=1.75*w-2*zeta*wn; K=(2.15*w^2-wn^2-2*zeta*wn*p)/wn^2; z=(w^3-p*wn^2)/(K*wn^2); nc=K*[1 z]; dc=[1 p]; sysc=tf(nc,dc); sys=series(sysc,sysg); syscl=feedback(sys,1); [num,den]=tfdata(syscl,'v'); sysp=tf([den(end)],num); step(syscl*sysp)  System: untitled1 Peak amplitude: 1.02 Overshoot (%): 1.98 At time (sec): 4.68  1.2  1  Amplitude  System: untitled1 Settling Time (sec): 7.54  0.8  0.6  0.4  0.2  0  0  2  4  6 Time (sec)  8  10  12  FIGURE CP5.5 Closed-loop system step response m-file script.  CP5.6  The unit step response is shown in Figure CP5.6. The performance numbers are as follows: Mp = 1.16, Tp = 0.73, and Ts = 1.62.  Step Response 1.4  System: sys_cl Peak amplitude: 1.16 Overshoot (%): 16.3 At time (sec): 0.73  1.2  1 System: sys_cl Settling Time (sec): 1.62  Amplitude  numg=[25]; deng=[1 5 0]; sys = tf(numg,deng); sys_cl = feedback(sys,[1]); t=[0:0.01:2]; step(sys_cl,t);  0.8  0.6  0.4  0.2  0  0  FIGURE CP5.6 Closed-loop system step response m-file script.  0.2  0.4  0.6  0.8  1 Time (sec)  1.2  1.4  1.6  1.8  2  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  229  Computer Problems  The m-file script and the simulations are shown in Figure CP5.7. % Part (a) numc=[2]; denc=[1]; sys_c = tf(numc,denc); nums=[-10]; dens=[1 10]; sys_s = tf(nums,dens); numg=[-1 -5]; deng=[1 3.5 6 0]; sys_g = tf(numg,deng); sysa = series(sys_c,sys_s); sysb = series(sysa,sys_g); sys = feedback(sysb,[1]); f=0.5*pi/180; % Convert to rad/sec t=[0:0.1:10]; u=f*t; [y,x]=lsim(sys,u,t);(y(length(t),1)-u(1,length(t)))*180/pi subplot(211) plot(t,y*180/pi,t,u*180/pi,'--'), grid xlabel('Time (sec)'),ylabel('theta') title('Constant gain C(s) = 2: theta (solid) & input (dashed)') % Part (b) numc=[2 1]; denc=[1 0]; sys_c = tf(numc,denc); [numa,dena]=series(numc,denc,nums,dens); sysa = series(sys_c,sys_s); sysb = series(sysa,sys_g); sys = feedback(sysb,[1]); [y,x]=lsim(sys,u,t);(y(length(t),1)-u(1,length(t)))*180/pi subplot(212), plot(t,y*180/pi,t,u*180/pi,'--'), grid xlabel('Time (sec)'),ylabel('theta') title('PI controller C(s) = 2 + 1/s: theta (solid) & input (dashed)') Constant gain C(s) = 2: theta (solid) & input (dashed) 5  theta  4 3 2 1 0  0  1  2  3  4  5 Time (sec)  6  7  8  9  10  8  9  10  PI controller C(s) = 2 + 1/s: theta (solid) & input (dashed) 6 5 4 theta  CP5.7  3 2 1 0  0  1  2  3  4  5 Time (sec)  6  7  FIGURE CP5.7 Closed-loop system response to a ramp input for two controllers.  For the constant gain controller, the attitude error after 10 seconds is ess = −0.3 deg. On the other hand, the PI controller has a zero steadystate error ess = 0 deg. So, we can decrease the steady-state error by  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  230  CHAPTER 5  The Performance of Feedback Control Systems  using a more sophisticated controller, in this case a PI controller versus a constant gain controller. CP5.8  The closed-loop characteristic equation is s3 + 12s2 + 610s + 500 = (s + 0.8324)(s2 + 11.1676s + 600.7027) = 0 . The natural frequency and damping ratio of the complex roots are ωn = 24.5 and ζ = 0.23. From this we predict Mp = 1.48, Ts = 0.72, and Tp = 0.13. The actual response is shown in Figure CP5.8. The differences Step Response From: U(1) 1.4  1.2  1  0.8 To: Y(1)  theta dot  numg=[100 100]; deng=[1 2 100]; sysg = tf(numg,deng); numc=[0.1 5]; denc=[1 0]; sysc = tf(numc,denc); sys_o = series(sysg,sysc); sys_cl = feedback(sys_o,[1]) t=[0:0.01:3]; step(sys_cl,t); ylabel('theta dot')  0.6  0.4  0.2  0  0  0.5  1  1.5  2  2.5  3  Time (sec.)  FIGURE CP5.8 Missile rate loop autopilot simulation.  can be explained by realizing that the system is not a second-order system. The closed-loop system actually has two zeros, one real pole, and two complex-conjugate poles: T (s) =  (s + 50)(s + 1) . (s + 0.8324)(s2 + 11.1676s + 600.7027)  The effect of the pole at s = −0.8324 is diminished by the zero at s = −1. The third pole and the zeros affect the overall response such that the analytic formulas for second-order systems are not exact predictors of the transient response.  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  231  Computer Problems  CP5.9  Figure CP5.9 shows an m-file to compute the closed-loop transfer function and to simulate and plot the step response. Step Response 1 System: sys Peak amplitude: 0.979 Overshoot (%): 95.7 At time (sec): 0.533  0.9 0.8  numg=[10]; deng=[1 10]; sysg = tf(numg,deng); numh=[0.5]; denh=[10 0.5]; sysh = tf(numh,denh); sys = feedback(sysg,sysh) step(sys);  Amplitude  0.7  Transfer function: 100 s + 5 --------------------10 s^2 + 100.5 s + 10  0.6  System: sys Settling Time (sec): 39.1  0.5 0.4 0.3 0.2 0.1 0  0  10  20  30 Time (sec)  40  50  60  FIGURE CP5.9 M-file to compute the transfer function and to simulate the step response.  CP5.10  Figure CP5.10 shows an m-file to compute the closed-loop transfer function and to simulate and plot the ramp response. The steady-state error Linear Simulation Results 100 90 80 70 Amplitude  numg=[10]; deng=[1 20 75 0]; sysg = tf(numg,deng); sys = feedback(sysg,1) t=[0:0.1:100]; u=t; % Unit ramp input lsim(sys,u,t);  60 50 40 30 20 10 0  0  10  20  30  40  50 Time (sec)  FIGURE CP5.10 M-file to compute the transfer function and to simulate the ramp response.  is 7.5.  60  70  80  90  100  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  232  CHAPTER 5  CP5.11  The Performance of Feedback Control Systems  Figure CP5.11 shows an m-file to compute the closed-loop transfer function and to simulate and plot the impulse, step, and ramp responses. Notice that the closed-loop system is unstable.  0  Amplitude  -10  0  2  4  6  0  2  4  6  0  2  4  6  8  10 Time (sec) Step Response  12  14  16  18  20  8 10 12 Time (sec) Linear Simulation Results  14  16  18  20  14  16  18  20  10 0 -10  Amplitude  numg=[1]; deng=[1 2 0]; sysg = tf(numg,deng); numc=[0.5 2]; denc=[1 0]; sysc = tf(numc,denc); syss=series(sysg,sysc); sys = feedback(syss,1) t=[0:0.1:20]; subplot(311) impulse(sys,t); subplot(312) step(sys,t); subplot(313) u=t; % Unit ramp input lsim(sys,u,t);  Amplitude  Impulse Response 10  40 20 0  8  10 Time (sec)  12  FIGURE CP5.11 M-file to compute the transfer function and to simulate the ramp response.  CP5.12  Figure CP5.12 shows an m-file to simulate and plot the step response for the original system and the 2nd-order approximation. For the original system, we find Ts = 2.28 and P.O. = 80.6%. For the 2nd-order approximation we find Ts = 2.16 and P.O. = 101%  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  233  Computer Problems  2.5  2nd order approximation 2  3rd order system response Step response  num=77*[1 2]; den=conv([1 7],[1 4 22]); sys = tf(num,den) na=(77/7)*[1 2]; da=[1 4 22]; sysa=tf(na,da); t=[0:0.01:5]; y=step(sys,t); ya=step(sysa,t); plot(t,y,t,ya,'--') xlabel('Time (s)'), ylabel('Step response')  1.5  1  0.5  0  0  1  2  3 Time (s)  FIGURE CP5.12 Step response.  4  5  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  C H A P T E R  6  The Stability of Linear Feedback Systems  Exercises E6.1  The Routh array is s3  1  K +1  s2  K  6  s1  b  0  so  6  where b=  K(K + 1) − 6 . K  For stability, we require K > 0 and b > 0. Therefore, using the condition that b > 0, we obtain K2 + K − 6 > 0 , and solving for K yields K > 2 and K < −3. We select K > 2, since we also have the condition that K > 0. E6.2  The Routh array is s3  1  2  s2  10  30  s1  -1  0  so  30  The system is unstable since the first column shows two sign changes. 234  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  235  Exercises  E6.3  The Routh array is s4  1  32  s3  10  37  s2  28.3  20  s1  29.9  s0  20  20  By the Routh-Hurwitz criterion, the system is stable (i.e., all the numbers in the first column are positive). E6.4  The closed-loop transfer function is T (s) =  s3  +  −K(s − 2) . + (4 − K)s + 2K  5s2  Therefore, the characteristic equation is s3 + 5s2 + (4 − K)s + 2K = 0 . The corresponding Routh array is given by s3  1  s2  4−K  5  2K  s1  b  0  so  2K  where b=  5(4 − K) − 2K 20 − 7K = . 5 5  For stability we require K > 0 and b > 0 . Thus, the range of K for stability is 0 < K < 20/7. E6.5  The closed-loop transfer function is T (s) =  s3  +  10s2  K . + 27s + 18 + K  When K = 20, the roots of the characteristic polynomial are s1,2 = −1.56 ± j1.76  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  236  CHAPTER 6  The Stability of Linear Feedback Systems  and s3 = −6.88 . E6.6  When K = 252, the roots of the characteristic equation are on the imaginary axis. The roots are s1,2 = ±j5.2  E6.7  and  s3 = −10 .  (a) The closed-loop system characteristic equation is 1 + GH(s) = 1 + or  K(s + 2) =0, s(s − 1)  s2 + (K − 1)s + 2K = 0 . √ We have the relationships ωn = 2K and 2ζωn = K − 1, where ζ = 0.707. Thus,   1 √ 2 √ 2K = K − 1 , 2  or   2 √ 2  2  =    K −1 √ 2K  2  ,  and K 2 − 6K + 1 = 0 . Solving for K yields K = 5.83 and K = 0.17. However, for stability we require K > 1 (from the Routh array), so we select K = 5.83. √ (b) The two roots on the imaginary axis when K = 1 are s1,2 = ±j 2. E6.8  The closed-loop system characteristic equation is 3  + 20s2 + (100 + K)s + 20K = 0 .  The corresponding Routh array is s3  1  (100 + K)  s2  20  20K  s1  b  0  so  20K  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  237  Exercises  where b=  20(100 + K) − 20K 20(100) = = 100 . 20 20  Therefore, the system is stable for all K > 0. E6.9  The characteristic equation is s3 + 2s2 + (K + 1)s + 8 = 0 , and the Routh array is given by s3  1  K +1  s2  2  8  s1  b  0  so  8  where b=  2(K + 1) − 8 =K −3 . 2  Setting b = 0, yields K − 3 = 0 or  K>3.  E6.10  Stable with your eyes open and (generally) unstable with your eyes closed.  E6.11  The system is unstable. The poles are s1 = −5.66, s2 = −0.90 and s3,4 = 0.28 ± j0.714.  E6.12  The characteristic equation is s2 + as + b = 0, so, the Routh array is s2  1  b  s1  a  0  so  b  The system is stable if and only if a > 0 and b > 0. For a second-order system, a necessary and sufficient condition for stability is that all the coefficients have the same sign. E6.13  The characteristic equation is s2 + (KD + 2)s + 4KP = 0.  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  238  CHAPTER 6  The Stability of Linear Feedback Systems  The Routh array is s2  1  4KP  s1  KD + 2  0  so  4KP  The system is stable if and only if KP > 0 and KD > −2. E6.14  The characteristic equation associated with the system matrix is s3 + 3s2 + 5s + 6 = 0 . The roots of the characteristic equation are s1 = −2 and s2,3 = −5±j1.66. The system is stable.  E6.15  The roots of q(s) are s1 = −4, s2 = −3, s3,4 = −1 ± j2 and s5,6 = ±j0.5. The system is marginally stable. The Routh array is s6  1  31.25  67.75  s5  9  61.25  14.75  s4  24.44  66.11  15  s3  31.909  9.2273  0  s2  60  15  s1  0  0  15  so The auxillary equation is 60s2 + 15 = 0 . Solving the auxillary equation yields two roots at s1,2 = ±j0.5. After accounting for the row of zeros, the completed Routh array verifies that the system has no poles in the right half-plane. E6.16  The Routh array is s4  1  45  s3  9  87  s2  35.33  50  s1  74.26  0  so  50  50  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  239  Exercises  The system is stable. The roots of q(s) are s1,2 = −3 ± j4, s3 = −2 and s4 = −1. E6.17  The characteristic equation is s3 + 7s2 + 36s + 24 = 0 . The system is stable. The roots of the characteristic equation are s1 = −0.77, s2,3 = −3.12 ± 4.64i.  E6.18  The roots of q(s) are s1 = −20 and s2,3 = ±j2.24. The system is marginally stable. The Routh array is s3  1  5  s2  20  100  s1  0  0  so The auxillary equation is 20s2 + 100 = 0 . The roots are s = ±j2.24. So, the system has roots at s = ±j2.24. Completing the Routh array (after accounting for the row of zeros) verifies that no poles lie in the right half-plane. E6.19  (a) Unstable. (b) Stable. (c) Stable.  E6.20  (a) The roots are s1,2 = −2 and s3 = −1. (b) The roots are s1,2,3 = −3.  E6.21  The characteristic equation is (sn − 2)3 + 10(sn − 2)2 + 29(sn − 2) + K = 0 or s3n + 4s2n + sn − 26 + K = 0 . The Routh array is  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  240  CHAPTER 6  The Stability of Linear Feedback Systems  s3  1  1  s2  4  s1  30−K 4  K − 26  so  K − 26  0  If K = 30, then the auxillary equation is 4s2n + 4 = 0 or sn = ±j. Therefore, s = sn − 2 implies s = −2 ± j. E6.22  This system is not stable. The output response to a step input is a ramp y(t) = kt.  E6.23  The characteristic polynomial is s3 + 4s2 + ks + 8 = 0 . The Routh array is s3  1  k  s2  4  8  s1  4k−8 4  so  8  So, k > 2 for stability. E6.24  The transfer function is G(s) = C(sI − A)−1 B + D    k  0  k  s+k    2  s + ks + k  = [ 1 0 0 ] −k          0      −1   0    s −1  = [ 1 0 0 ]  0 s  −ks    1  s+k    s2 + ks      1  0      s   0   −ks − k s2    1  1  ∆(s)  where ∆(s) = s3 + ks2 + ks + k. Thus, the transfer function is G(s) = The Routh array is  s3  +  ks2  1 . + ks + k  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  241  Exercises  s3  1  k  s2  k  k  s1  k−1  so  k  For stability k > 1. E6.25  The closed-loop transfer function is T (s) =  Ks + 1 . s2 (s + p) + Ks + 1  Therefore, the characteristic equation is s3 + ps2 + Ks + 1 = 0 . The Routh array is s3  1  K  s2  p  1  s1  (pK − 1)/p  so  1  We see that the system is stable for any value of p > 0 and pK − 1 > 0. E6.26  The closed-loop transfer function is T (s) =  10 . 2s2 + (K − 20)s + 10 − 10K  Therefore, the characteristic equation is 2s2 + (K − 20)s + 10 − 10K = 0 . The Routh array is s2  2  s1  K − 20  so  10-10K  10 − 10K  We see that the system is stable for any value of K > 20 and any K < 1. Therefore, the system is unstable for all K > 0 since the gain K cannot be simultaneously greater than 20 and less than 1.  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  242  CHAPTER 6  The Stability of Linear Feedback Systems  Problems P6.1  (a) Given s2 + 5s + 2 , we have the Routh array s2  1  2  s1  5  0  so  2  Each element in the first column is positive, thus the system is stable. (b) Given s3 + 4s2 + 8s + 4 , we have the Routh array s3  1  8  s2  4  4  s1  7  0  so  4  Each element in the first column is positive, thus the system is stable. (c) Given s3 + 2s2 − 6s + 20 , we determine by inspection that the system is unstable, since it is necessary that all coefficients have the same sign. There are two roots in the right half-plane. (d) Given s4 + s3 + 2s2 + 12s + 10 , we have the Routh array s4  1  2  10  s3  1  12  0  s2  -10  10  0  s1  13  0  so  10  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  243  Problems  There are two sign changes in the first column, thus the system is unstable with two roots in the right half-plane. (e) Given s4 + s3 + 3s2 + 2s + K , we have the Routh array s4  1  3  K  s3  1  2  0  s2  1  K  s1  2−K  0  so  K  Examining the first column, we determine that the system is stable for 0 < K < 2. (f) Given s5 + s4 + 2s3 + s + 6 , we know the system is unstable since the coefficient of the s2 term is missing. There are two roots in the right half-plane. (g) Given s5 + s4 + 2s3 + s2 + s + K , we have the Routh array s5  1  2  1  s4  1  1  K  s3  1  s2  1−K  K  K  s1  −K  0  so  K  Examining the first column, we determine that for stability we need K > 0 and K < 0. Therefore the system is unstable for all K. P6.2  (a) The closed-loop characteristic polynomial is s4 + 27.88s3 + 366.4s2 + 1500s + 1500ka = 0 . The Routh array is  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  244  CHAPTER 6  The Stability of Linear Feedback Systems  s4  1  366.4  s3  27.88  1500  s2  312.6  1500ka  s1  b  so  1500ka  1500ka  where b = 1500 − 133.78ka . Examining the first column of the Routh array, we find that b > 0 and 1500ka > 0 for stability. Thus, 0 < ka < 11.21 . (b) With Ts = 1.5 =  4 , ζωn  we determine that ζωn = 2.67 . So, shift the axis by s = so − 2.67, and (so − 2.67)4 + 27.88(so − 2.67)3 + 366.4(so − 2.67)2 + 1500(so − 2.67) + 1500ka = s4o + 17.2s3o + 185.85s2o + 63.55so − 1872.8 + 1500ka . The Routh array is s4  1  185.85  s3  17.2  63.55  s2  182.16  1500ka -1872.8  s1  b  so  1500ka -1872.8  1500ka -1872.8  where b = 240.38 − 141.63ka . Examining the first column of the Routh array, we find that b > 0 and 1500ka − 1872.8 > 0. Thus, 1.25 < ka < 1.69.  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  245  Problems  P6.3  (a) Given G(s) =  K , (s + 1)(s + 2)(0.5s + 1)  and H(s) =  1 , 0.005s + 1  the closed-loop transfer function is T (s) =  0.0025s4  K(0.005s + 1) . + 0.5125s3 + 2.52s2 + 4.01s + 2 + K  Therefore, the characteristic equation is 0.0025s4 + 0.5125s3 + 2.52s2 + 4.01s + (2 + K) = 0 . The Routh array is given by s4  0.0025  2.52  2+K  s3  0.5125  4.01  0  s2  2.50  2+K  s1  3.6 − 0.205K  0  so  2+K  Examining the first column, we determine that for stability we require −2 < K < 17.6 . (b) Using K = 9, the roots of the characteristic equation are s1 = −200 ,  s2,3 = −0.33 ± 2.23j ,  and  s4 = −4.35 .  Assuming the complex roots are dominant, we compute the damping ratio ζ = 0.15. Therefore, we estimate the percent overshoot as √ 2 P.O. = 100e−πζ/ 1−ζ = 62% . The actual overshoot is 27%, so we see that assuming that the complex poles are dominant does not lead to accurate predictions of the system response. P6.4  (a) The closed-loop characteristic equation is 1 + GH(s) = 1 +  K(s + 40) =0, s(s + 10)(s + 20)  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  246  CHAPTER 6  The Stability of Linear Feedback Systems  or s3 + 30s2 + 200s + Ks + 40K = 0 . The Routh array is s3  1  200 + K  s2  30  40K  s1  200 −  so  K 3  0  40K  Therefore, for stability we require 200 − K/3 > 0 and 40K > 0. So, the range of K for stability is 0 < K < 600 . (b) At K = 600, the auxilary equation is 30s2 + 40(600) = 0  or  s2 + 800 = 0 .  The roots of the auxiliary equation are s = ±j28.3 . (c) Let K = 600/2 = 300. Then, to the shift the axis, first define so = s + 1. Substituting s = so − 1 into the characteristic equation yields (so −1)3 +30(so −1)2 +500(so −1)+12000 = s3o +27s2o +443so +11529 . The Routh array is s3  1  443  s2  27  11529  s1  16  0  so  11529  All the elements of the first column are positive, therefore all the roots lie to left of s = −1. We repeat the procedure for s = so − 2 and obtain s3o + 24s2o + 392so + 10992 = 0 . The Routh array is  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  247  Problems  s3  1  392  s2  24  10992  s1  -66  0  so  10992  There are two sign changes in the first column indicating two roots to right of s = −2. Combining the results, we determine that there are two roots located between s = −1 and s = −2. The roots of the characteristic equation are s1 = −27.6250  and  s2,3 = −1.1875 ± 20.8082j .  We see that indeed the two roots s2,3 = −1.1875±20.8082j lie between -1 and -2. P6.5  (a) Given the characteristic equation, s3 + 3s2 + 4s + 2 = 0 , we compute the roots s1 = −1, and s2,3 = −1 ± j.  (b) The roots of the characteristic equation  s4 + 9s3 + 30s2 + 42s + 20 = 0 are s1 = −1, s2 = −2, and s3,4 = −3 ± j1.  (c) The roots of the characteristic equation  s3 + 19s2 + 110s + 200 = 0 are s1 = −4, s2 = −5, and s3 = −10. P6.6  (a) The characteristic equation is 1 + G(s) = 0 , or s3 + s2 + 10s + 2 = 0 . The roots are: s1 = −0.2033, and s2,3 = −0.3984 ± j3.1112.  (b) The characteristic equation is  s4 + 10s3 + 35s2 + 50s + 24 = 0 . The roots are s1 = −1, s2 = −2, s3 = −3, and s4 = −4.  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  248  CHAPTER 6  The Stability of Linear Feedback Systems  (c) The characteristic equation is s3 + 11s2 + 29s + 6 = 0 . The roots are s1 = −0.2258, s2 = −3.8206 and s3 = −6.9536. P6.7  (a) The closed-loop characteristic equation is s3 + 101s2 + (100 + 10KKa )s + 100KKa = 0 . The Routh array is s3  1  100 + 10KKa  s2  101  100KKa  s1  b  so  100KKa  where b = 100 +  910 KKa > 0 . 101  Thus, examing the first column, we determine that KKa > 0 stabilizes the system. (b) The tracking error is e(s) = lim s(1 − T (s)) s→0  100 100 = . 2 s KKa  We require E(s) < 1o = 0.01745. So, KKa >  100 = 5729 . 0.01745  When KKa = 5729, the roots of the characteristic polynomial are s1 = −10.15 P6.8  and  s2,3 = −45.43 ± j233.25 .  (a) The closed-loop characteristic equation is 1+  K =0, (0.5s + 1)(s + 1)( 14 s + 1)  or s3 + 7s2 + 14s + 8(1 + K) = 0 . The Routh array is  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  249  Problems  s3  1  14  s2  7  8(1 + K)  s1  b  so  8(1 + K)  where 7(14) − 8(1 + K) . 7  b=  For stability, we require b > 0 and 8(1 + K) > 0. Therefore, the range of K for stability is −1 < K < 11.25 . (b) Let K = 11.25/3 = 3.75. Then, the closed-loop transfer function is T (s) =  s3  +  3.37 . + 14s + 38  7s2  The settling time to a step input is Ts ≈ 6 seconds. (c) We want Ts = 4 sec, so Ts = 4 =  4 ζωn  implies  ζωn = 1 .  Our desired characteristic polynomial is (s + b)(s2 + 2ζωn s + ωn2 ) = s3 + (2 + b)s2 + (ωn2 + 2b)s + bωn2 where we have used the fact that ζωn = 1 and ωn and b are to be determined. Our actual characteristic polynomial is s3 + 7s2 + 14s + 8(1 + K) = 0 . Comparing the coefficients of the actual and desired characteristic polynomials, we find the following relationships: ωn2  2+b=7 + 2b = 14 bωn2 = 8(1 + K) .  Solving these three equations yields b=5,  ωn = 2 and K = 1.5 .  The actual settling time is Ts = 4.17 sec. This is not exactly our  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  250  CHAPTER 6  The Stability of Linear Feedback Systems  desired Ts since we have the contribution of the additional pole at s = −5. The closed-loop poles are s1 = −5 and s2,3 = −1 ± 1.73j . P6.9  (a) The closed-loop characteristic equation is 1 + GH(s) = 1 +  10K , (s + 100)(s + 20)2  or s3 + 140s2 + 4400s + 40000 + 10K = 0 . The Routh array is s3  1  4400  s2  140  40000 + 10K  s1  b  so  40000 + 10K  where b=  140(4400) − (40000 + 10K) . 140  Examining the first column and requiring all the terms to be positive, we determine that the system is stable if −4000 < K < 57600 . (b) The desired characteristic polynomial is (s+b)(s2 +1.38ωn s+ωn2 ) = s3 +(1.38ωn +b)s2 +(ωn2 +1.38ωn b)s+bωn2 where we have used the fact that ζ = 0.69 to achieve a 5% overshoot, and ωn and b are to be determined. The actual characteristic polynomial is s3 + 140s2 + 4400s + 40000 + 10K = 0 . Equating the coefficients of the actual and desired characteristic polynomials, and solving for K, b, and ωn yields b = 104.2 ,  ωn = 25.9  So, a suitable gain is K = 3003.  and  K = 3003 .  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  251  Problems  P6.10  (a) The closed-loop characteristic equation is s4 + 7s3 + 20s2 + (24 + K)s + 10K = 0 . The Routh array is s4  1  20  10K  s3  7  24 + K  0  s2  116−K 7  10K  s1  b  so  10K  where b=    116−K 7    (24 + K) − 70K    Setting b > 0 yields  116−K 7    .  2784 − 398K − K 2 > 0 , which holds when −404.88 < K < 6.876 . Examining the first column, we also find that K < 116 and K > 0 for stability. Combining all the stability regions, we determine that for stability 0 < K < 6.876 . (b) When K = 6.876, the roots are s1,2 = −3.5 ± 1.63j , P6.11  and  s3,4 = ±2.1j .  Given s3 + (1 + K)s2 + 10s + (5 + 15K) = 0 , the Routh array is s3  1  10  s2  1+K  5 + 15K  s1  b  so  5 + 15K  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  252  CHAPTER 6  The Stability of Linear Feedback Systems  where b=  5 − 5K (1 + K)10 − (5 + 15K) = . 1+K 1+K  Given that K > 0, we determine that the system is stable when 5−5K > 0 or 0              0, ab − c > 0 and c > 0. When a > 0 and c > 0, we know that b > 0. So, a necessary condition for stability is that all coefficients a, b, and c be positive. The necessary and sufficient conditions for stability also require that b > c/a, in addition to a > 0 and c > 0. P6.13  The characteristic equation is s3 + (p + 2ζωn )s2 + (2ζωn p + Kωn2 )s + Kωn2 z = 0. The conditions for stability (see P6.12) are p + 2ζωn > 0, 2ζωn p + Kωn2 > (Kωn2 z)/(p + 2ζωn ), and Kωn2 z > 0. Since we know that K > 0, ζ > 0, and ωn > 0, it follows that for stability z > 0, p > −2ζωn , and 2ζωn p + Kωn2 >  P6.14  Kωn2 z . p + 2ζωn  The system has the roots s1,2 = ±j ,  s3,4 = ±j ,  and s5,6 = −1 ± 3j ,  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  253  Problems  Therefore, the system is not stable since there are repeated roots on the jω-axis. P6.15  (a) Neglecting the zeros and poles, we have the characteristic equation s4 + 30s3 + 325s2 + 2500s + K = 0 . The Routh array is s4  1  325  K  s3  30  2500  0  s2  241.67  K  s1  b  so  K  where b=  604166.67 − 30K . 241.67  Therefore, the system is stable for 0 < K < 20139. (b) Without neglecting the zeros and poles, the closed-loop characteristic equation is s6 + 90s5 + 5525s4 + 12400s3 + (1255000 + K)s2 + (8500000 + 30K)s + 1125K = 0 . This is stable for 0 < K < 61818 . We see that the additional poles and zero makes the system stable for a much larger gain K. P6.16  (a) The Routh array is s3  1  5  s2  5  6  s1  3.8  so  6  Examining the first column of the Routh array, we see no sign changes. So, the system is stable. (b) The roots of the system are s1 = −0.3246 and s2,3 = −2.3377 ± 3.6080j.  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  254  CHAPTER 6  The Stability of Linear Feedback Systems  (c) The step response is shown in Figure P6.16.  Step Response 0.18 0.16 0.14  Amplitude  0.12 0.1 0.08 0.06 0.04 0.02 0  0  5  10  15  Time (sec )  FIGURE P6.16 Unit step response.  P6.17  The closed-loop transfer function is T (s) =  s3  +  K +1 . + 3s + K + 1  3s2  The Routh array is s3  1  3  s2  3  K+1  s1  8−K 3  so  K +1  So, for stability we require −1 < K < 8. P6.18  The system characteristic equation is s2 + (h − k)s + ab − kh = 0 . For stability we require h > k and ab > kh. If k > h, the system is unstable.  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  255  Problems  P6.19  (a) The characteristic equation is s3 + 9s2 + (K − 10)s + 2K = 0 . The Routh array is s3  1  s2  K − 10  9  2K  s1  7K−90 9  so  2K  For stability K > 90/7 . (b) When K = 90/7, the system is marginally stable. The roots are q  s1,2 = ±j 20/7 , at the jω-axis crossing. P6.20  The closed-loop characteristic equation is q(s) = s5 + s4 + 9s3 + Ks2 + 2Ks + K . The range of stability for the vertical-liftoff vehicle is 5.177 < K < 7.823 . Therefore, for K = 6, the system is stable. When K = 6 we have q(s) = s5 + s4 + 9s3 + 6s2 + 12s + 6 The Routh array is s5  1  9  12  s4  1  6  6  s3  3  6  s2  4  6  s1  1.5  so  6  All entries in the first column are positive, so the system is stable.  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  256  CHAPTER 6  P6.21  The Stability of Linear Feedback Systems  The state transition matrix is     (k − p1 )e−p1 t − (k2 − p2 )e−p2 t e−p1 t − e−p2 t 1   2 Φ(t, 0) = −p t −p t −p t −p t p2 − p1 1 2 1 2 −k1 e + k1 e −p1 e + p2 e where p1 p2 = k1 and p1 + p2 = k2 . We assume that p1 6= p2 . In the case when p1 = p2 , the state transition matrix will change, but the factors e−p1 t and e−p2 t will remain. The eigenvalues of A are given by the solution to det |λI − A| = λ2 + k2 λ + k1 = 0 . q  Therefore, the eigenvalues are λ1,2 = −k2 /2 ± k22 − 4k1 . If k2 > 0 and k1 > 0, then the eigenvalues are in the left half-plane, and the system is stable. The transfer function is given by G(s) = C (sI − A)−1 B = −  s−1 . s2 + k2 s + k1  Therefore the characteristic equation is s2 + k2 s + k1 = 0 and the poles q are s1,2 = −k2 /2 ± k22 − 4k1 . If k2 > 0 and k1 > 0, then the poles are in the left half-plane, and the system is stable. Notice that the values of λ1,2 and s1,2 are the same. Also, the eigenvalues are the same as the values of −p1 and −p2 . So, if the eigenvalues are negative, then the elements of the state transition matrix will decay exponentially.  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  257  Advanced Problems  Advanced Problems The Routh array is s4  1  K1  s3  20  4  s2  20K1 −4 20  K2  s1  b  0  so  K2  K2  where b=  20K1 − 4 − 100K2 . 5K1 − 1  For stability, we require K2 > 0, K1 > 0.2, and b > 0. Therefore, using the condition that b > 0, we obtain K2 < 0.2K1 − 0.04 . The stability region is shown in Figure AP6.1. 0. 4  0.35  0. 3  0.25  K2  AP6.1  0. 2  0.15  0. 1  0.05 STABLE REGION 0 0. 2  0. 4  0. 6  0. 8  1  1. 2 K1  FIGURE AP6.1 Stability region.  1. 4  1. 6  1. 8  2  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  258  CHAPTER 6  AP6.2  The Stability of Linear Feedback Systems  The Routh array is s4  1  30  s3  9  s2  310−K 9  K − 40  s1  b  so  K  K  K 0  where b=  (310 − K)(K − 40) − 81K . 310 − K  Therefore, using the condition that b > 0, we obtain the stability range for K: 59.07 < K < 209.94 . AP6.3  (a) The steady-state tracking error to a step input is ess = lim s(1 − T (s))R(s) = 1 − T (0) = 1 − α . s→0  We want |1 − α| < 0.05 . This yields the bounds for α 0.95 < α < 1.05 . (b) The Routh array is s3  1  α  s2  1+α  1  s1  b  0  so  1  where b=  α2 + α − 1 . 1+α  Therefore, using the condition that b > 0, we obtain the stability range for α: α > 0.618 .  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  259  Advanced Problems  (c) Choosing α = 1 satisfies both the steady-state tracking requirement and the stability requirement. The closed-loop transfer function is T (s) =  s3  K . + (p + 1)s2 + ps + K  The Routh array is s3  1  p  s2  1+p  K  s1  b  0  so  K  where b=  p2 + p − K . 1+p  Therefore, using the condition that b > 0, we obtain the the relationship K < p2 + p . The plot of K as a function of p is shown in Figure AP6.4.  120  100  80  K  AP6.4  60  40 STABLE REGION 20  0 0  1  2  3  4  5 p  FIGURE AP6.4 Stability region.  6  7  8  9  10  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  260  CHAPTER 6  The closed-loop transfer function is T (s) =  30K1 K2 . (s + 1 + K1 K3 )(s − 10)(2s + K2 K3 − 4) + 30K1 K2 K4  The Routh array is s3  2  a  s2  b  c  s1  d  0  so  c  where a = −9K2 K3 + 16 + K1 K2 K32 − 24K1 K3 , b = 2K1 K3 + K2 K3 − 22, and c = −10K2 K3 + 40 − 10K1 K2 K32 + 40K1 K3 and d = (ab − 2c)/b . The conditions for stability are 2K1 K3 + K2 K3 − 22 > 0 −10K2 K3 + 40 − 10K1 K2 K32 + 40K1 K3 > 0 −2(−10K2 K3 + 40 − 10K1 K2 K32 + 40K1 K3 ) + (9K2 K3 +16 + K1 K2 K32 − 24K1 K3 )(2K1 K3 + K2 K3 − 22) > 0 Valid values for the various gains are: K1 = 50, K2 = 30, K3 = 1, and K4 = 0.3. The step response is shown in Figure AP6.5.  Step Response 350  300  250  Amplitude  AP6.5  The Stability of Linear Feedback Systems  200  150  100  50  0  FIGURE AP6.5 Stability region.  0  5  10  15 Time (sec)  20  25  30  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  261  Advanced Problems  AP6.6  The characteristic equation is s3 + 7s2 + (KD + 14)s + KP + 8 = 0. For stability we require that KP > −8 and KD >  KP + 8 − 14. 7  The relationship between KD and KP is shown in Figure AP6.6. 2  0  −2  STABLE REGION  K  D  −4  −6  −8  UNSTABLE REGION −10  −12  −14  0  10  20  30  40  50  60  70  80  90  100  K  P  FIGURE AP6.6 Stability region.  AP6.7  The characteristic equation is 0.1s4 + 2.05s3 + s2 + 8KP s + 8KI = 0. From the Routh array we find the conditions for stability are 0 < KI < 0.3125 1.2812 −  p  1.6416 − 5.2531KI < KP < 1.2812 +  p  1.6416 − 5.2531KI  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  262  CHAPTER 6  The Stability of Linear Feedback Systems  Design Problems CDP6.1  The plant model with parameters given in Table CDP2.1 in Dorf and Bishop is given by: θ(s) 26.035 = . Va (s) s(s + 33.142) In the above transfer function we have neglected the motor inductance Lm . The closed-loop transfer function from the input to the output is 26.035Ka θ(s) = 2 . R(s) s + 33.142s + 26.035Ka The Routh array is s2  1  26.035Ka  s1  33.142  0  s0  26.035Ka  Stability is achieved for any 0 < Ka < ∞. DP6.1  The closed-loop characteristic polynomial is 1 1 1 s3 + s2 (5 + p + K) + s( Kp + K + 5p) + K = 0 . 5 5 5 (i) When p = 2, we have 1 3 s3 + s2 (7 + K) + s(10 + K) + K = 0 . 5 5 The Routh array is s3  1  s2  7+  s1  b  so  K  10 + 35 K K 5  K  where b=  (7 + K/5)(10 + 3K/5) − K . 7 + 51 K  When −32.98 < K < −17.69, we find that b > 0. Examining the other terms in the first column of the array, we find that the system  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  263  Design Problems  is stable for any K > 0. (ii) When p = 0, we have 1 1 s3 + s2 (5 + K) + s( K) + K = 0 . 5 5 The Routh array is s3  1  1 5K  s2  5 + 15 K  K  s1  b  so  K  where b=  (5 + 15 K) 15 K − K K 2 /25 = . (5 + K/5) (5 + K/5)  Again, examination of the first column reveals that any K > 0 results in a stable system. So, we just need to select any K > 0; e.g. K = 10. DP6.2  (a) The closed-loop characteristic equation is 1+  20(Ks + 1) =0, s2 (s + 20)  or s3 + 20s2 + 20Ks + 20 = 0 . The Routh array is s3  1  20K  s2  20  20  s1  b  so  1  where b=  20K − 1 . 1  For stability, we require K > 0.05. (b) The desired characteristic polynomial is (s2 + as + b)(s + 5) = s3 + s2 (a + 5) + s(5a + b) + 5b = 0 . Equating coefficients with the actual characteristic equation we can  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  264  CHAPTER 6  The Stability of Linear Feedback Systems  solve for a, b and K, yielding b = 4, a = 15, and K=  5a + b 79 = . 20 20  (c) The remaining two poles are s1 = −14.73 and s2 = −0.27. (d) The step response is shown in Figure DP6.2.  1 0.9 0.8 0.7  y(t)  0.6 0.5 0.4 0.3 0.2 0.1 0  0  2  4  6  8  10 time (sec)  12  14  16  18  FIGURE DP6.2 Mars guided vehicle step response.  DP6.3  (a) The closed-loop characteristic equation is 2τ s3 + (τ + 2)s2 + (K + 1)s + 2K = 0 . The Routh array is s3  2τ  K+1  s2  τ +2  2K  s1  b  so  2K  where b=  (τ + 2)(K + 1) − 4Kτ . (τ + 2)  20  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  265  Design Problems  Examining the first column of the Routh array, we determine that for stability τ > 0, K > 0 and setting b > 0 yields the relationships: (1) K <  τ +2 2 when τ > 3τ − 2 3  (2) K > 0 when 0 < τ ≤  2 . 3  The plot of τ versus K is shown in Figure DP6.3a.  5 4.5 4 3.5  tau  3 2.5 2 1.5 1 0.5  0  STABLE REGION 1  2  3  4  5  6  7  K  FIGURE DP6.3 (a) The plot of τ versus K.  (b) The steady-state error is ess =  A , Kv  where Kv = 2K .  So, ess 1 = . A 2K We require that ess ≤ 0.25A, therefore K≥2. One solution is to select τ = 0.5, then we remain within the stable region. (c) The step response is shown in Figure DP6.3b. The percent overshoot is P.O. = 57%.  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  266  CHAPTER 6  The Stability of Linear Feedback Systems  1.6 P.O. = 56.77 % 1.4 1.2  y(t)  1 0.8 0.6 0.4 0.2 0 0  2  4  6  8  10  12  14  16  18  20  time (sec)  FIGURE DP6.3 CONTINUED: (b) Closed-loop system step response.  DP6.4  (a) The closed-loop characteristic polynomial is s3 + Ks2 + [(2 + m)K − 1]s + 2mK = 0 . The Routh array is s3  1  s2  2K + mK − 1  K  2mK  s1  b  so  2mK  Examining the first column of the Routh array, we see that for stability we require m > 0, K > 0, and b > 0, where b=  (2K + mK − 1)K − 2mK = (2 + m)K − (1 + 2m) > 0 , K  or K>  1 + 2m . 2+m  The plot of K vs m is shown in Figure DP6.4a.  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  267  Design Problems  1.6  1.4 STABLE REGION 1.2  K  1  0.8  0.6  0.4  0  0.5  1  1.5  2  2.5  3  3.5  6  7  4  4.5  5  m  FIGURE DP6.4 (a) The plot of K versus m.  1.8 P.O. = 64.3208 % 1.6  1.4  1.2  y(t)  1  0.8  0.6  0.4  0.2  0  0  1  2  3  4  5 time (sec)  FIGURE DP6.4 CONTINUED: (b) Shuttle attitude control step response.  (b) The steady-state error is ess 1 1 = < 0.10 , = A Kv 2mK  8  9  10  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  268  CHAPTER 6  The Stability of Linear Feedback Systems  or mK > 5. For example, we can select m = 0.5 and K = 2. (c) See Figure DP6.4b for the step response where P.O. = 64.3%. DP6.5  The closed-loop transfer function is T (s) =  s3  +  K . + 20s + K  10s2  The range of K for stability is 0 < K < 200. If we let K = Km /N where Km = 200, then N = 6.25 results in a step response with P.O. = 15.7% and Ts = 1.96 seconds. DP6.6  The closed-loop system is given by   ẋ =   0  1  2 − K1 −2 − K2      x+  0 1    r  The characteristic polynomial is s2 +(2+K2 )s+K1 −2h = 0. Soithe system is stable for K1 > 2 and K2 > −2. Selecting K = 10 2 results in closed-loop eigenvalues at s = −2 ± 2j. The closed-loop step response has a settling time of 2.11 s and a percent overshoot of 4.32%. Im(s)  sin-1 ζ = sin-1 0.69=43.63ο  Re(s) desired region for eigenvalues  ζωn = -1  FIGURE DP6.6 Performance region.  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  269  Design Problems  DP6.7  (a) The inner loop closed-loop transfer function is 20s Y (s) = 3 . 2 U (s) s + 10s + 20s + 20K1 The Routh array is s3  1  20  s2  ω  20K1  s1  200−20K1 10  so  20K1  For stability 0 < K1 < 10. (b) The fastest response (that is, the quickest settling time) occurs when K1 = 2.2 (c) With K1 = 2.2, the closed-loop transfer function is Y (s) 20K2 s = 3 . R(s) s + 10s2 + (20 + 20K2 )s + 44 The Routh array is s3  1  20(K2 + 1)  s2  10  44  s1  200K2 +156 10  so  44  For stability, we require 200K2 + 156 > 0 . Therefore, K2 > −0.78. DP6.8  The closed-loop characteristic equation is s2 + 4KD s + 4(KP + 1) = 0. So, it is possible to find KP and KD to stabilize the system. For example, any KP > 0 and KD > 0 leads to stability. Choosing KP ≥ 9 results in a steady-state tracking error √ less than 0.1 due to a unit step input. Then, the damping ratio ζ = 2/2 is achieved by selecting √ √ 2 KP + 1 KD = . 2  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  270  CHAPTER 6  The Stability of Linear Feedback Systems  Computer Problems CP6.1  The m-file script is shown in Figure CP6.1. ans = -0.6063 + 2.7322i -0.6063 - 2.7322i -1.7874 pa=[1 3 10 14]; roots(pa) pb=[1 8 24 32 16]; roots(pb) pc=[1 0 2 1]; roots(pc)  ans = -2.0004 -2.0000 + 0.0004i -2.0000 - 0.0004i -1.9996 ans = 0.2267 + 1.4677i 0.2267 - 1.4677i -0.4534  FIGURE CP6.1 Computing the polynomial roots with the rootsfunction.  CP6.2  The m-file script is shown in Figure CP6.2. K1=1;K2=2;K3=5; den=[1 2 1]; num1=K1*[1 -1 2];num2=K2*[1 -1 2];num3=K3*[1 -1 2]; sys1 = tf(num1,den); sys2 = tf(num2,den); sys3 = tf(num3,den); sys1_cl=feedback(sys1,[1]); sys2_cl=feedback(sys2,[1]); sys3_cl=feedback(sys3,[1]); p1 = pole(sys1_cl), p2 = pole(sys2_cl), p3 = pole(sys3_cl)  ans = -2.5000e -01 + 1.1990e+00i -2.5000e -01 - 1.1990e+00i  ans =  ans = 2.5000e -01 + 1.3307e+00i 2.5000e -01 - 1.3307e+00i 0 + 1.2910e+00i 0 - 1.2910e+00i  FIGURE CP6.2 K = 1 is stable;K = 2 is marginally stable; and K = 5 is unstable.  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  271  Computer Problems  CP6.3  The closed-loop transfer function and the roots of the characteristic equation are shown in Figure CP6.3. Transfer function: s+1 ---------------------s^3 + 4 s^2 + 7 s + 11  numg=[1 1]; deng=[1 4 6 10]; sysg = tf(numg,deng); sys = feedback(sysg,[1]) r=pole(sys)  r= -2.8946 -0.5527 + 1.8694i -0.5527 - 1.8694i  FIGURE CP6.3 Closed-loop transfer function and roots.  CP6.4  There are no poles in the right half-plane, but the system is unstable since there are multiple poles on the jω-axis at s = ±j and s = ±j (see Figure CP6.4). Step Response From: U(1) 25  20  -2.0000 0.0000 + 1.0000i 0.0000 - 1.0000i -0.0000 + 1.0000i -0.0000 - 1.0000i  10  5 To: Y(1)  ans =  15  Amplitude  num=[1]; den=[1 2 2 4 1 2]; sys = tf(num,den); pole(sys) t = 0:0.1:100; step(sys,t)  0  -5  -10  -15  -20  -25  0  10  20  30  40  50  Time (sec.)  FIGURE CP6.4 Unstable system step response.  60  70  80  90  100  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  272  CHAPTER 6  CP6.5  The Stability of Linear Feedback Systems  The closed-loop system poles for the slow/fast pilots are shown in Figure CP6.5. The maximum allowable time delay is 0.2045seconds. At the maximum allowable time delay, the system has roots on the jω-axis at s = ±2.6j. The slow pilot destabilizes the aircraft.  nume=[-10]; dene=[1 10]; syse = tf(nume,dene); numg=[-1 -6]; deng=[1 3 6 0]; sysg = tf(numg,deng); % % Fast pilot % tau=0.1; tau1=2; K=1; tau2=0.5; nump=-K*[tau1*tau tau-2*tau1 -2]; denp=[tau2*tau tau+2*tau2 2]; sysp = tf(nump,denp); sysa = series(sysp,syse); sysb = series(sysa, sysg); sys = feedback(sysb,[1]); fast_pilot=pole(sys) % % Slow pilot % tau=0.6; tau1=2; K=1; tau2=0.5; nump=-K*[tau1*tau tau-2*tau1 -2]; denp=[tau2*tau tau+2*tau2 2]; sysp = tf(nump,denp); sysa = series(sysp,syse); sysb = series(sysa, sysg); sys = feedback(sysb,[1]); slow_pilot = pole(sys) % % Maximum pilot time delay, tau = 0.2045 sec % tau=0.2045; tau1=2; K=1; tau2=0.5; nump=-K*[tau1*tau tau-2*tau1 -2]; denp=[tau2*tau tau+2*tau2 2]; sysp = tf(nump,denp); sysa = series(sysp,syse); sysb = series(sysa, sysg); sys = feedback(sysb,[1]); max_pilot_delay=pole(sys)  closed-loop system poles  fast_pilot = -19.6267 -10.7712 -3.8885 -0.1697 + 2.7880i -0.1697 - 2.7880i -0.3742  slow_pilot = -9.4526 -4.5228 + 2.2595i -4.5228 - 2.2595i 0.2793 + 2.0314i 0.2793 - 2.0314i -0.3937  max_pilot_delay = -10.0433 + 2.2684i -10.0433 - 2.2684i -4.3153 0.0001 + 2.6040i 0.0001 - 2.6040i -0.3783  FIGURE CP6.5 Closed-loop system poles for an aircraft with a pilot in-the-loop.  CP6.6  The closed-loop transfer function is T (s) =  s3  +  5s2  1 . + (K − 3)s + K + 1  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  273  Computer Problems  Utilizing the Routh-Hurwitz approach, for stability we determine that K>4. When K = 4, the roots of the characteristic equation are s1 = −5 and s2,3 = ±j . The m-file script which generates a plot of the roots of the characteristic equation as a function of K is shown in Figure CP6.6. K=[0:0.1:5]; n=length(K); for i=1:n numg=[1]; deng=[1 5 K(i)-3 K(i)]; sys_o = tf(numg,deng); sys_cl = feedback(sys_o,[1]); p(:,i)=pole(sys_cl); end plot(real(p),imag(p),'x'), grid text(-0.9,0.95,'K=4 -->'); text(-0.2,1.3,'K=5'); text(0,0.2,'K=0') % From a Routh-Hurwitz analysis we find that % minimum K for stability is K=4 Kmax=4; numg=[1]; deng=[1 5 Kmax-3 Kmax]; sysg = tf(numg,deng); sys_cl = feedback(sysg,[1]); pole(sys_cl) 1.5  K=5 1  K=4 -->  0.5 K=0 0  -0.5  -1  -1.5 -6  -5  -4  -3  -2  -1  0  FIGURE CP6.6 Roots of the characteristic equation as a function of K, where 0 < K < 5.  1  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  274  CHAPTER 6  CP6.7  The Stability of Linear Feedback Systems  The characteristic equation is p(s) = s3 + 10s2 + 14s + 12 .  A=[0 1 0;0 0 1;-12 -14 -10]; b=[0;0;12]; c=[1 1 0]; d=[0]; sys = ss(A,b,c,d); % % Part (a) % p=poly(A) % % Part (b) % roots(p) % % Part (c) % step(sys)  p= 1.0000 10.0000 14.0000 12.0000 ans = -8.5225 -0.7387 + 0.9286i -0.7387 - 0.9286i  Step Response 1.4  1.2  Amplitude  1  0.8  0.6  0.4  0.2  0  0  1  2  3  4 5 Time (sec)  6  7  8  9  FIGURE CP6.7 Characteristic equation from the state-space representation using the poly function.  The roots of the characteristic equation are s1 = −8.5225  and s2,3 = −0.7387 ± 0.9286j .  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  275  Computer Problems  The system is stable since all roots of the characteristic equation are in the left half-plane. The unit step response and associated m-file script are shown in Figure CP6.7. CP6.8  The characteristic equation is s3 + 10s2 + 10s + 5K1 = 0 . (a) The Routh array is s3  1  10  s2  10  5K1  s1  100−5K1 10  so  5K1  From the Routh-Hurwitz criterion, we obtain the limits 0 < K1 < 20 for stability. (b) The plot of the pole locations is 0 < K1 < 30 is shown in Figure CP6.8. As seen in Figure CP6.8, when K1 > 20, the pole locations move into the right half-plane. Root Locus 4 3  Imaginary Axi s  2 1  k=20  0 ?-1 ?-2 ?-3 ?-4 ?-12  ?-10  ?-8  ?-6  ?-4  ?-2  0  Real Axi s  FIGURE CP6.8 Pole locations for 0 < K1 < 30.  CP6.9  (a) The characteristic equation is s3 + 2s2 + s + k − 4 = 0 .  2  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  CHAPTER 6  The Stability of Linear Feedback Systems  The Routh array is s3  1  1  s2  2  s1  6−k 2  k−4  so  k−4  For stability, we obtain 4 < k < 6. (b) The pole locations for 0 < k < 10 are shown in Figure CP6.9. We see that for 0 < k < 4 the system is unstable. Similarly, for 6 < k < 10, the system is unstable. Root Locus 2  k=10  pole locations when k=0 1.5 1 0.5  pole location when k=0 increasing k  k=10  0  k=6  inc  rea  ?-1 ?-1.5 ?-2 ?-3  k=4  gk  ?-0.5  sin  Imaginary Axi s  276  ?-2  ?-1 Real Axi s  FIGURE CP6.9 Pole locations for 0 < k < 10.  0  1  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  C H A P T E R  7  The Root Locus Method  Exercises (a) For the characteristic equation 1+K  s(s + 4) =0, + 2s + 2  s2  the root locus is shown in Figure E7.1.  4 3 2 1  Imag Axis  E7.1  0  x  o  o  -1  x  -2 -3 -4 -4  -3  -2  -1  0  1  2  3  4  Real Axis  FIGURE E7.1 s(s+4) Root locus for 1 + K s2 +2s+2 = 0.  277  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  278  CHAPTER 7  The Root Locus Method  (b) The system characteristic equation can be written as (1 + K)s2 + (2 + 4K)s + 2 = 0 . Solving for s yields −(1 + 2K) s= ± (1 + K)  p  (2 + 4K)2 − 8(1 + K) . 2(1 + K)  When (2 + 4K)2 − 8(1 + K) = 0 , then we have two roots at s1,2 = − (1+2K) 1+K . Solving for K yields K = 0.31. (c) When K = 0.31, the roots are s1,2 =  −(1 + 0.62) = −1.24 . (1.31)  (d) When K = 0.31, the characterisitc equation is s2 + 2.472s + 1.528 = (s + 1.24)2 = 0 . Thus, ωn = 1.24 and ζ = 1, the system is critically damped. The settling time is Ts ≈ 4 sec. E7.2  (a) The root locus is shown in Figure E7.2. When K = 6.5, the roots of the characteristic equation are s1,2 = −2.65 ± j1.23  and s3,4 = −0.35 ± j0.8 .  The real part of the dominant root is 8 times smaller than the other two roots. (b) The dominant roots are (s + 0.35 + j0.8)(s + 0.35 − j0.8) = s2 + 0.7s + 0.7625 . From this we determine that ωn = 0.873  and ζ =  0.7 = 0.40 . 2(0.873)  Thus, the settling time is 4 4 = = 11.43 sec . ζωn 0.35 √ 2 The percent overshoot is P.O. = e−πζ/ 1−ζ = 25.4%. Ts =  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  279  Exercises  4 3 * K=6.5  2 *  Imag Axis  1  x *  0  x  -1  x  *  x *  -2 -3 -4 -4  -3  -2  -1  0  1  2  3  4  Real Axis  FIGURE E7.2 Root locus for 1 + K s(s+2)(s12 +4s+5) = 0.  The root locus is shown in Figure E7.3. The roots are s1 = −8.7, s2,3 = −1.3 ± j2.2 when K = 7.35 and ζ = 0.5. 4 zeta=0.5  3 2  o  *  <----- K=7.35  1  Imag Axis  E7.3  0  x  *  x  -1 -2  o *  -3 -4 -10  -8  -6  -4  -2 Real Axis  FIGURE E7.3 2 Root locus for 1 + K ss2+4s+8 = 0. (s+4)  0  2  4  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  280  CHAPTER 7  E7.4  The Root Locus Method  The root locus is shown in Figure E7.4.  2 1.5 1  x  Imag Axis  0.5 0  o  -0.5 -1  x  -1.5 -2 -4  -3.5  -3  -2.5  -2  -1.5  -1  -0.5  0  0.5  1  Real Axis  FIGURE E7.4 s+1 = 0. Root locus for 1 + K s2 +4s+5  The departure angles and entry points are θd = 225o , −225o and σb = −2.4 . E7.5  (a) The root locus is in Figure E7.5. The breakaway points are σb1 = −13.0 , σb2 = −5.89 . (b) The asymptote centroid is σcent = −18 , and φasym = ±90o . (c) The gains are K1 = 1.57 and K2 = 2.14 at the breakaway points.  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  281  Exercises 40  30  20  Imag Axis  10  0  -10  < asymptote  -20  -30  -40  -15  -10  -5  0  5  Real Axis  FIGURE E7.5 s2 +2s+10 Root locus for 1 + K (s4 +38s3 +515s 2 +2950s+6000) = 0.  The system is unstable for K > 75. 10 8  System: sys Gain: 75 Pole: −0.000981 + 8.66i Damping: 0.000113 Overshoot (%): 100 Frequency (rad/sec): 8.66  6 4 Imaginary Axis  E7.6  2 0 −2 −4 −6 −8 −10 −10  −8  −6  FIGURE E7.6 15K Root locus for 1 + K s(s2 +15s+75) = 0.  −4 Real Axis  −2  0  2  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  282  CHAPTER 7  E7.7  The Root Locus Method  The root locus is shown in Figure E7.7. The characteristic equation has 20  15  asymptote −−−>  Imaginary Axis  10  5  0  System: sys Gain: 27.3 Pole: −1.44 + 1.11i Damping: 0.792 Overshoot (%): 1.7 Frequency (rad/sec): 1.81  −5  −10  −15  −20 −25  −20  −15  −10 −5 Real Axis  0  5  10  FIGURE E7.7 s+8 = 0. Root locus for 1 + K s(s+4)(s+6)(s+9)  4 poles and 1 zero. The asymptote angles are φ = +60o , −60o , −180o centered at σcent = −3.7. When K = 27.35 then ζ = 0.8 for the complex roots. E7.8  The characteristic equation is 1+K  (s + 1) =0, s2 (s + 9)  or s3 + 9s2 + Ks + K = 0 . For all the roots to be equal and real, we require (s + r)3 = s3 + 3rs2 + 3r 2 s + r 3 = 0 . Equating terms and solving for K yields K = 27. All three roots are equal at s = −3, when K = 27. The root locus is shown in Figure E7.8.  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  283  Exercises  8 6 4  Imag Axis  2 3 roots at s=-3 0  x  o  x  -2 -4 -6 -8 -15  -10  -5  0  5  Real Axis  FIGURE E7.8 Root locus for 1 + K s2s+1 (s+9) = 0.  E7.9  The characteristic equation is 1+K  1 =0 s(s2 + 2s + 5)  or s3 + 2s2 + 5s + K = 0 . (a) The system has three poles at s = 0 and −1 ± j2. The number of asymptotes is np − nz = 3 centered at σcent = −2/3, and the angles are φasymp at ±60o , 180o .  (b) The angle of departure, θd , is 90o +θd +116.6o = 180o , so θd = −26.6o . (c) The Routh array is  s3  1  5  s2  2  K  s1  b  so  K  where b = 5 − K/2. So, when K = 10 the roots lie on the imaginary  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  284  CHAPTER 7  The Root Locus Method  axis. The auxilary equation is 2s2 + 10 = 0  √ s1,2 = ±j 5 .  which implies  (d) The root locus is shown in Figure E7.9.  4 3 2  x  Imag Axis  1 asymptote ---> 0  x  -1 -2  x  -3 -4 -4  -3  -2  -1  0  1  2  3  4  Real Axis  FIGURE E7.9 1 = 0. Root locus for 1 + K s(s2 +2s+5)  E7.10  (a) The characteristic equation is 1+  K(s + 2) =0. s(s + 1)  Therefore, K=−  (s2 + s) , (s + 2)  and dK s2 + 4s + 2 =− =0. ds (s + 2)2 Solving s2 +4s+2 = 0 yields s = −0.586 and −3.414. Thus, the system  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  285  Exercises  breakaway and entry points are at s = −0.586 and s = −3.414.  (b) The desired characteristic polynomial is  (s + 2 + aj)(s + 2 − aj) = s2 + 4s + 4 + a2 = 0 , where a is not specified. The actual characteristic polynomial is s2 + (1 + K)s + 2K = 0 . Equating coefficients and solving for K yields√K = 3 and a = Thus, when K = 3, the roots are s1,2 = −2 ± 2j.  √  2.  (c) The root locus is shown in Figure E7.10.  2 K=3, s=-2+1.414j  1.5  *  1  Imag Axis  0.5 s=-3.41  0  o  x  s=-0.58  x  -0.5 -1 *  -1.5 -2 -4  -3.5  -3  -2.5  -2  -1.5  -1  -0.5  0  0.5  1  Real Axis  FIGURE E7.10 s+2 Root locus for 1 + K s(s+1) = 0.  E7.11  The root locus is shown in Figure E7.11 for the characteristic equation 1+  K(s + 2.5) =0. (s2 + 2s + 2)(s2 + 4s + 5)  From the root locus we see that we can only achieve ζ = 0.707 when K = 0.  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  286  CHAPTER 7  The Root Locus Method  5 4 3 2  Imag Axis  1 0  x  x  x  x  -2  -1  <---- zeta=0.707 & K=0  o  -1 -2 -3 -4 -5 -5  -4  -3  0  1  2  3  4  5  Real Axis  FIGURE E7.11 Root locus for 1 +  E7.12  K(s+2.5) (s2 +2s+2)(s2 +4s+5)  = 0.  (a) The root locus is shown in Figure E7.12 for the characteristic equation 1+  K(s + 1) =0. s(s2 + 6s + 18)  (b) The roots of the characteristic equation are (i) K = 10: s1,2 = −2.8064 ± 4.2368j and s3 = −0.3872  (ii) K = 20: s1,2 = −2.7134 ± 5.2466j and s3 = −0.5732  (c) The step response performance of the system is summarized in Table E7.12.  K  10  20  Ts (sec)  9.0  5.5  0  0  4.8  2.6  P.O. Tr (sec) TABLE E7.12  System performance when K = 10 and K = 20.  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  287  Exercises  Root Locus 15  10  Imaginary Axis  5  0  −5  −10  −15 −3.5  FIGURE E7.12 Root locus for 1 +  E7.13  −3  −2.5  K(s+1) s(s2 +6s+18)  −2  −1.5 Real Axis  −1  −0.5  0  0.5  = 0.  (a) The characteristic equation is s(s + 1)(s + 3) + 4s + 4z = 0 . Rewriting with z as the parameter of interest yields 1+z  4 =0. s(s + 1)(s + 3) + 4s  The root locus is shown in Figure E7.13a. (b) The root locations for z = 0.6 , 2.0 , and 4.0 are shown in Figure E7.13a. When z = 0.6, we have ζ = 0.76 and ωn = 2.33. Therefore, the predicted step response is P.O. = 2.4%  and Ts = 2.3 sec  (ζ = 0.6) .  When z = 2.0, we have ζ = 0.42 and ωn = 1.79. Therefore, the predicted step response is P.O. = 23%  and  Ts = 5.3 sec  (ζ = 2.0) .  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  CHAPTER 7  The Root Locus Method  4 3 * z=0.6 o z=2.0 + z=4.0  +  2  x  o  *  Imag Axis  1 0  +  o  x  *  -1 x  *  o  -2  +  -3 -4 -4  -3  -2  -1  0  1  2  3  4  Real Axis  FIGURE E7.13 4 = 0. (a) Root locus for 1 + z s(s+1)(s+3)+4s  Finally, when z = 4.0, we have ζ = 0.15 and ωn = 2.19. Therefore, the predicted step response is P.O. = 62% and Ts = 12 sec. (c) The actual step responses are shown in Figure E7.13b.  1.6  1.4  ___ z=0.6 - - - z=2.0  1.2  ..... z=4.0  1 y(t)  288  0.8  0.6  0.4  0.2  0  0  2  4  6  8 time (sec)  10  FIGURE E7.13 CONTINUED: (b) Step Responses for z = 0.6, 2.0, and 4.0.  12  14  16  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  289  Exercises  (a) The root locus is shown in Figure E7.14 for the characteristic equation K(s + 10) =0. s(s + 5)  1+  The breakaway point is sb = −2.93; the entry point is se = −17.1.  10 8 6  K=5, s=-5+5j *  4 2  Imag Axis  E7.14  0  s=-17.1  o  x  -10  -5  s=-2.93  x  -2 -4 -6 -8 -10 -20  -15  0  5  Real Axis  FIGURE E7.14 Root locus for 1 +  K(s+10) s(s+5)  = 0.  √ (b) We desire ζ = 1/ 2 = 0.707. So, the desired characteristic polynomial is 1 s2 + 2 √ ωn s + ωn2 = 0 . 2     Comparing the desired characteristic polynomial to the actual we find the relationships √ 2ωn = 5 + K . ωn2 = 10K and Solving for K and ωn yields K = 5 and ωn = 7.07. The roots are s1,2 = −5 ± j5 when K = 5.  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  290  CHAPTER 7  (a) The characteristic equation 1+K  (s + 10)(s + 2) =0 s3  has the root locus in Figure E7.15.  Root Locus 15 10  Imaginary Axi s  E7.15  The Root Locus Method  5  K=1.67  0 ?-5 ?-10 ?-15 ?-25  ?-20  ?-15  ?-10  -5  0  5  Real Axi s  FIGURE E7.15 Root locus for 1 +  K(s+10)(s+2) s3  = 0.  (b) The Routh array is s3  1  12K  s2  K  20K  s1  b  so  20K  when b = 12K − 20. For stability, we require all elements in the first column to be positive. Therefore, K > 1.67 . (c) When K > 3/4, we have ess = lim sE(s) = lim s s→0  s→0  1 1 s2 · 2 = lim 3 =0. s→0 s + K(s + 1)(s + 3) 1 + GH(s) s  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  291  Exercises  The expansion for e−T s is e−T s = 1 − T s +  (T s)2 − ... 2!  If (T s) << 1, then e−T s ≈ 1 − T s =  a + bs , c + ds  where a, b, c and d are constants to be determined. Using long division,  40 30 K=21  20  *  10  Imag Axis  E7.16  0  x  x  o  -10 -20 -30 -40 -40  -30  -20  -10  0  10  20  30  40  50  60  Real Axis  FIGURE E7.16 Root locus for 1 +  K(20−s) (s+1)(20+s)  = 0.  we expand (a + bs)/(c + ds) and match as many coefficients as possible. In this case, we determine that a = c = (2/T ) and also that b = −d = −1. In this case, with T = 0.1, we have e−T s =  20 − s −(s − 20) = . 20 + s (s + 20)  So, the characteristic equation is 1+  −K(s − 20) , (s + 1)(s + 20)  and the root locus is shown in Figure E7.16. Using a Routh-Hurwitz  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  292  CHAPTER 7  The Root Locus Method  analysis with the characteristic polynomial s2 + (21 − K)s + 20 + 20K = 0 , we determine that the system is stable for −1 < K < 21. (a) The root locus is in Figure E7.17a.  2 1.5 1 0.5  Imag Axis  E7.17  0  x  x  -0.5 -1 -1.5 -2 -2  -1.5  -1  -0.5  0  0.5  1  1.5  2  Real Axis  FIGURE E7.17 (a) Root locus for 1 +  K s(s−1)  = 0.  The root locus is always in the right half-plane; the system is unstable for K > 0. (b) The characteristic equation is 1+  K(s + 2) =0, s(s − 1)(s + 20)  and the root locus is shown in Figure E7.17b. The system is stable for K > 22.3 and when K = 22.3, the roots are s1,2 = ±j1.53  and s3 = −19 .  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  293  Exercises  10 8 6 4  Imag Axis  2  *  0  x  o  K=22.3  x x  -2 -4 -6 -8 -10 -30  -25  -20  -15  -10  -5  0  5  10  Real Axis  FIGURE E7.17 CONTINUED: (b) Root locus for 1 +  = 0.  The root locus is shown in Figure E7.18.  6  4  2  Imag Axis  E7.18  K(s+2) s(s+20)(s−1)  x  +  0  x  +  K=8.15  x +  x  +  -2  -4  -6 -6  -4  -2  0  2  4  6  Real Axis  FIGURE E7.18 Root locus for 1 +  K s(s+3)(s2 +2s+2)  = 0.  When K = 8.15, the roots are s1,2 = ±j1.095 and s3,4 = −2.5 ± j0.74.  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  294  CHAPTER 7  E7.19  The Root Locus Method  The characteristic equation is 1+  K =0, + 6s + 64)  3)(s2  s(s +  and the root locus is shown in Figure E7.19. When K = 1292.5, the roots are s1,2 = ±j4.62  and s3,4 = −4.49 ± j6.36 .  15  10 x +  Imag Axis  5  +  0  x  K=1292.5  x  +  -5 + x  -10  -15 -15  -10  -5  0  5  10  15  Real Axis  FIGURE E7.19 Root locus for 1 +  E7.20  K s(s+3)(s2 +6s+64)  = 0.  The characteristic equation is 1+  K(s + 1) =0, s(s − 2)(s + 6)  and the root locus is shown in Figure E7.20. The system is stable for K > 16 . The maximum damping ratio of the stable complex roots is ζ = 0.25 .  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  295  Exercises 20  15  10  Imaginary Axis  5  ζmax = 0.25 0                    0  1  2  Real Axis  FIGURE E7.20 Root locus for 1 +  = 0.  The gain is K = 10.8 when the complex roots have ζ = 0.66.  10  5 K=10.8  Imag Axis  E7.21  K(s+1) s(s−2)(s+6)  +  0  x  +  x o  +  x  -5  -10 -10  -5  0 Real Axis  FIGURE E7.21 Root locus for 1 +  Ks s3 +5s2 +10  = 0.  5  10  3  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  296  CHAPTER 7  E7.22  The Root Locus Method  The root locus is shown in Figure E7.22. The characteristic equation is 1+  K(s2 + 18)(s + 2) =0. (s2 − 2)(s + 12)  Root Locus 5 4 3  Imaginary Axis  2 1 0 −1 −2 −3 −4 −5 −14  FIGURE E7.22 Root locus for 1 +  E7.23  −12  −10  K(s2 +18)(s+2) (s2 −2)(s+12)  −8  −6 Real Axis  −4  −2  0  = 0.  The characteristic equation is 5s2 + as + 4 = 0 , which can rewritten as 1+  as =0. +4  5s2  The roots locus (with a as the parameter) is shown in Figure E7.23.  2  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  297  Exercises  1.5 1  x  Imag Axis  0.5 0  o  -0.5 x  -1 -1.5 -1.5  -1  -0.5  0  0.5  1  1.5  Real Axis  FIGURE E7.23 Root locus for 1 +  E7.24  as 5s2 +4  = 0.  The transfer function is G(s) = C(sI − A)−1 B + D   = [ 1 0 ] =  s2  s  −1  4 s+k  1 . + ks + 4  −1      0 1     Therefore, the characteristic equation is s2 + ks + 4 = 0 , or 1+k  s2  s =0. +4  The root locus for 0 < k < ∞ is shown in Figure E7.24. The closed-loop system is stable for all 0 < k < ∞.  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  298  CHAPTER 7  The Root Locus Method 2.5 2 1.5  Imaginary Axis  1 0.5 0   5   5   5    5     5     5     5  0  0.5  Real Axis  FIGURE E7.24 Root locus for 1 + k s2s+4 = 0.  The characteristic equation is 1+K  10 =0. s(s + 25)  The root locus shown in Figure E7.25 is stable for all 0 < K < ∞. 15  10  5 Imaginary Axis  E7.25  0              Real Axis  FIGURE E7.25 10 Root locus for 1 + K s(s+25) = 0.      0  5  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  299  Exercises  E7.26  The characteristic polynomial is   det   or  s  −1  s+K −3 s+K +2 1+K    =0  s+1 =0. s2 + 2s − 3  The root locus shown in Figure E7.26 is stable for all 0 < K < 3.  Root Locus 0.8  0.6  Imaginary Axis  0.4  0.2  0  −0.2  −0.4  −0.6  −0.8 −12  −10  −8  −6 −4 Real Axis  −2  0  2  FIGURE E7.26 s+1 Root locus for 1 + K s2 +2s−3 = 0.  E7.27  The characteristic equation is 1+p  s2  s =0. + 4s + 40  The root locus shown in Figure E7.27 is stable for all 0 < p < ∞.  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  300  CHAPTER 7  The Root Locus Method 8  6  Imaginary Axis  4  2  0                  0  2  Real Axis  FIGURE E7.27 s = 0. Root locus for 1 + p s2 +4s+40  The characteristic equation is 1+K  s(s2  s−1 =0. + 2s + 2)  The system is stable for −1.33 < K < 0. 1.5  1  0.5 Imaginary Axis  E7.28  0  #5    #5       !  " Real Axis  FIGURE E7.28 Root locus for 1 + K s(s2s−1 = 0. +2s+2)  0  2  4  6  8  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  301  Problems  Problems P7.1 Root Locus 30  20  Imaginary Axis  10  0  −10  −20  −30 −50  −40  −30  −20 −10 Real Axis  0  10  20  Root Locus 5 4 3  Imaginary Axis  2 1 0 −1 −2 −3 −4 −5 −7  FIGURE P7.1 (a) Root locus for 1 +  −6  −5  K s(s+10)(s+8)  −4  −3 −2 Real Axis  = 0, and (b) 1 +  −1  0  1  K (s2 +2s+2)(s+1)  2  = 0.  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  302  CHAPTER 7  The Root Locus Method  Root Locus 40  30  Imaginary Axis  20  10  0  −10  −20  −30  −40 −12  −10  −8  −6 −4 Real Axis  −2  0  2  Root Locus 4  3  Imaginary Axis  2  1  0  −1  −2  −3  −4 −4  −3.5  −3  −2.5  FIGURE P7.1 CONTINUED: (c) Root locus for 1 +  P7.2  −2 −1.5 Real Axis  K(s+5) s(s+2)(s+7)  −1  −0.5  = 0, and (d)1 +  0  0.5  K(s2 +4s+8) s2 (s+7)  = 0.  The root locus is shown in Figure P7.2 for the characteristic equation 1+  10Kv (s + 10) =0. s(s + 1)(s + 100)  The damping ratio is ζ = 0.6 when Kv = 0.8, 135 and 648. The roots of the characteristic equation are: (a) Kv = 0.8 : s1 = −99.9, s2,3 = −0.54 ± j0.71  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  303  Problems  (b) Kv = 135 : s1 = −85.9, s2,3 = −7.5 ± j10  (c) Kv = 648 : s1 = −11.7, s2,3 = −44.6 ± j59.5  30  20  Imag Axis  10  0  x  o  xx  -10  -20  -30  -100  -80  -60  -40  -20  Real Axis  FIGURE P7.2 Root locus for 1 +  P7.3  10Kv (s+10) s(s+1)(s+100)  = 0.  (a) The breakaway point is s = −0.88 at K = 4.06. (b) The characteristic equation can be written as  s(s + 2)(s + 5) + K = 0 . The Routh array is s3  1  10  s2  7  K  s1  b  0  so  K  where b=  70 − K . 7  0  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  304  CHAPTER 7  The Root Locus Method  √ When K = 70, the system has roots on jω-axis at s = ±j 10. (c) When K = 6, the roots are s1,2 = −0.83 ± j0.66, s3 = −5.34. (d) The characteristic equation 1+  K =0 s(s + 2)(s + 5)  has the root locus shown in Figure P7.3. 10 8 6 4  Imag Axis  2 0 -2 -4 -6 -8 -10 -10  -8  FIGURE P7.3 Root locus for 1 +  P7.4  -6  -4  K s(s+2)(s+5)  -2  0 Real Axis  2  4  6  8  10  = 0.  The characteristic equation for the large antenna is 1 + G1 G(s) = 1 +  100ka =0, (0.1s + 1)(s2 + 14.4s + 100)  or 1+  1000ka =0. (s + 10)(s2 + 14.4s + 100)  The root locus is shown in Figure P7.4. Using Routh's criteria, we find that the system is stable for −1 < ka < 4.83 .  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  305  Problems  20 *  15  <-- K=4.827  10 x  Imag Axis  5 0  x  -5 x  -10 -15  *  -20 -20  -15  -10  -5  0  5  10  15  20  Real Axis  FIGURE P7.4 Root locus for 1 +  1000ka (s2 +14.14s+100)(s+10)  = 0.  When ka = 4.83, we have s1,2 = ±j15.53. P7.5  (a) The characteristic equation for hands-off control is 1+  25K2 (s + 0.03)(s + 1) =0. (s + 0.4)(s2 − 0.36s + 0.16)(s + 9)  The root locus is shown in Figure P7.5a. The damping ratio is ζ = 0.707 when K2 = 1.6 or K2 = 0.74. (b) The transfer function from Td (s) to Y (s) is Y (s) =  G2 (s)Td (s) , 1 + G2 (s)Gf (s)  where Gf (s) =  K2 (s + 1) . s+9  Using the final value theorem, we determine that yss = lim s s→0  G2 (s) 1 11.7   = 3.8 , = 1 + G2 (s)Gf (s) s 1 + 11.7 K92  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  CHAPTER 7  The Root Locus Method  5 4  <-- K=1.6  3 2 <-- K=0.74  Imag Axis  1  x  0  x  o  x o x  -1 -2 -3 -4 -5 -10  -8  -6  -4  -2  0  2  4  Real Axis  FIGURE P7.5 (a) Root locus for 1 +  25K2 (s+0.03)(s+1) (s+9)(s2 −0.36s+0.16)(s+0.4)  = 0.  20 15 10 5  Imag Axis  306  x  0  x  o  x  xo  x  -5 -10 -15 -20 -20  -15  -10  -5  0  5  10  15  20  Real Axis  FIGURE P7.5 CONTINUED: (b) Root locus for 1 +  25K1 (s+0.03)(s+9) (s+0.045)(s2 +12s+1)(s+1.33)(s2 +7.66s+29.78)  = 0.  where we have selected K2 = 1.6. For K2 = 0.74, we find that yss = 5.96.  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  307  Problems  (c) The closed-loop characteristic equation with the pilot loop added is 1+  25K1 (s + 0.03)(s + 9) =0. (s + 0.045)(s + 1.33)(s2 + 7.66s + 29.78)(s2 + 12s + 1)  The root locus is shown in Figure P7.5b. (d) Using K1 = 2, we determine that ess = 0.44 . (a) The characteristic equation is 1+  K(s + 0.20)(s2 + 4s + 6.25) =0. (s + 0.9)(s − 0.6)(s − 0.1)(s + 4)  The root locus is shown in Figure P7.6.  4 zeta=0.5 zeta*wn=-1/3  3 2 o  K=4 -->  1  Imag Axis  P7.6  0  x  x  o  x  x  -1 o  -2 -3 -4 -6  -5  -4  -3  -2  -1  0  1  2  Real Axis  FIGURE P7.6 Root locus for 1 +  K(s+0.2)(s2 +4s+6.25) (s+0.9)(s−0.6)(s−0.1)(s+4)  = 0.  (b) For Ts < 12 sec, we require ζωn > 1/3. Also, we want ζ > 0.5. So, we seek roots for a stable system with ζωn > 1/3 and ζ > 0.5. This occurs when K > 4.  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  308  CHAPTER 7  (a) The characteristic equation for the speed control system is 1+  K =0, (s + 4)2 (s + δ)  where K=  0.004 R  and  δ=  0.75 = 0.0001875 . 4000  The root locus is shown in Figure P7.7. At ζ = 0.6, we have K = 19.1,  6  4  2 <-- K=19.1  Imag Axis  P7.7  The Root Locus Method  0  x  x  -2  -4  -6 -6  -5  -4  -3  -2  -1  0  1  2  3  Real Axis  FIGURE P7.7 Root locus for 1 +  K (s+4)2 (s+1.875e−04)  = 0.  therefore R = 0.00021 . When K = 19.1 the roots are s1,2 = −1.1 ± j1.43  and  s3 = −5.80 .  (b) The steady-state error is lim s∆ω(s) = lim s  s→0  s→0  (0.25s + 1)2 ∆L(s) (0.25s + 1)2 (Js + b) + 1/R  4  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  309  Problems  =  1 ∆L ≈ ∆LR , b + 1/R  when R < 0.1. (a) The characteristic equation for the speed control system with the hydroturbine is 1+  K(−s + 1) =0, (s + 4)(s + 2)(s + δ)  where K=  0.002 R  and  δ=  0.75 = 0.0001875 . 4000  The root locus is shown in Figure P7.8. At ζ = 0.6, we have K = 2.85,  2 1.5 1 K=2.85 -->  0.5  Imag Axis  P7.8  0  x  x  x  o  0  1  -0.5 -1 -1.5 -2 -6  -5  -4  -3  -2  -1  2  Real Axis  FIGURE P7.8 Root locus for 1 +  K(−s+1) (s+4)(s+2)(s+δ)  = 0.  therefore R = 0.0007 . When K = 2.85 the roots are −0.45 ± j0.60, and -5.1.  3  4  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  310  CHAPTER 7  The Root Locus Method  (b) The steady-state error is lim s∆ω(s) = lim s  s→0  s→0  =  (0.25s + 1)(0.5s + 1) ∆L(s) (0.25s + 1)(0.5s + 1)(Js + f ) + (−s + 1)/R  1 ∆L ≈ ∆LR , f + 1/R  when R < 0.1. The characteristic equation is 1+K  (s + 0.5)(s + 0.1)(s2 + 2s + 289) =0 s(s + 30)2 (s − 0.4)(s + 0.8)(s2 + 1.45s + 361)  where K = K1 K2 . The root locus is shown in Figure P7.9. When K = 4000 , the roots are s1,2 = −0.82 ± j19.4 50  40  30  20  10 Imag Axis  P7.9  0  -10  -20  -30  -40  -50  -35  FIGURE P7.9 Root locus for 1 +  -30  -25  -20  -15 Real Axis  -10  K(s+0.5)(s+0.1)(s2 +2s+289) s(s+30)2 (s−0.4)(s+0.8)(s2 +1.45s+361)  -5  = 0.  0  5  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  311  Problems  s3 s4 s5 s6 s7  (a) The characteristic equation is 1+  K1 K2 (s + 2)2 =0. (s + 10)(s + 100)(s2 + 1.5s + 6.25)  The root locus is shown in Figure P7.10.  10 8 6 4 x  2  Imag Axis  P7.10  = −39.8 = −14.9 = −5.0 = −0.38 = −0.14 .  0  x  x  -2  o  x  -4 -6 -8 -10 -120  -100  -80  -60  -40  Real Axis  FIGURE P7.10 Root locus for 1 +  K1 K2 (s+2)2 (s+10)(s+100)(s2 +1.5s+6.25)  = 0.  (b) The gain K1 K2 = 1620 when ζ = 0.707. Therefore, K2 = 81000 ,  -20  0  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  312  CHAPTER 7  The Root Locus Method  since K1 = 0.02 at medium weight cruise condition. (c) At lightweight cruise condition K1 = 0.2 . Using K2 = 81000, we find the roots are s1,2 = −54 ± j119 s3,4 = −2 ± j0.6 . The roots s3,4 become negligible and the roots at s1,2 become highly oscillatory. Hence, in this case ζ = 0.41 . (a) The closed-loop characteristic equation is 1+  20Ka (s2 + s + 0.02) =0, s(s + 1)2 (s2 + 2s + 0.8)  where K2 = 10 . Then, the root locus is shown in Figure P7.11a.  3  2  1 Ka=0.035 -->  Imag Axis  P7.11  0  x  xo  x  ox  -1  -2  -3 -3  -2  -1  0  1  2  Real Axis  FIGURE P7.11 20s2 +20s+0.4 (a) Root locus for 1 + Ka s(s+1) 2 (s2 +2s+0.8) = 0, where K2 = 10.  3  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  313  Problems  (b) When Ka < 0.035 , all the roots have a damping greater than or equal to 0.60. (c) Select Ka = 0.035 . Then, the characteristic equation with K2 as the parameter is 1 + K2  0.07(s2 + s) =0. s5 + 4s4 + 5.8s3 + 3.6s2 + 0.8s + 0.014  The root locus is shown in Figure P7.11b.  3 Ka=0.035 2  Imag Axis  1  0  x  x o x  x  xo  -1  -2  -3 -3  -2  -1  0  1  2  3  Real Axis  FIGURE P7.11 0.07s(s+1) CONTINUED: (b) Root locus for 1+K2 s(s+1)2 (s2 +2s+0.8)+0.014 = 0, where Ka = 0.035.  P7.12  (a) The closed-loop transfer function is T (s) =  1.8s2 (s  Ka Km (s + 25)(s + 15) . + 2) + Ka Km (s + 25)(s + 15) + 1.6Km s(s + 2)  So, with E(s) = R(s) − Y (s), we have E(s) = (1 − T (s))R(s) and ess = lim sE(s) = 1 − T (0) = 0 . s→0  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  CHAPTER 7  The Root Locus Method  Therefore, when the system is stable, it has zero steady-state error. (b) The characteristic equation is s3 + (3.6 + Ka )s2 + (3.2 + 40Ka )s + 375Ka . The Routh array is s3  1  3.2 + 40Ka  s2  3.6 + Ka  375Ka  s1  b  so  375K  Solving for b > 0 leads to 0 < Ka < 0.05 or Ka > 5.64 for stability. (c) The characteristic equation can be written as 1+  Ka (s + 25)(s + 15) =0. s(s + 2)(s + 1.6)  The root locus is shown in Figure P7.12. (d) When K > 40 ,  40 30 20 10  Imag Axis  314  0  o  o  xx x  -10 -20 -30 -40 -70  -60  -50  -40  -30  -20  Real Axis  FIGURE P7.12 (s+25)(s+15) Root locus for 1 + Ka s(s+2)(s+1.6) = 0, where Km = 1.8.  -10  0  10  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  315  Problems  the roots are s1 = −123  and  s2,3 = −15.6 ± j31.2 .  From the step response we find P.O. = 5% Tp = 0.67 sec Ts = 0.25 sec . (a) The characteristic equation is 1+  s(s +  3)(s2  K =0. + 4s + 7.84)  The root locus is shown in Figure P7.13. The breakaway point is s = −1.09 at K = 9.72.  (b) When K = 13.5, the roots are  s1,2 = −0.84 ± j0.84 s3,4 = −2.66 ± j1.55 . 6  4  2  x +  Imag Axis  P7.13  +  0  x  x + +  -2  x  -4  -6 -6  -4  -2  0 Real Axis  FIGURE P7.13 Root locus for 1 +  K s(s+3)(s2 +4s+7.84)  = 0.  2  4  6  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  316  CHAPTER 7  The Root Locus Method  (c) The roots s = −0.84 ± j0.84 are dominant roots. (d) For the dominant roots, we determine that ζ = 0.7 and ωn = 1.19. Therefore, the settling time is Ts =  sec .  The characteristic equation is 1+  K(s + 2.5)(s + 3.2) =0. + 1)(s + 10)(s + 30)  s2 (s  The root locus is shown in Figure P7.14. When K = 559.3, the roots are s1 = −30.75  s2 = −8.48  s3 = −1.78  s4,5 = ±j3.11 .  s3 = −2.21  s4,5 = ±j10.23 .  When K = 4321, the roots are s1 = −34.45  s2 = −4.35  The crossover points are s = ±j3.11  and  s = ±j10.23 .  25 20 15 10 Imaginary Axis  P7.14  4 = 4.8 ζωn  5 0  $& $)' $)& $(' $(& $%&  $%'  $(&  $('  $)& Real Axis  FIGURE P7.14 (s+2.5)(s+3.5) Root locus for 1 + K s2 (s+1)(s+10)(s+30) = 0.  $)'  $&  0  5  10  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  317  Problems  Therefore, the system is stable for 559.3 < K < 4321 . The characteristic equation is 1+  K(s2 + 30s + 625) . s(s + 20)(s2 + 20s + 200)(s2 + 60s + 3400)  The root locus is shown in Figure P7.15. When K = 30000, the roots are s1 = −18.5  s2 = −1.69  s3,4 = −9.8±j8.9  s5,6 = −30.1±j49.9.  The real root near the origin dominates, and the step response is overdamped.  100 80 60 x  40 20  Imag Axis  P7.15  o x  0  x  x x  -20  o  -40 x  -60 -80 -100 -100  -80  -60  -40  -20  0  20  40  Real Axis  FIGURE P7.15 s2 +30s+625 Root locus for 1 + K s(s+20)(s2 +20s+200)(s 2 +60s+3400) = 0.  60  80  100  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  318  CHAPTER 7  (a) Let τ = 0. Then, first reduce the motor and rolls to an equivalent G(s) as follows: G(s) =  1  0.25 s(s+1) 0.25 + s(s+1)  =  0.25 0.25 = . s(s + 1) + 0.25 (s + 0.5)2  The loop transfer function is then L(s) =  2(s + 0.5)Ka (0.25) 0.5Ka = . 2 2 s(s + 1) (s + 0.5) s(s + 1)2 (s + 0.5)  The characteristic equation is 1 + Ka  0.5 =0. s(s + 1)2 (s + 0.5)  The root locus is shown in Figure P7.16.  2 1.5 1 0.5  Imag Axis  P7.16  The Root Locus Method  + +  0  x  x  x +  +  -0.5 -1 -1.5 -2 -2  -1.5  -1  -0.5  0  0.5  1  1.5  Real Axis  FIGURE P7.16 Root locus for 1 +  0.5Ka s(s+1)2 (s+0.5)  = 0.  (b) When K = 0.123, the roots of the characteristic equation are s1,2 = −1.1 ± j0.27 s3,4 = −0.15 ± j0.15 .  2  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  319  Problems  The roots at s = −0.15 ± j0.15 have a damping ratio of ζ = 0.707.  (c) When τ becomes nonnegligible, the root locus will have an additional pole, and the root locus will change accordingly. The characteristic equation is 2 (M1 s2 + bs + k1 + k12 )(M2 s2 + k12 ) − k12 =0. 2 is negligible If we let M1 = k1 = b = 1, and assume k12 < 1 so that k12 and k1 + k12 ≈ k1 , then the characteristic equation is  (s2 + s + 1)(M2 s2 + k12 ) = 0  or  1+  k =0, s2  where k=  k12 . M2  The root locus is shown in Figure P7.17. All the roots lie on the jω axis. If we select s  k12 = ωo , M2  then we cancel the vibration.  3  2  root locus -->  1  Imag Axis  P7.17  0  x  -1  -2  -3 -3  -2.5  -2  -1.5  -1 Real Axis  FIGURE P7.17 Root locus for 1 +  k s2  = 0.  -0.5  0  0.5  1  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  320  CHAPTER 7  The characteristic equation is βs3 + (1 + 2β)s2 + (2 + 4α)s + 4 = 0 . When β = 0 we have 1+  s2  4αs =0. + 2s + 4  The root locus for β = 0 is shown in Figure P7.18.  3 <-- zeta=0.6  beta=0  2 x *  1  Imag Axis  P7.18  The Root Locus Method  0  o  -1 * x  -2  -3 -3  -2  -1  0  1  Real Axis  FIGURE P7.18 4s = 0, where β = 0. Root locus for 1 + α s2 +2s+4  For α = 0.3, the poles are s = −1.6 ± j1.2 . Then, we have 1+  β(s + 2)s2 =0. s2 + (2 + 4α)s + 4  When β = 0.121 s1,2 = −1.51 ± j1.51 s3 = −7.24 .  2  3  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  321  Problems  Thus, ζ = 0.707  and ζωn = 1.5 .  So, the performance specs are met. Also, Gc (s) = P7.19  0.3s + 1 2.48(s + 3.33) = . 0.121s + 1 (s + 8.26)  The characteristic equation is 1+  Ka (s2 + 4s + 100) =0. s(s + 2)(s + 6)  The root locus is shown in Figure P7.19. 10 8 6  Imaginary Axis  4 2 0 −2 −4 −6 −8 −10 −10  −8  −6  −4 Real Axis  −2  0  FIGURE P7.19 s2 +4s+100 Root locus for 1 + Ka s(s+2)(s+6) = 0.  When Ka = 0.094;, the roots are s1,2 = −0.85 ± j0.85 s3 = −6.38 . Thus, the complex roots have a damping ratio of ζ = 0.707. P7.20  The characteristic equation is s3 + (2 +  1 2 2 4 )s + ( + K)s + = 0 , β β β  2  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  322  CHAPTER 7  The Root Locus Method  where K=  4α β  α = 0.3  β = 0.121 .  The root sensitivity to changes in K is found to be r1 ∼ SK =  ∆r1 = 1.186 −149.75o . ∆K/K  The root sensitivity to changes in the pole at s = −2 is found to be r1 ∼ ∆r1 S∆ = 1.656 −137o , = ∆/2  (a) Let the pole be (s + 4 + ∆) and neglect ∆2 terms. Then, the characteristic equation is 1+∆  2s2 + (8 + 2δ)s + 8δ =0 s3 + (8 + δ)s2 + (16 + 8δ)s + 16δ + K  where δ = 0.000788 and K = 19.1. 3  2 x  1  Imag Axis  P7.21  where the pole is s + 2 + ∆.  0  x  o  o  -1 x  -2  -3 -10  -8  -6  -4  -2  0  2  Real Axis  FIGURE P7.21 2s2 +(8+2δ)s+8δ Root locus for 1 + ∆ s3 +(8+δ)s2 +(16+8δ)s+16δ+K = 0, (δ = 0.000788 and K = 19.1).  The root sensitivity is determined to be r1 ∼ ∆r1 S∆ = 3.3146 −132o . = ∆/4  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  323  Problems  (b) Let R = Ro + ∆R, where R = 0.00021. Then, r1 ∼ ∆r1 SR = 1.31 6 −107o . = ∆R/R  P7.22  The characteristic equation is s3 + 2s2 + s + K , where K = 0.24 for ζ = 0.707. The root sensitivity to changes in the pole at s = −1 is found to be r1 ∼ ∆r1 S∆ = 0.956 −126o , = ∆  where the pole is s + 1 + ∆. P7.23  The characteristic equation is s3 + 5s2 + (6 + K)s + K , where K = 6.3 for ζ = 0.707. The root sensitivity to changes in the pole at s = −2 is found to be r1 ∼ ∆r1 S∆ = 1.256 −169.4o , = ∆/2  where the pole is s + 2 + ∆. The root sensitivity to changes in the zero at s = −1 is found to be r1 ∼ ∆r1 S∆ = 0.556 34.4o , = ∆  where the zero is s + 1 + ∆. P7.24  The root locus for each of the four cases shown is shown in Figure P7.24. The four open-loop transfer functions are (a) KF (s) =  s+8 + + 296s3 + 1170s2 + 1575s 1 s2 + 6s + 6.75 KF (s) = 6 (d) KF (s) = 3 5 4 s + 2s + s s + 5s2 + 4s  (b) KF (s) = (c)  s2 + 7s + 8.25 s3 + 6s2 + 5s s5  30s4  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  324  CHAPTER 7  The Root Locus Method  (a)  5  (b)  10  0  ox  Imag Axis  Imag Axis  5 ox x  0  x  ox x x  x  -5 -5 -10  -5 Real Axis  -10 -20  0  (c)  2  5  -10 Real Axis  0  (d)  0  x  Imag Axis  Imag Axis  1 x  0  ox  ox x  -1 -2 -2  0  -5 -10  2  Real Axis  -5  0  Real Axis  FIGURE P7.24 Root locus for the four cases.  P7.25  The characteristic equation is 1 + KGc (s)G(s) = 0 , therefore, KGc (s)G(s) = −1 . Squaring both sides yields K 2 G2c (s)G2 (s) = 1 and 1 − K 2 G2c (s)G2 (s) = 0 . The root locus with 0 < K 2 < ∞ is shown in Figure P7.25. The value of K 2 for which the locus crosses the imaginary axis is K 2 = 2/3 , p  therefore K = 2/3 = 0.8165 corresponds to the jω-axis crossing (at s = 0). You can check that 1 + KGc (s)G(s) = 0 for K = 0.8165 and s = 0.  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  325  Problems  3  2  Imag Axis  1  0  -1  -2  -3 -3  -2  -1  0 Real Axis  1  2  3  FIGURE P7.25 Root locus for the equation 1 − K 2 G2c (s)G2 (s) = 0.  P7.26  (a) The characteristic equation is 1+  K(s + 2)2 =0. s(s2 + 1)(s + 8)  The root locus is shown in Figure P7.26. (b) Using Routh's criteria, we determine that K > 14 for stability. (c) From the Routh array, we determine that for K = 14, we have two purely imaginary poles at √ s = ±j 8 . (d) When K > 50, the real part of the complex roots is approximately equal to the real part of the two real roots and therefore the complex roots are not dominant roots.  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  326  CHAPTER 7  The Root Locus Method  15  10  Imag Axis  5  0  -5  -10  -15 -15  -10  -5  0 Real Axis  5  10  15  FIGURE P7.26 (s+2)2 Root locus for 1 + K s(s2 +1)(s+8) = 0.  P7.27  The characteristic equation is 1+  K(s2 + 0.1) =0. s(s2 + 2)  The root locus is shown in Figure P7.27a. The locus enters the axis at s = −1.26 and leaves the axis at s = −0.36 . Define p(s) = K =  −(s3 + 2s) . s2 + 0.1  Then, a plot of p(s) vs s is shown in Figure P7.27b, where it can be seen that p(s) has two inflection points at s = −1.28  and s = −0.36 .  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  327  Problems  Root Locus 2 1.5  Imaginary Axis  1 0.5 0 −0.5 −1 −1.5 −2 −3  −2.5  −2  −1.5 −1 Real Axis  −0.5  0  0.5  3.5  3  2.5  p(s)  2  1.5  1  0.5  0 −2  −1.8  −1.6  −1.4  −1.2  −1 s  −0.8  −0.6  −0.4  −0.2  0  FIGURE P7.27 s2 +0.1 s3 +2s (a) Root locus for 1 + K s(s 2 +2) = 0. (b) Plot of p(s) = − s2 +0.1 versus s.  P7.28  The characteristic equation is 1 + L(s) = 1 +  K(s2 + 12s + 20) =0. s3 + 10s2 + 25s  The root locus is shown in Figure P7.28. The breakaway point is s = −5.0  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  328  CHAPTER 7  The Root Locus Method  6  4  Imag Axis  2  0  -2  -4  -6 -20  -15  -10  -5  0  Real Axis  FIGURE P7.28 (s2 +12s+20) Root locus for 1 + K s3 +10s2 +25s = 0.  and the entry point is s = −15.6. When K = 2, the roots are s1 = −1.07 s2,3 = −5.46 ± j2.75 . When K = 2, the roots are s1 = −1.07 s2,3 = −4.36 ± j1.68 . The predicted step response when K = 2 is Ts = 9 sec and P O ≈ 0%. P7.29  The characteristic equation is 1+K  s2 + 10s + 30 =0. s2 (s + 10)  The root locus is shown in Figure P7.29. When ζ = 0.707, the necessary gain is K = 16. The corresponding roots are s1 = −18.87 and s2,3 = −3.56 ± j3.56.  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  329  Problems  Root Locus 4 System: sys Gain: 16 Pole: −3.56 + 3.57i Damping: 0.707 Overshoot (%): 4.34 Frequency (rad/sec): 5.04  3  Imaginary Axis  2 1 0 −1 −2 −3 −4 −20  −15  −10  −5  0  5  Real Axis  FIGURE P7.29 2 = 0. Root locus for 1 + K ss+10s+30 2 (s+10)  P7.30  The transfer function is Z(s) =  LCRs2 + Ls Rs2 + s = . LCs2 + CRs + 1 s2 + Rs + 1  So, R r1 = − + 2  !1  R2 −1 4  2  .  Thus, the nominal r1o = − 21 . Simultaneously, R r2 = − − 2  !1  R2 −1 4  2  .  Thus, the nominal r2o = −2. We see that there is a difference by a factor of 4. Also, ri SR  ∂r1 = ∂R  Ro  Ro Ro2 · Ro = − + 2 4  !− 1  Ro2 −1 4  2  =  5 , 6  where Ro = 2.5. And r2 SR  ∂r2 = ∂R  Ro Ro2 Ro = − − 2 4 Ro  !− 1  Ro2 −1 4  2  =  −10 . 3  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  330  CHAPTER 7  The Root Locus Method  r2 r1 So, the magnitude of |SR | = 4|SR |.  P7.31  The characteristic equation is 1+K  s(s +  0.16)(s2  s+4 =0. + 14.6s + 148.999)  The root locus is shown in Figure P7.31. When K = 1350, the roots are  20 15 10  (+) K=326 -->  x  *  +  <-- K=1350 (*)  Imag Axis  5 + *  0  o *  xx +  -5 + x  -10  *  -15 -20 -20  -15  -10  -5  0  5  10  Real Axis  FIGURE P7.31 s+4 Root locus for 1 + K s(s+0.16)(s2 +14.6s+148.999) = 0.  s1,2 = ±j9.6 s3,4 = −7.4 ± j1.9 . When K = 326, the roots are s1,2 = −6.5 ± j8.7 P7.32  s3,4 = −0.9 ± j3.2 .  The characteristic equation is 1+  K(s + 1)(s + 5) =0. s(s + 1.5)(s + 2)  15  20  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  331  Problems  4 3  *  2  *  Imag Axis  1 0  o  x  x  o**  x  -1 -2  *  -3  *  -4 -10  -8  -6  -4  -2  0  2  4  Real Axis  FIGURE P7.32 (s+1)(s+5) Root locus for 1 + K s(s+1.5)(s+2) = 0.  K  TABLE P7.32  ζ  Ts (sec)  P.O. (%)  1.57  0.707  0.98  1.4  3.48  0.707  1.1  5.8  2.35  0.69  1.3  4.0  Step Response Results for K = 1.57, K = 3.48, and K = 2.35.  (a) The breakaway point is s = −1.73; the entry point is s = −8.62.  (b) The damping ratio ζ = 0.707 when K = 1.57 and again when K = 3.46. (c) The minimum damping ratio ζ = 0.69 is achieved when K = 2.35. (d) The results are summarized in Table P7.32.The best choice of gain is K = 1.57. P7.33  (a) The root locus for the V-22 is shown in Figure P7.33a. The system is stable when 0 < K < 0.48 and K > 136.5. (b) The unit step input response (for K = 280) is shown in Figure P7.33b. The step response has a P.O. = 90% and Ts ≈ 50 sec.  (c) The plot of y(t) for a unit step disturbance is shown in Figure P7.33b.  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  CHAPTER 7  The Root Locus Method  2 1.5 1  Imag Axis  0.5 0  x  o  o  -1  -0.5  xxx  -0.5 -1 -1.5 -2 -3  -2.5  -2  -1.5  0  0.5  1  Real Axis  (i) Unit step input response 2 y(t) w/o prefilter .... (dotted line) y(t) with prefilter ____ (solid line)  y(t)  1.5 1 0.5 0  4  0 x 10  10  20  10  20  -3  30  40 50 Time (sec) (ii) Unit step disturbance response  60  70  80  60  70  80  3 y(t)  332  2 1 0 -1  0  30  40 Time (sec)  50  FIGURE P7.33 s2 +1.5s+0.5 (a) Root locus for 1 + K s(20s+1)(10s+1)(0.5s+1) = 0. (b) (i) Unit step input response with and without prefilter; (ii) Unit step disturbance response.  The response to the disturbance is oscillatory, but the maximum value of oscillation is about 0.003; so it is negligible. (d) The effect of adding a prefilter can be seen in Figure P7.33b. With  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  333  Problems  the prefilter we find P O = 7% and Ts ≈ 40 sec. P7.34  The characteristic equation is 1+  K(s + 2) =0. (s + 1)(s + 2.5)(s + 4)(s + 10)  The root locus is shown in Figure P7.34a. The roots, predicted and actual percent overshoot for K = 400, 500, and 600 are summarized in Table P7.34. The actual unit step input responses are shown in Figure P7.34b.  roots  ζ  predicted P.O. (%)  actual P.O. (%)  400  -13.5,-1.00 ± 5.71j,-1.98  0.173  57.6  51.6  500  -14.0,-0.75 ± 6.24j,-1.98  0.120  68.4  61.2  600  -14.4,-0.53 ± 6.71j,-1.98  0.079  77.9  69.6  TABLE P7.34  Summary for K = 400, 500, 600.  Root Locus 20  15  10  Imaginary Axis  K  5  0  −5  −10  −15  −20 −30  −25  −20  −15  −10 Real Axis  −5  FIGURE P7.34 s+2 (a) Root locus for 1 + K (s+1)(s+2.5)(s+4)(s+10) = 0.  0  5  10  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  334  CHAPTER 7  The Root Locus Method  1.6  K=400 .... (dotted line)  1.4  K=500 −−− (dashed line) K=600 ___ (solid line)  1.2  y(t)  1  0.8  0.6  0.4  0.2  0  0  2  4  6  8  10 Time (sec)  12  14  16  18  20  FIGURE P7.34 CONTINUED (b) Unit step input responses for K = 400, 500, 600.  (a) The root locus is shown in Figure P7.35 for the characteristic equation 1+  K(s + 1)2 =0. s(s2 + 1)  3  K=4.52  2  *  1  Imag Axis  P7.35  x  0  o  -1  x  x  -2  -3 -5  *  *  -4  -3  -2 Real Axis  FIGURE P7.35 (s+1)2 Root locus for 1 + K s(s2 +1) = 0.  -1  0  1  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  335  Problems  (b) When K = 4.52, the roots are s1 = −0.58 s2,3 = −1.96 ± j1.96 . The complex roots have ζ = 0.707. (c) The entry point is s = −3.38 when K = 7.41.  (d) The predicted P.O. = 4.5% The characteristic equation is 1+  K(s + 1)(s + 2)(s + 3) =0. s3 (s − 1)  (a) The root locus is shown in Figure P7.36.  8 6 4 2  Imag Axis  P7.36  (ζ = 0.707) and the actual P.O. = 17%.  0  o  o  o  x  x  -2 -4 -6 -8 -10  -8  -6  -4  -2  Real Axis  FIGURE P7.36 (s+1)(s+2)(s+3) Root locus for 1 + K = 0. s3 (s−1)  (b) When K = 2.96, the roots are s1,2 = ±j4.08 s3,4 = −0.98 ± j0.33 .  0  2  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  336  CHAPTER 7  The Root Locus Method  (c) When K = 20, the roots are s1 = −1.46 s2 = −1.07 s3,4 = −8.23 ± j2.99 . When K = 100, the roots are s1 s2 s3 s4  = −92.65 = −3.51 = −1.82 = −1.01 .  (d) When K = 20, the damping ratio is ζ = 0.94. Therefore, the predicted P.O. = 0.02%. The actual overshoot is P.O. = 23%. P7.37  Since we know that ess = 0 for a step input, we know that a = 0 or b = 0. Select a = 0. Also, ωn = 2π/T = 20 rad/sec. The desired characteristic polynomial is (s + r1 )(s + j20)(s − j20) = s3 + r1 s2 + 400s + 400r1 = 0 . The actual characteristic polynomial is 1+  2K =0, s(s + b)(s + 40)  s3 + (40 + b)s2 + 40bs + 2K = 0 .  or  Comparing the coefficients in the desired and actual characteristic polynomials, we determine that b = 10, r1 = 50, and K = 10000. P7.38  (a) The characteristic equation is 1+  K(s + 1) =0. s(s − 3)  √ The system is stable for K > 3. When K = 3, the roots are s = ±j 3.  (b) The root locus is shown in Figure P7.38a. (c) When K = 10 , the roots are s1 = −2 s2 = −5 .  Since both roots are real and stable, we expect that there will be zero overshoot. The actual response has a 40% overshoot, as seen in Figure P7.38b.  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  337  Problems  6  4  Imag Axis  2  0  o  x  x  -2  -4  -6 -6  -4  -2  0  2  4  6  2  2.5  3  Real Axis 1.4  1.2  Amplitude  1  0.8  0.6  0.4  0.2  0 0  0.5  1  1.5 Time (secs)  FIGURE P7.38 s+1 (a) Root locus for 1 + K s(s−3) = 0. (b) Unit step response.  P7.39  The loop transfer function is Gc (s)G(s) =  22K . (s + 1)(s2 + 8s + 22)  When K = 0.529, the closed-loop poles are s1,2 = −3.34 ± 1.83j and s3 = −2.32 and have the maximum damping ζ = 0.877. The root locus is shown in Figure P7.39a. The step response is shown in Figure P7.39b.  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  CHAPTER 7  The Root Locus Method  Root Locus 10 8 6  Imaginary Axis  4 2 0 −2 −4 −6 −8 −10 −14  −12  −10  −8  −6 Real Axis  −4  −2  0  2  Step Response 0.35  0.3  0.25  Amplitude  338  0.2  0.15  0.1  0.05  0  0  FIGURE P7.39 (a) Root locus for  0.5  1  22K (s+1)(s2 +8s+22)  1.5 2 Time (sec)  2.5  = 0. (b) Unit step response.  3  3.5  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  339  Advanced Problems  Advanced Problems The characteristic equation is 1+K  s+6 =0. s(s + 4)(s2 + 4s + 8)  The root locus is shown in Figure AP7.1. The gain at maximum ζ is  10  5  Imag Axis  AP7.1  x +  0  o  x+  +  x +  x  -5  -10 -10  -5  0  5  10  Real Axis  FIGURE AP7.1 s(s+4) Root locus for 1 + K s2 +2s+2 = 0.  K = 3.7 . The roots at K = 3.7 are s1 = −3.6424  s2,3 = −1.3395 ± +1.3553j  s4 = −1.6786 .  Using Figure 5.13 in Dorf & Bishop, the predicted percent overshoot and settling time are P.O. = 5%  and Ts = 3 sec ,  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  340  CHAPTER 7  The Root Locus Method  since ζ = 0.7 and a 6 = = 4.5 . ωn ζ 1.9(0.7) The actual percent overshoot and settling time are P.O. = 1% and Ts = 2.8 sec. The characteristic equation is 1+K  (s + 1)(s + 4) =0. s(s − 1)(s + 5)(s + 10)  The root locus is shown in Figure AP7.2a. The selected gain is K = 43.7. 15  Imaginary Axis  10  5  0  −5  −10  −15 −12  −10  −8  −6 −4 Real Axis  −2  0  2  1.5 System: syscl Peak amplitude: 1.48 Overshoot (%): 48.3 At time (sec): 0.857 System: syscl Settling Time (sec): 2.31  1 Amplitude  AP7.2  0.5  0  0  0.5  1  1.5  2 2.5 Time (sec)  3  3.5  4  4.5  FIGURE AP7.2 (s+1)(s+4) (a) Root locus for 1 + K s(s−1)(s+5)(s+10) = 0; (b) Step response for K = 43.7.  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  341  Advanced Problems  The actual percent overshoot (see Figure AP7.2b) is P.O. = 48.3%. AP7.3  The characteristic equation (with p as the parameter) is 1+p  s3  s(s + 1) =0. + s2 + 10  The root locus is shown in Figure AP7.3.  5 4 3 2  x  Imag Axis  1 +  0  x  o  o +  -1 x  -2 -3 -4 -5 -5  -4  -3  -2  -1  0  1  2  3  4  5  Real Axis  FIGURE AP7.3 s(s+1) Root locus for 1 + p s3 +s2 +10 = 0.  When p = 21 the dominant roots have a damping ratio of ζ = 0.707. AP7.4  The characteristic equation (with α as the parameter) is 1+α  s(s + 1) =0. + s2 + 1  s3  The root locus is shown in Figure AP7.4a. The steady-state error is 1 =1−α . s→0 1 + G(s)  ess = lim sE(s) = lim s→0  To meet the steady-state error specification, we require 0.9 < α < 1.1 . The step responses for α = 0.9, 1 and 1.1 are shown in Figure AP7.4b.  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  342  CHAPTER 7  The Root Locus Method  3  2  Imag Axis  1  x  0  x  o  o  x  -1  -2  -3 -3  -2  -1  0  1  2  3  Real Axis alpha=0.9 (solid); alpha=1.0 (dashed); alpha=1.1 (dotted) 1.8 1.6 1.4  Amplitude  1.2 1 0.8 0.6 0.4 0.2 0 0  5  10  15  20  25  30  35  40  45  50  Time (sec)  FIGURE AP7.4 s(s+1) (a) Root locus for 1 + p s3 +s2 +10 = 0. (b) Step responses for α = 0.9, 1 and 1.1.  AP7.5  The root locus is shown in Figure AP7.5. When K = 20.45, ζ = 0.707. The r1 ∼ root sensitivity is SK = ∆r1 /(∆K/20.45) = 3.156 87.76o . When K = 88, the complex roots lie on the jω-axis—a 330% increase in the gain.  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  343  Advanced Problems  5 4 3 2  Imag Axis  1 +  0  +x  x  x +  -1 -2 -3 -4 -5 -15  -10  -5  0  5  Real Axis  FIGURE AP7.5 Root locus for 1 + K s3 +10s21+7s−18 = 0.  A gain of K = 13 provides an acceptable response of Ts < 1 and P.O. < 7.5%. The root locus is shown in Figure AP7.6.  Root Locus 2.5 2 1.5 1 Imaginary Axis  AP7.6  0.5 0 −0.5 −1 −1.5 −2 −2.5 −3  −2.5  −2  FIGURE AP7.6 s2 +3s+6 Root locus for 1 + K s3 +2s 2 +3s+1 = 0.  −1.5 −1 Real Axis  −0.5  0  0.5  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  344  CHAPTER 7  AP7.7  The Root Locus Method  The root locus for the positive feedback system is shown in Figure AP7.7. 15  10  Imag Axis  5  0  x  x  -5  -10  -15 -15  -10  -5  0  5  10  15  Real Axis  FIGURE AP7.7 −1 Root locus for 1 + K s2 +12s+32 = 0.  The root locus is shown in Figure AP7.8a. When k = 0.448, all the roots  30  20  10  Imag Axis  AP7.8  x  0  x  o x  -10  -20  -30 -30  -20  -10  0 Real Axis  FIGURE AP7.8 (a) Root locus for 1 + k s3 +19s120s 2 +34s+120 = 0.  10  20  30  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  345  Advanced Problems  of the characteristic equation are real—the step response is shown in Figure AP7.8b. 1 0.9 0.8 0.7  Amplitude  0.6 0.5 0.4 0.3 0.2 0.1 0 0  0.5  1  1.5  2  2.5  3  Time (secs)  FIGURE AP7.8 CONTINUED (b) Step response with k = 0.448.  The root locus for each controller is shown in Figure AP7.9.  AP7.9  (a)  (b) 5 Imaginary Axis  Imaginary Axis  5  0  −5 −15  −10  −5 Real Axis  0  0  −5 −15  5  −10  (c)  0  5  0  5  (d)  15  5  10  Imaginary Axis  Imaginary Axis  −5 Real Axis  5 0 −5  0  −10 −15 −15  −10  −5 Real Axis  FIGURE AP7.9 Root locus for the various controllers.  0  5  −5 −15  −10  −5 Real Axis  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  346  CHAPTER 7  AP7.10  The Root Locus Method  The characteristic equation (with K as the parameter) is 1+K  s2 + 7s + 20 =0. s(s2 + 7s + 10)  The root locus is shown in Figure AP7.10. The steady-state value of the 10 8 6 4  Imag Axis  2 0 -2 -4 -6 -8 -10 -10  -8  -6  -4  -2  0 Real Axis  2  4  6  8  10  FIGURE AP7.10 s2 +7s+20 Root locus for 1 + K s(s 2 +7s+10) = 0.  step response for any K is 0.5. With K = 15 the closed-loop transfer function is T (s) =  10s + 150 . s3 + 22s2 + 115s + 300  The step response has the following characteristics: P.O. = 4.8% AP7.11  and  Ts = 2 seconds .  The root locus is shown in Figure AP7.11a. A suitable gain is K = 500. The step response is shown in Figure AP7.11b.  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  347  Advanced Problems  Root Locus 80 60  Imaginary Axis  40 20 0 −20 −40 −60 −80 −100  −80  −60  −40 −20 Real Axis  0  20  40  FIGURE AP7.11 (s+2)2 (a) Root locus for 1 + K s(s+10)(s+20)(s2 +3s+3.5) = 0.  Step Response 1.4 System: sys_cl Peak amplitude: 1.09 Overshoot (%): 9.01 At time (sec): 0.945  1.2  Amplitude  1 System: sys_cl Settling Time (sec): 2.39  0.8  0.6  0.4  0.2  0  0  0.5  1  1.5  2 Time (sec)  2.5  3  3.5  4  FIGURE AP7.11 CONTINUED: (b) Step response with K = 500.  AP7.12  The root locus is shown in Figure AP7.12a. The PI controller can be written as Gc (s) =  Kp s + KI s  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  CHAPTER 7  The Root Locus Method 8  6  4  Imag Axis  2  0  -2  -4  -6  -8 -7  -6  -5  -4  -3  -2  -1  0  1  2  Real Axis  Step Response From: U(1) 1.4  1.2  1  0.8 To: Y(1)  Amplitude  348  0.6  0.4  0.2  0  0  5  10  15  Time (sec.)  FIGURE AP7.12 (s+0.2) (a) Root locus for 1 + Kp s(s2 +7s+10) = 0. (b) Step response with Kp = 5.54.  and setting KI = 0.2Kp , the characteristic equation can be written as 1 + Kp  (s + 0.2) =0 + 7s + 10)  s(s2  A suitable gain is Kp = 5.55. The step response is shown in Figure AP7.12b.  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  349  Advanced Problems  AP7.13  The characteristic equation is 1 + K1 K2  1 = 0. (s + 5)(s − 1)  The root locus is shown in Figure AP7.12a. The fastest expected settling  Root Locus 4  3  Imaginary Axis  2  1  0  −1  −2  −3  −4 −6  −5  −4  −3  −2 Real Axis  −1  0  1  2  FIGURE AP7.13 1 Root locus for 1 + K1 K2 (s+5)(s−1) = 0.  time is Ts = 4/ωn ζ = 2 seconds since maximum |ωn ζ| = 2. AP7.14  The root locus of the uncompensated transfer function is shown in Figure AP7.14a. It can be seen that the system is unstable for Ku = 131.25 with a period of Tu = 0.72, as illustrated in FigureAP7.14b. Using the Ziegler-Nichols design formulas yields KP = 0.6Ku = 78.75, KI = 1.2Ku /Tu = 218.75, and KD = 0.6Ku Tu = 7.0875 where  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  CHAPTER 7  The Root Locus Method  Root Locus 30 System: sysg Gain: 131 Pole: 0.0153 + 8.66i Damping: −0.00176 Overshoot (%): 101 Frequency (rad/sec): 8.66  Imaginary Axis  20  10  0  −10  −20  −30 −40  −30  −20  −10 Real Axis  0  10  20  8  10  12  FIGURE AP7.14 10 = 0. (a) Root locus for 1 + Ku s(s+10)(s+7.5)  Step Response 2 1.8 1.6 1.4 Amplitude  350  1.2 1 0.8 0.6 0.4 0.2 0  0  2  4  6 Time (sec)  FIGURE AP7.14 CONTIUED: (b) Step response at the ultimate gain Ku = 131.  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  351  Advanced Problems  Step Response 1.6 System: sys_cl Peak amplitude: 1.6 Overshoot (%): 59.5 At time (sec): 0.445  1.4  Amplitude  1.2 1 System: sys_cl Settling Time (sec): 2.1  0.8 0.6 0.4 0.2 0  0  0.5  1  1.5 2 Time (sec)  2.5  3  3.5  FIGURE AP7.14 CONTINUED: (c) Step response with the Ziegler-Nichols tuned PID controller.  Step Response  −3  14  x 10  12 10  Amplitude  8 6 4 2 0 −2 −4  0  0.5  1  1.5 2 Time (sec)  2.5  3  3.5  FIGURE AP7.14 CONTINUED: (d) Disturbance response with the Ziegler-Nichols tuned PID controller.  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  352  CHAPTER 7  The Root Locus Method  Design Problems CDP7.1  The closed-loop transfer function from the input to the output is 26.035Ka θ(s) = 2 , R(s) s + (33.1415 + 26.035Ka K1 )s + 26.035Ka where we consider for the first time the tachometer feedback (see Figure CDP4.1 in Dorf and Bishop). The characteristic equation is 1 + K1  26.035Ka s =0. s2 + 33.1415s + 26.035Ka  The root locus is shown below. In accordance with the discussion in Chap30  20  Imag Axis  10  0  -10  -20  -30 -30  -20  -10  0 Real Axis  10  20  30  ter 5, we continue to use Ka = 22. This allows us to meet the overshoot specification (P.O. < 5%) without the tachometer feedback and to provides good steady-state tracking errors to a step input. To meet the design specifications of both P.O. and Ts we want the closed-loop poles to the left of −ζω = −4/0.3 = −13.33 and ζ > 0.69. A reasonable selection is K1 = 0.012. This places the closed-loop poles at s = −20 ± j13. DP7.1  (a) The characteristic equation is 1+  (s2  18K(s + 0.015)(s + 0.45) =0. + 1.2s + 12)(s2 + 0.01s + 0.0025)  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  353  Design Problems  Since we want a negative feedback system, we have Gc (s) = −K. When ωn > 2 and ζ = 0.15, the gain K = 0.12. The root locus is shown in Figure DP7.1a.  6  4 x  Imag Axis  2  o  oxx  -0.5  0  0  -2 x  -4  -6 -4  -3.5  -3  -2.5  -2  -1.5  -1  0.5  1  Real Axis  FIGURE DP7.1 18(s+0.015)(s+0.45) (a) Root locus for 1 + K (s2 +1.2s+12)(s2 +0.01s+0.0025) = 0.  (b) The unit step response is shown in Figure DP7.1b. The percent overshoot is P.O. = 100% .  (c) The characteristic equation with the anticipatory controller is 1+  18K(s + 2)(s + 0.015)(s + 0.45) =0. (s2 + 1.2s + 12)(s2 + 0.01s + 0.002s)  The root locus is shown in Figure DP7.1c. If we select K = 9.2/18 , then the complex roots have a damping ζ = 0.90. The roots are at s1 = −0.253 s2 = −0.019 s3,4 = −5.07 ± j2.50 .  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  CHAPTER 7  The Root Locus Method  0.7  0.6  Amplitude  0.5  0.4  0.3  0.2  0.1  0 0  20  40  60  80  100  120  140  160  180  200  Time (secs)  FIGURE DP7.1 CONTINUED: (b) Unit step response for gain controller.  6  4 x  2  Imag Axis  354  0  xx o o  o  -2 x  -4  -6 -6  -4  -2  0  2  4  6  Real Axis  FIGURE DP7.1 18(s+2)(s+0.015)(s+0.45) CONTINUED: (c) Root locus for 1 + K (s2 +1.2s+12)(s2 +0.01s+0.0025) = 0.  (d) The unit step response for the system with the anticipatory controller is shown in Figure DP7.1d.  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  355  Design Problems  1 0.9 0.8  Amplitude  0.7 0.6 0.5 0.4 0.3 0.2 0.1 0 0  20  40  60  80  100  120  140  160  180  200  Time (secs)  FIGURE DP7.1 CONTINUED: (d) Unit step response for anticipatory controller.  DP7.2  The characteristic equation is 1+  10K(s + 1) =0. s(s2 + 4.5s + 9)  (a) The root locus is shown in Figure DP7.2a. When K = 0.435, we have ζ = 0.6 and the roots are s1 = −0.368 s2,3 = −2.1 ± j2.75 . (b) The response to a step input is shown in Figure DP7.2b. The performance results are P.O. = 0% Tss = 10 sec ess = 0 . (c) We have ζ = 0.41 when K = 1.51. The step response is shown in Figure DP7.2b. The performance results to the step input are P.O. = 0% Ts = 4 sec ess = 0 .  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  356  CHAPTER 7  The Root Locus Method  5 4 3 2  x  Imag Axis  1 0  o  x  -1  0  -1 -2  x  -3 -4 -5 -5  -4  -3  -2  1  2  3  4  5  Real Axis  FIGURE DP7.2 10(s+1) (a) Root locus for 1 + K s(s2 +4.5s+9) = 0.  1 0.9 K=0.435 ____ (solid line)  0.8  K=1.510 ---- (dashed line)  Amplitude  0.7 0.6 0.5 0.4 0.3 0.2 0.1 0 0  2  4  6  8  10  12  Time (sec)  FIGURE DP7.2 CONTINUED: (b) Unit step responses for K = 0.425, 1.51.  DP7.3  The characteristic equation is 1+  K(s2 + 6.5s + 12) =0. s(s + 1)(s + 2)  14  16  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  357  Design Problems  (a) The root locus is shown in Figure DP7.3.  6  4  2  Imag Axis  o  0  x  x  x  -2  -1  0  o  -2  -4  -6 -6  -5  -4  -3  1  Real Axis  FIGURE DP7.3 s2 +6.5s+12 = 0. Root locus for 1 + K s(s+1)(s+2)  When K = 41, the roots are s1 = −37.12 and s2,3 = −3.44 ± j1.19 .  (b) The percent overshoot is P.O. ≈ 1% when ζ = 0.82 at K = 0.062. (c) Select K > 300. DP7.4  The characteristic equation is 1+K  10(0.01s + 1) =0. s(s2 + 10s + 10K1 )  If we choose K1 = 2.5, then the root locus will start at s = 0, −5 and -5. This is shown in Figure DP7.4. The root locus then has a nice shape so that we can select K to place the complex poles where desired and the one real root will be farther in the left half-plane; thus the notion of dominant poles will be valid. So, if we desire a P.O. < 5%, we want ζ > 0.69. This occurs when K ≈ 3. Thus, our design is K1 = 2.5  and K = 3 .  The unit step response is shown in Figure DP7.4. The settling time is less than 3.5 sec and the P O < 4%. The response to a unit step disturbance is also shown in Figure DP7.4. The steady-state error magnitude to the disturbance is 0.33.  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  358  CHAPTER 7  The Root Locus Method 4  3  2 K=3 -->  Imag Axis  1  0  -1  -2  -3  -4 -20  -15  -10  -5 Real Axis  0  5  10  1.4  Input step response Disturbance step response 1.2  1  y(t)  0.8  0.6  0.4  0.2  0  0  0.5  1  1.5  2  2.5 Time (sec)  3  3.5  4  4.5  5  FIGURE DP7.4 10(0.01s+1) (a) Root locus for 1 + K s(s2 +10s+25) = 0. (b) System response to step input and disturbance.  DP7.5  The characteristic equation is 1+K  s+1 =0. s(s − 0.1)(s2 + 10s + 41)  The root locus is shown in Figure DP7.5a. The system is stable for 5 < K < 300. The step response with K =  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  359  Design Problems  Root Locus 10 8 System: sysgc Gain: 90.5 Pole: −1.42 + 2.24i Damping: 0.536 Overshoot (%): 13.6 Frequency (rad/sec): 2.66  6  Imaginary Axis  4 2 0 −2 −4 −6 −8 −10 −12  −10  −8  −6 −4 Real Axis  −2  0  2  Step Response 1.6 System: sys_cl Peak amplitude: 1.57 Overshoot (%): 57 At time (sec): 1.24  1.4  Amplitude  1.2  System: sys_cl Time (sec): 3.39 Amplitude: 0.98  1 0.8 0.6 0.4 0.2 0  0  0.5  1  1.5  2 2.5 Time (sec)  3  3.5  4  4.5  FIGURE DP7.5 s+1 (a) Root locus for 1 + K s(s−0.1)(s 2 +10s+41) = 0. (b) Step response with K = 875.  90.5 is shown in Figure DP7.5b. We choose K = 90.5 to minimize the settling time. The damping of the dominant poles is ζ = 0.54, so that the estimated percent overshoot is P.O. = 13%. The actual percent overshoot and settling time are P.O. = 57% and Ts = 3.4 seconds. The match between the actual and predicted percent overshoot can be improved by selecting a much higher gain K, but then the step response becomes overy oscillatory and the settling time increases too much for a typical high-performance aircraft.  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  360  CHAPTER 7  The characteristic equation is 1+K  s+2 =0. s(s + 10)(s − 1)  The maximum damping is ζ = 0.46 at K = 55. The root locus is shown in Figure DP7.6a; the step response is shown in Figure DP7.6b. The percent overshoot and settling time are P.O. = 61.3% and Ts = 2 seconds.  20 15 10  Imag Axis  5 +  0  x  +  o  x x  +  -5 -10 -15 -20 -20  -15  -10  -5  0  5  10  15  20  Real Axis 1.8 1.6 1.4 1.2  Amplitude  DP7.6  The Root Locus Method  1 0.8 0.6 0.4 0.2 0 0  0.5  1  1.5  2  2.5  3  Time (secs)  FIGURE DP7.6 s+2 (a) Root locus for 1 + K s(s+10)(s−1) = 0. (b) Step response with K = 55.  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  361  Design Problems  DP7.7  The loop transfer function is Gc (s)G(s) =  KP s + KI . s(s + 1)(0.5s + 1)  One possible set of PI controller gains are KP = 0.82 and KI = 0.9. The step response is shown in Figure DP7.7.  Step Response 1.4 System: syscl Peak amplitude: 1.05 Overshoot (%): 4.59 At time (sec): 3.57  1.2  Amplitude  1  System: syscl Settling Time (sec): 4.94  0.8  0.6  0.4  0.2  0  0  1  2  3  4 Time (sec)  5  6  7  8  FIGURE DP7.7 Step response for with PI controller Gc (s) = (0.82s + 0.9)/s.  DP7.8  The closed-loop transfer function is T (s) =  Vo (s) G(s) = . V (s) 1 + KG(s)  The dc gain is T (0) =  G(0) 1 ≈ . 1 + KG(0) K  The root locus is shown in Figure DP7.8. The maximum value of K for stability is K = 0.062 .  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  362  CHAPTER 7  The Root Locus Method  x10 7 2 1.5 1  +  Imag Axis  0.5 0+  x  x  -0.5 -1  +  -1.5 -2 -2  -1.5  -1  -0.5  0  0.5  1  1.5  Real Axis  2 x10 7  FIGURE DP7.8 3.142K1 ×1017 Root locus for 1 + K (s+3142)(s+10 7 )2 = 0.  Therefore, the minimum dc gain is about 1/0.062=16. Selecting K = 0.05  and R1 = 10 K  yields R2 = 19R1 = 190 K . DP7.9  The closed-loop transfer function (with Gp (s) = 1 and K = 1) is T (s) =  2s3 + 6s2 + 14s + 10 . s4 + 6s3 + 13s2 + 26s + 6  So, if we select Gp (s) = 1/T (0) = 0.6, the step response (with K = 1) will have a zero steady-state tracking error. The root locus is shown in Figure DP7.9a. The step responses for K = 1, 1.5 and 2.85 are shown in Figure DP7.9b. For K = 1, we have P.O. = 0%, Tr = 7.8 and Ts = 13.9; for K = 1.5, we have P.O. = 0%, Tr = 5.4 and Ts = 9.6; and for K = 2.85, we have P.O. = 5.2%, Tr = 0.5 and Ts = 7.3. The best gain selection is K = 2.85.  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  363  Design Problems  8 6 4  Imag Axis  2  x  0  x  o  -2  x  x  -4 -6 -8 -8  -6  -4  -2  0  2  4  6  8  Real Axis K=1 (solid); K=1.5 (dashed); K=2.85 (dotted) 1.2  1  Amplitude  0.8  0.6  0.4  0.2  0 0  2  4  6  8  10  12  14  16  18  20  Time (sec)  FIGURE DP7.9 6(s+1) (a) Root locus for 1 + K s(s+4)(s2 +2s+5) = 0. (b) Step responses with K = 1, 1.5, 2.85.  DP7.10  A suitable selection of the various parameters is ζ = 0.5  and q = 3/5 .  With q = 3/5, the open-loop zeros are real and equal. Then, it follows that λ=  2q =3. 1−q  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  364  CHAPTER 7  The Root Locus Method  The root locus is shown in Figure DP7.10. A reasonable choice of gain is K = 30 . The resulting step response is extremely fast with no overshoot. The closed-loop transfer function is approximately given by T (s) ≈  1923 . s + 1923  6  4 x  Imag Axis  2  0  o  x  -2 x  -4  -6 -6  -4  -2  0  2  4  6  Real Axis  FIGURE DP7.10 4s2 +4s+1 Root locus for 1 + K 0.0625s 3 +0.25s2 +s = 0.  DP7.11  The characteristic equation (with K as the parameter) is 1+K  10(s2 + 10) =0. s3 + 20s  The root locus is shown in Figure DP7.11a. To maximize the closed-loop system damping we choose K = 0.513. The step response is shown in Figure DP7.11b.  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  365  Design Problems 5  4  3  2  Imag Axis  1  0  -1  -2  -3  -4  -5 -2  -1.5  -1  -0.5 Real Axis  0  0.5  1  Step Response From: U(1) 1.4  1.2  0.8 To: Y(1)  Amplitude  1  0.6  0.4  0.2  0  0  1  2  3  4  5  Time (sec.)  FIGURE DP7.11 10(s2 +10) (a) Root locus for 1 + K s3 +20s = 0. (b) Step response with K = 0.513.  DP7.12  The characteristic equation is 1+K  s + 1.5 =0. (s + 1)(s + 2)(s + 4)(s + 10)  The root locus is shown in Figure DP7.12a.  6  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  CHAPTER 7  The Root Locus Method  10 8 6 4  Imag Axis  2 0  x  x  x o x  -2 -4 -6 -8 -10 -15  -10  -5  0  5  Real Axis K=100 (solid); K=300 (dashed); K=600 (dotted) 1.6 1.4 1.2 1  Amplitude  366  0.8 0.6 0.4 0.2 0 0  1  2  3  4  5  6  7  8  9  10  Time (sec)  FIGURE DP7.12 s+1.5 (a) Root locus for 1 + K (s+1)(s+2)(s+4)(s+10) = 0. (b) Step response with K = 100, 300, 600.  The closed-loop system roots are: K = 100 : s1 = −11.38 K = 300 : s1 = −12.94 K = 600 : s1 = −14.44  s2,3 = −2.09 ± 3.10j s2,3 = −1.29 ± 5.10j s2,3 = −0.53 ± 6.72j  The step responses are shown in Figure DP7.12b.  s4 = −1.45 s4 = −1.48 s4 = −1.49  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  367  Design Problems  DP7.13  The closed-loop transfer function is T (s) =  s3  Ka . + Ka K2 s + Ka  s2  +  A suitable choice of gains is Ka = 0.52  and K2 = 3 .  The step response is shown in Figure DP7.13.  1.2  1  Amplitude  0.8  0.6  0.4  0.2  0  0  2  4  6  8  10  12  14  16  18  20  Time (secs)  FIGURE DP7.13 Step response with Ka = 0.52 and K2 = 3.  DP7.14  The characteristic equation is s2 + 10KD s + 10(KP + 1) = 0 . In the Evans form we have 1 + KD  10(s + τ ) =0. s2 + 10  The root locus is shown in Figure DP7.14 for τ = 6. As τ → 0,√the dominant closed-loop pole approaches s = 0 as KD√→ ∞. As τ → 10, the dominant closed-loop pole approaches s = − 10 as KD → ∞. A viable controller is KP = 72 and KD = 12 when τ = 6.  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  CHAPTER 7  The Root Locus Method  Root Locus 8  6  4  Imaginary Axis  368  2  0  −2  −4  −6  −8 −25  FIGURE DP7.14 Root locus when τ = 6.  −20  −15  −10 Real Axis  −5  0  5  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  369  Computer Problems  Computer Problems The root locus for parts (a)-(d) are shown in Figures CP7.1a - CP7.1d.  num=[30]; den=[1 14 43 30]; rlocus(sys) 30  Imaginary Axis  20  10  0  −10  −20  −30 −40  −30  −20  −10 Real Axis  0  10  20  num=[1 20]; den=[1 4 20]; rlocus(sys) 20 15 10 Imaginary Axis  CP7.1  5 0 −5 −10 −15 −20 −70  −60  −50  −40  −30 Real Axis  −20  −10  0  10  FIGURE CP7.1 s+20 (a) Root locus for 1 + k s3 +14s230 = 0. (b) Root locus for 1 + k s2 +4s+20 = 0. +43s+30  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  CHAPTER 7  The Root Locus Method  num=[1 1 2]; den=[1 6 10 0]; rlocus(sys) 1.5  Imaginary Axis  1  0.5  0  −0.5  −1  −1.5 −6  −5  −4  −3 −2 Real Axis  −1  0  1  num=[1 4 6 10 6 4]; den=[1 4 4 1 1 10 1]; rlocus(sys) 1.5  1  Imaginary Axis  370  0.5  0  −0.5  −1  −1.5 −7  −6  −5  −4  −3 Real Axis  −2  −1  0  1  FIGURE CP7.1 2 +s+2 CONTINUED: (c) Root locus for 1 + k s(ss2 +6s+10) = 0. (d) Root locus for 1 + 5  4  3  2  +4s +6s +10s +6s+4 k s6s+4s 5 +4s4 +s3 +s2 +10s+1 = 0.  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  371  Computer Problems  CP7.2  The maximum value of the gain for stability is k = 0.791. The m-file script and root locus is shown in Figure CP7.2.  Select a point in the graphics window num=[1 -2 2]; den=[1 3 2 0]; sys = tf(num,den);  selected_point =  rlocus(sys) rlocfind(sys)  -0.0025 + 0.6550i ans = 0.8008 1  0.8  0.6  0.4  Imag Axis  0.2  0  −0.2  −0.4  −0.6  −0.8  −1 −3  −2.5  −2  −1.5  −1  −0.5 Real Axis  0  0.5  1  1.5  2  FIGURE CP7.2 Using the rlocfind function.  The value of k = 0.8008 selected by the rlocfind function is not exact since you cannot select the jω-axis crossing precisely. The actual value is determined using Routh-Hurwitz analysis. CP7.3  The partial fraction expansion of Y (s) is Y (s) =  s(s2  s+6 0.1667 1.6667 1.5 = − + . + 5s + 4) s+4 s+1 s  The m-file script and output is shown in Figure CP7.3.  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  372  CHAPTER 7  The Root Locus Method r= 0.1667 -1.6667 1.5000 p=  num=[1 6]; den=[1 5 4 0]; [r,p,k]=residue(num,den)  -4 -1 0 k= []  FIGURE CP7.3 Using the residue function.  The characteristic equation is 1+p  s−1 =0. s2 + 5s + 10  The root locus is shown in Figure CP7.4. The closed-loop system is stable for 0 < p < 10 .  n*+,-. /.01 23n,-. 4 .501 6789*:;:<:= 2.5 2 1.5 1 Imaginary Axis  CP7.4  0.5 0 −0.5 −1 −1.5 −2 −2.5 −8  −7  −6  FIGURE CP7.4 s−1 Root locus for 1 + p s2 +5s+10 = 0.  −5  −4  −3 −2 Real Axis  −1  0  1  2  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  373  Computer Problems  CP7.5  The characteristic equation is 1+k  s+1 =0. s2  The root locus is shown in Figure CP7.5. For k = 2 we obtain s1,2 = −1 ± j, that is, we have ζ = 0.707. num=[1 1]; den=[1 0 0]; sys = tf(num,den); hold off, clf rlocus(sys); hold on plot([0 -2],[0 2*tan(acos(0.707))],'--') plot([0 -2],[0 -2*tan(acos(0.707))],'--') plot([-1 -1],[1 -1],'*') 2.5  2  1.5  1  Imag Axis  0.5  0  -0.5  -1  -1.5  -2  -2.5 -3  -2.5  -2  -1.5  -1 Real Axis  -0.5  0  0.5  1  FIGURE CP7.5 Root locus for 1 + k s+1 = 0. s2  CP7.6  We choose a controller with two real poles and two real zeros selected to meet the steady-state specification. The characteristic equation is 1+K  (s + 5.5)(s + 0.01) 10 =0. (s + 6.5)(s + 0.0001) s3 + 15s2 + 50s  The m-file and root locus is shown in Figure CP7.4a. From the root locus we can select the value of the gain K that results in an estimated P.O. ≤ 5% and a ζωn ≤ −2 to meet the settling time specification. We  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  374  CHAPTER 7  The Root Locus Method  select K = 8.58. The step response is shown in Figure CP7.6b showing the percent overshoot and settling time specifications are satisfied. The velocity constant is kv = 145.2 which implies a steady-state error to a ramp input of ess = 1/kv = 0.0069. 25 20 15  System: untitled1 Gain: 8.58 Pole: Damping: 0.807 Overshoot (%): 1.37 Frequency (rad/sec): 2.58  10 Imaginary Axis  ng=10; dg=conv([1 10 0],[1 5]); s ysg=tf(ng,dg); nh=conv([1 0.01],[1 5.5]); dh=conv([1 6.5],[1 0.0001]); sysh=tf(nh,dh); figure(1) rlocus(sysg*sysh) K=8.58; sysh=tf(K*nh,dh); sys=series(sysg,sysh);syscl=feedback(sys,1) figure(2) subplot(121) step(syscl); Kv=10*8.58*0.01*5.5/10/6.5/0.0001/5 systd=feedback(sysg,sysh); subplot(122) step(systd)  5 0 −5 −10 −15 −20 −25 −40  −35  −30  −25  Step Response  −15 −10 Real Axis  −5  0  5  10  Disturbance Response  1.4  0.14  System: syscl Peak 1.2amplitude: 1.02 Overshoot (%): 1.77 At time (sec): 2.14  0.12  1  0.1 System: syscl Settling Time (sec): 1.51  0.8  y(t)/Q  Amplitude  −20  0.08  0.6  0.06  0.4  0.04  0.2  0.02  0  0  1  2 3 Time (sec)  4  5  0  0  200 400 Time (sec)  600  FIGURE CP7.6 (a) Root locus. (b) Step response and disturbance response.  CP7.7  The m-file script to generate the root locus for each controller in parts (a)-(c) is shown in Figure CP7.7. The performance region is indicated on each root locus in Figures CP7.7b - CP7.7d. For part (a), the controller gain is found to be Gc (s) = 11.3920. The integral controller in part (b)  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  375  Computer Problems  numg=[1]; deng=[1 5 6]; sysg = tf(numg,deng); t=[0:0.1:15]; % % Part (a) % ÈSelect a point in the graphics window sys1 = sysg; rlocus(sys1), grid selected_point = hold on plot([-0.4 -0.4],[-6 6],'--',... -2.5030 + 3.3380i [0 -6*tan(36.2*pi/180)],[0 6],'--',... [0 -6*tan(36.2*pi/180)],[0 -6],'--') ans = hold off [kp,poles] = rlocfind(sys1) 11.3920 % % Part (b) % numc=[1]; denc=[1 0]; sysc = tf(numc,denc); sys2 = series(sysc,sysg); Select a point in the graphics window figure rlocus(sys2), grid selected_point = hold on plot([-0.4 -0.4],[-6 6],'--',... -0.6690 + 0.8210i [0 -6*tan(36.2*pi/180)],[0 6],'--',... [0 -6*tan(36.2*pi/180)],[0 -6],'--') ans = hold off [ki,poles] = rlocfind(sys2) 4.0930 % % Part (c) % Plot performance region boundaries on graph. figure numc=[1 1]; denc=[1 0]; sysc = tf(numc,denc); Select a point in the graphics window sys3 = series(sysc,sysg); rlocus(sys3), grid selected_point = hold on plot([-0.4 -0.4],[-6 6],'--',... -2.0695+ 2.7387i [0 -6*tan(36.2*pi/180)],[0 6],'--',... [0 -6*tan(36.2*pi/180)],[0 -6],'--') ans = hold off [kpi,poles] = rlocfind(sys3) 9.2516 % % Part (d) % figure sys1_o = kp*sys1; sys1_cl = feedback(sys1_o,[1]); sys2_o = ki*sys2; sys2_cl = feedback(sys2_o,[1]); sys3_o = kpi*sys3; sys3_cl = feedback(sys3_o,[1]); [y1,t]=step(sys1_cl,t); [y2,t]=step(sys2_cl,t); [y3,t]=step(sys3_cl,t); plot(t,y1,t,y2,'--',t,y3,':'),grid xlabel('time [sec]'),ylabel('y(t)') title('Gc(s): proportional (solid), integral (dashed) & PI (dotted)')  FIGURE CP7.7 (a) Script to generate the root locus for each controller.  is determined to be Gc (s) =  4.093 . s  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  CHAPTER 7  The Root Locus Method 6  4  Imag Axis  2  0  -2  -4  -6 -4  -3  -2  -1 Real Axis  0  1  2  FIGURE CP7.7 CONTINUED: (b) Root locus for proportional controller with selected K = 11.3920.  The proportional integral (PI) controller in part (c) is  6  4  2  Imag Axis  376  0  -2  -4  -6 -4  -3  -2  -1 Real Axis  0  1  2  FIGURE CP7.7 CONTINUED: (c) Root locus for integral controller with selected K = 4.0930.  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  377  Computer Problems  Gc (s) =  9.2516(s + 1) . s  The proportional controller is stable for all K > 0 but has a significant 6  4  Imag Axis  2  0  -2  -4  -6 -4  -3  -2  -1 Real Axis  0  1  2  FIGURE CP7.7 CONTINUED: (d) Root locus for PI controller with selected K = 9.2516.  steady-state error. The integral controller has no steady-state error, but is stable only for K < 30. The PI controller has zero steady-state error and is stable for all K > 0. Additionally, the PI controller has a fast transient response. The step responses for each controller is shown in Figure CP7.7e.  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  378  CHAPTER 7  The Root Locus Method Gc(s): proportional (solid), integral (dashed) & PI (dotted)  1.4  1.2  1  y(t)  0.8  0.6  0.4  0.2  0  0  5  10  15  time [sec]  FIGURE CP7.7 CONTINUED: (e) Step responses for each controller.  CP7.8  The loop transfer function can be written as Gc (s)G(s) =  K1 + K2 s s+5 = K̄2 2 2 Js s  where K̄2 = K2 /J . The parameter of interest for the root locus is K̄2 . The root locus is shown in Figure CP7.8. The selected value of K̄2 = 7.1075 . Therefore, K2 = 7.1075 J  and  K1 = 35.5375 . J  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  379  Computer Problems num=[1 5]; den=[1 0 0]; sys=tf(num,den); rlocus(sys); rlocfind(sys) 10 8 6 +  4  Imag Axis  2 0  o  x  -2 -4 +  -6 -8 -10 -10  -8  -6  -4  -2  0  2  4  6  8  10  Real Axis  FIGURE CP7.8 Root locus to determine K̄2 .  The value of K that results in a damping ratio of ζ = 0.707 is K = 5.2. The root locus is shown in Figure CP7.9. Root Locus 5 4 3 2 Imaginary Axis  CP7.9  s = -0.68 + 0.68j  1 0 −1  s = -6.63 s = -0.68 - 0.68j  −2 −3 −4 −5 −10  −5  0 Real Axis  FIGURE CP7.9 Root locus for 1 + K s3 +8s21+10s+1 = 0.  5  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  380  CHAPTER 7  (a) The characteristic equation is s3 + (2 + k)s2 + 5s + 1 = 0 . (b) The Routh array is s3  1  5  s2  2+k  1  s1  5k+9 2+k  so  1  Root Locus 2 1.5 1 Imaginary Axi s  CP7.10  The Root Locus Method  0.5 0 ?-0.5 ?-1 ?-1.5 ?-2 ?-2.5  ?-2  ?-1.5  ?-1  ?-0.5  Real Axi s  FIGURE CP7.10 2 Root locus for 1 + k s3 +2ss2 +5s+1 = 0.  For stability we require 2 + k > 0 or  k > −2  and 5k + 9 > 0 or  k > −9/5 .  Therefore, the stability region is defined by k > −1.8 .  0  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  381  Computer Problems  (c) Rearranging the characteristic equation yields 1+k  s2 s3 + 2s2 + 5s + 1 = 0 .  The root locus is shown in Figure CP7.10. We see that the system is stable for all k > 0.  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  C H A P T E R  8  Frequency Response Methods  Exercises E8.1  Given the loop transfer function L(s) =  4 , (s + 2)2  we determine that |L(jω)| =  4 4 + ω2  and φ(ω) = −2 tan−1 ω/2 .  The frequency response is shown in Figure E8.1.  Bode Diagram  Magnitude (dB)  0 −20 −40 −60  Phase (deg)  −80 0 −45 −90 −135 −180 −2 10  FIGURE E8.1 Frequency response for L(s) =  382  −1  10  0  10 Frequency (rad/sec)  4 . (s+2)2  1  10  2  10  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  383  Exercises  The magnitude and phase angle for ω = 0, 0.5, 1, 2, 4, ∞ are summarized in Table E8.1.  ω  0  0.5  1  2  4  ∞  |L(jω)|  1  0.94  0.80  0.50  0.20  0  φ (deg)  0  -28.07  -53.13  -90  –126.87  -180  TABLE E8.1  E8.2  Magnitude and phase for L(s) =  4 . (s+2)2  The transfer function is G(s) =  5000 . (s + 70)(s + 500)  The frequency response plot is shown in Figure E8.2. The phase angle is computed from φ = − tan−1  ω ω − tan−1 . 70 500  The phase angles for ω = 10, 100 and 700 are summarized in Table E8.2.  ω  TABLE E8.2  10  200  700  |G(jω)|  -16.99  -27.17  -41.66  φ (deg)  -9.28  -92.51  -138.75  Magnitude and phase for G(s) =  5000 . (s+70)(s+500)  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  384  CHAPTER 8  Frequency Response Methods  Bode Diagram  Magnitude (dB)  0 −20 −40 −60 −80  Phase (deg)  −100 0 −45 −90 −135 −180 0 10  1  10  FIGURE E8.2 Frequency response for G(s) =  E8.3  2  3  10 Frequency (rad/sec)  4  10  10  5000 (s+70)(s+500) .  The loop transfer function is L(s) =  300(s + 100) . s(s + 10)(s + 40)  The phase angle is computed via φ(ω) = −90o − tan−1  ω ω ω − tan−1 + tan−1 . 10 40 100  At ω = 28.3, we determine that φ = −90o − 70.5o − 35.3o + 15.8o = 180o . Computing the magnitude yields 1  |L(jω)| =  ω 2 2 300(100)(1 + ( 100 ) ) 1  1  ω 2 2 ω 2 2 ω10(1 + ( 10 ) ) 40(1 + ( 40 ) )  when ω = 28.3. We can also rewrite L(s) as L(s) =  s 75( 100 + 1) . s s s( 10 + 1)( 40 + 1)  = 0.75 ,  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  385  Exercises  Then, the magnitude in dB is ω 2 ω ) ) − 10 log 10 (1 + ( )2 ) 100 10 ω − 10 log 10 (1 + ( )2 ) − 20 log10 ω = −2.5 dB , 40  20 log 10 |L| = 20 log 10 (75) + 10 log 10 (1 + (  at ω = 28.3. E8.4  The transfer function is G(s) =  Ks . (s + a)(s + 10)2  Note that φ = 0o at ω = 3, and that φ = +90o − tan−1  ω ω − 2 tan−1 . a 10  Substituting ω = 3 and solving for a yields a=2. Similarly, from the magnitude relationship we determine that K = 400 . E8.5  The lower portion for ω < 2 is 20 log  K = 0 dB , ω  at ω = 8. Therefore, 20 log  K = 0 dB 8  which occurs when K=8. We have a zero at ω = 2 and another zero at ω = 4. The zero at ω = 4 yields a = 0.25 . We also have a pole at ω = 8, and a second pole at ω = 24. The pole at ω = 24 yields b = 1/24 .  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  386  CHAPTER 8  Frequency Response Methods  Therefore, G(s) = E8.6  8(1 + s/2)(1 + s/4) . s(1 + s/8)(1 + s/24)(1 + s/36)  The loop transfer function is L(s) =  10 . s(s/5 + 1)(s/100 + 1)  The Bode diagram is shown in Figure E8.6. When 20 log 10 |L(jω)| = 0 dB, we have ω = 9.4 rad/sec . Bode Diagram  Magnitude (dB)  50 0 −50 −100  Phase (deg)  −150 −90 −135 −180 −225 −270 −1 10  0  10  FIGURE E8.6 Bode Diagram for L(s) =  E8.7  1  2  10 10 Frequency (rad/sec)  3  10  4  10  10 . s(s/5+1)(s/100+1)  The transfer function is T (s) =  4 . (s2 + s + 1)(s2 + 0.4s + 4)  (a) The frequency response magnitude is shown in Figure E8.7. The frequency response has two resonant peaks at ωr1 = 0.8 rad/sec  and ωr2 = 1.9 rad/sec .  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  387  Exercises  10  Gain dB  5 0 -5 -10 10-1  100 Frequency (rad/sec)  101  Amplitude  1.5 1 0.5 0 0  2  4  FIGURE E8.7 (a) Bode Diagram for T (s) =  6  8  10 12 Time (secs)  4 (s2 +s+1)(s2 +0.4s+4) .  14  16  18  20  (b) Unit step response.  (b) The percent overshoot is P.O. = 35% , and the settling time is Ts ≈ 16 sec . (c) The step response is shown in Figure E8.7. E8.8  (a) The break frequencies are ω1 = 1 rad/sec, ω2 = 5 rad/sec, and ω3 = 20 rad/sec . (b) The slope of the asymptotic plot at low frequencies is 0 dB/decade. And at high frequencies the slope of the asymptotic plot is -20 dB/decade. (c) The Bode plot is shown in Figure E8.8.  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  388  CHAPTER 8  Frequency Response Methods  Bode Diagram 20  Magnitude (dB)  10 0 −10 −20 −30 180  Phase (deg)  135 90 45 0 −45 −90 −2 10  −1  10  FIGURE E8.8 Bode Diagram for Gc (s)G(s) =  1  2  10  3  10  100(s−1) . s2 +25s+100  The Bode diagram for G(s) is shown in Figure E8.9. 40  Gain dB  20 0 -20 -40 10-1  100  101  102  103  102  103  Frequency (rad/sec)  50  Phase deg  E8.9  0  10 10 Frequency (rad/sec)  0 -50 10-1  100  101 Frequency (rad/sec])  FIGURE E8.9 Bode Diagram for G(s) =  (s/5+1)(s/20+1) . (s+1)(s/80+1)  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  389  Exercises  E8.10  The frequency response has two peaks; the first peak at f ≈ 1.8 and the second peak at f ≈ 3.1. One possible G(jω) is 1  G(jω) =    (jωτ + 1) 1 +    2ζ1 ωn1    jω +    jω ωn1  2    1+    2ζ2 ωn2    jω +    jω ωn2  2  ,  where τ=  1 , 2π(0.2)  ωn1 = 2π(1.8 × 103 )  ζ1 = 0.15;  ζ2 = 0.15;  ωn2 = 2π(3.1 × 103 ) .  The damping ratios are estimated using Figure 8.10 in Dorf & Bishop. E8.11  The Bode plot is shown in Figure E8.11. The frequency when 20 log10 |GC G(ω)| = 0 is ω = 9.9 rad/sec. Bode Diagram 20  Magnitude (dB)  0 −20 −40 −60 −80 −100 −120 0  Phase (deg)  −45 −90 −135 −180 −225 −270 −1 10  0  10  FIGURE E8.11 Bode Diagram for Gc (s)G(s) =  E8.12  1  10 Frequency (rad/sec)  2  10  1000 . (s2 +10s+100)(s+2)  (a) The transfer function is G(s) = C(sI − A)−1 B + D = (b) The Bode plot is shown in Figure E8.12.  −5(s − 1) . s2 + 3s + 2  3  10  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  390  CHAPTER 8  Frequency Response Methods Bode Diagram  Phase (deg)  Magnitude (dB)  10 0 -10 -20  270 180 90 -2 10  10  FIGURE E8.12 Bode Diagram for G(s) =  0  10 Frequency (rad/sec)  10  1  10  2  −5(s−'1) . s2 +3s+2  The closed-loop transfer function is T (s) =  s3  +  100 . + 20s + 110  11s2  The Bode plot of T (s) is shown in Figure E8.13, where ωB = 4.9 rad/sec.  Bode Diagram  Magnitude (dB)  50  Phase (deg)  E8.13  -1  0  -3 dB  -50 -100  -45 -90 -135 -180 -225 -270 -1 10  FIGURE E8.13 Bode Diagram for T (s) =  ωb=4.9  10  0  1  10 Frequency (rad/sec)  100 s3 +11s2 +20s+110 .  10  2  10  3  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  391  Exercises  E8.14  The loop transfer function is L(s) =  20 . (s2 + 1.4s + 1)(s + 10)  The Bode plot of L(s) is shown in Figure E8.14. The frequency when 20 log 10 |L(ω)| = 0 is ω = 1.32 rad/sec. Bode Diagram  Magnitude (dB)  50  0  −50  −100  −150 0  Phase (deg)  −45 −90 −135 −180 −225 −270 −2 10  FIGURE E8.14 Bode Diagram for L(s) =  E8.15  −1  0  10  1  10 10 Frequency (rad/sec)  2  10  3  10  20 . (s2 +1.4s+1)(s+10)  The closed-loop transfer function is T (s) =  3s + 5 . s2 + s + K + 6  The bandwidth as a function of K is shown in Figure E8.15. The bandwidth as a function of K is: (a) K = 1 and ωb = 7.0 rad/sec. (b) K = 2 and ωb = 7.9 rad/sec. (c) K = 10 and ωb = 14.7 rad/sec. The bandwidth increases as K increases.  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  CHAPTER 8  Frequency Response Methods  24 22 20 18 ωb (rad/s)  392  16 14 12 10 8 6  0  FIGURE E8.15 Bandwith of T (s) =  2  4  3s+5 s2 +s+K+6 .  6  8  10 K  12  14  16  18  20  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  393  Problems  Problems (a) The transfer function is 1 , (1 + 0.25s)(1 + 3s)  Gc (s)G(s) = and  1 . (1 − 0.75ω 2 ) + j3.25ω  Gc (jω)G(jω) =  The polar plot is shown in Figure P8.1a. A summary of the magnitude and phase angles for ω = 0, 0.5, 1, 2, 5 and ∞ can be found in Table P8.1a. Nyquist Diagram 0.8 0.6 0.4 Imaginary Axis  P8.1  0.2 0 −0.2 −0.4 −0.6 −0.8 −1  −0.8  −0.6  FIGURE P8.1 (a) Polar plot for Gc (s)G(s) =  −0.4  −0.2  0 Real Axis  0.2  0.4  0.6  0.8  1  1 . (1+0.25s)(1+3s)  ω  0  0.5  1  2  5  ∞  |Gc (jω)G(jω)| (dB)  1.00  0.55  0.31  0.15  0.04  0  φ (deg)  0  -63.4  -85.6  -107.1  -137.51  -180  TABLE P8.1  (a) Magnitudes and phase angles for Gc (s)G(s) =  1 . (1+0.25s)(1+3s)  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  394  CHAPTER 8  Frequency Response Methods  (b) The transfer function is Gc (s)G(s) =  5(s2 + 1.4s + 1) (s − 1)2  and 5 (1 − ω 2 ) + 1.4jω Gc (jω)G(jω) = . (1 − ω 2 ) − 2jω   The polar plot is shown in Figure P8.1b. A summary of the magnitude and phase angles for ω = 0, 0.25, 0.5, 1, 2, 8, 16 and ∞ can be found in Table P8.1b. Nyquist Diagram 5 4 3  Imaginary Axis  2 1 0 −1 −2 −3 −4 −5 −4  −3  −2  FIGURE P8.1 CONTINUED: (b) Polar plot for Gc (s)G(s) =  −1  0  1 Real Axis  2  3  4  5  6  5(s2 +1.4s+1) . (s−1)2  ω  0  0.25  0.5  1  2  8  16  ∞  |Gc (jω)G(jω)| (dB)  5.00  4.71  4.10  3.50  4.10  4.92  4.98  5.00  φ (deg)  0  48.5  96.1  180.0  -96.2  -24.3  -12.2  0  TABLE P8.1  CONTINUED: (b) Magnitudes and phase angles for Gc (s)G(s) =  5(s2 +1.4s+1) . (s−1)2  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  395  Problems  (c) The transfer function is Gc (s)G(s) =  (s2  (s − 8 . + 6s + 8)  The polar plot is shown in Figure P8.1c. A summary of the magnitude and phase angles for ω = 0, 1, 2, 3, 4, 5, 6, ∞ can be found in Table P8.1c.  Nyquist Diagram 0.8 0.6  Imaginary Axis  0.4 0.2 0 −0.2 −0.4 −0.6 −0.8 −1  −0.8  −0.6  FIGURE P8.1 CONTINUED: (c) Polar plot for Gc (s)G(s) =  −0.4 −0.2 Real Axis  0  0.2  0.4  s−8 . s2 +6s+8  ω  0  1  2  3  4  5  6  ∞  |Gc (jω)G(jω)| (dB)  1.00  0.87  0.65  0.47  0.35  0.27  0.22  0.00  φ (deg)  180.0  132.3  94.4  66.3  45.0  28.5  15.3  -90.0  TABLE P8.1  CONTINUED: (c) Magnitudes and phase angles for Gc (s)G(s) =  s−8 . s2 +6s+8  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  396  CHAPTER 8  Frequency Response Methods  (d) The transfer function is Gc (s)G(s) =  20(s + 8) . s(s + 2)(s + 4)  The polar plot is shown in Figure P8.1d. A summary of the magnitude and phase angles for ω = 1, 0.1, 0.8, 1.6, 3.2, 12.8, ∞ can be found in Table P8.1d. Nyquist Diagram 20 15  Imaginary Axis  10 5 0 −5 −10 −15 −20 −20  −15  −10  −5  0  5  Real Axis  FIGURE P8.1 CONTINUED: (d) Polar plot for Gc (s)G(s) =  20(s+8) . s(s+2)(s+4)  ω  0  0.1  0.8  1.6  3.2  12.8  ∞  |Gc (jω)G(jω)| (dB)  ∞  199.70  22.87  9.24  2.79  0.14  0.00  φ (deg)  0  -93.6  -117.4  -139.1  -164.8  174.3  180.0  TABLE P8.1  CONTINUED: (d) Magnitudes and phase angles for Gc (s)G(s) =  20(s+8) . s(s+2)(s+4)  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  397  Problems  P8.2  (a) The Bode plot is shown in Figure P8.2a. A summary of the magnitude and phase angles for ω = 0.25, 0.5, 1, 2, 4, 8, 16 can be found in Table P8.2a.  Bode Diagram  Magnitude (dB)  0 −20 −40 −60  Phase (deg)  −80 0 −45 −90 −135 −180 −2 10  FIGURE P8.2 (a) Bode plot for Gc (s)G(s) =  −1  10  0  1  10 Frequency (rad/sec)  2  10  10  1 . (1+0.25s)(1+3s)  ω  0.25  0.5  1.0  2.0  4.0  8.0  16.0  |Gc (jω)G(jω)| (dB)  -1.95  -5.19  -10.26  -16.65  -24.62  -34.60  -45.93  φ (deg)  -40.5  -63.4  -85.6  -107.1  -130.2  -151.0  -164.8  TABLE P8.2  (a) Magnitudes and phase angles for Gc (s)G(s) =  1 . (1+0.25s)(1+3s)  (b) The transfer function is Gc (s)G(s) =  5(s2 + 1.4s + 1) (s − 1)2  The Bode plot is shown in Figure P8.2b. A summary of the magnitude  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  398  CHAPTER 8  Frequency Response Methods  and phase angles for ω = 0.25, 0.5, 1, 2, 4, 8, 16 can be found in Table P8.2b.  Bode Diagram  Magnitude (dB)  14 13 12 11  Phase (deg)  10 0 −90 −180 −270 −360 −1 10  0  1  10 Frequency (rad/sec)  FIGURE P8.2 CONTINUED: (b) Bode plot for Gc (s)G(s) =  10  5(s2 +1.4s+1) . (s−1)2  ω  0.25  0.5  1.0  2.0  4.0  8.0  16.0  |Gc (jω)G(jω)| (dB)  13.46  12.26  10.88  12.26  13.46  13.84  13.95  φ (deg)  48.5  96.2  180.0  -96.2  -48.5  -24.3  -12.2  TABLE P8.2  CONTINUED: (b) Magnitudes and phase angles for Gc (s)G(s) =  5(s2 +1.4s+1) . (s−1)2  (c) The transfer function is Gc (s)G(s) =  (s2  (s − 8) . + 6s + 8)  The Bode plot is shown in Figure P8.2c. A summary of the magnitude and phase angles for ω = 0.6, 1, 2, 3, 4, 5, 6, ∞ can be found in Table P8.2c.  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  399  Problems  Bode Diagram 0  Magnitude (dB)  −10 −20 −30 −40 −50 −60 180 Phase (deg)  135 90 45 0 −45 −90 −2 10  −1  0  10  1  2  10 10 Frequency (rad/sec)  FIGURE P8.2 CONTINUED: (c) Bode plot for Gc (s)G(s) =  3  10  10  s−8 s2 +6s+8 .  ω  0.6  1  2  3  4  5  6  ∞  |Gc (jω)G(jω)| (dB)  -0.45  -1.17  -3.72  -6.49  -9.03  -11.26  -13.18  -120.00  φ (deg)  150.5  132.3  94.4  66.3  45.0  28.5  15.3  -90.0  TABLE P8.2  CONTINUED: (c) Magnitudes and phase angles for Gc (s)G(s) =  s−8 s2 +6s+8 .  (d) A summary of the magnitude and phase angles for ω = 0.2, 0.8, 3.2, 6.4, 12.8, 25.6, 51.2 can be found in Table P8.2d. The Bode plot is shown in Figure P8.2d.  ω  0.2  0.8  3.2  6.4  12.8  25.6  51.2  |Gc (jω)G(jω)| (dB)  39.95  27.19  8.90  -3.98  -17.35  -30.0355  -42.28  φ (deg)  -97.1  -117.4  -164.8  178.0  174.2  176.0  177.8  TABLE P8.2  CONTINUED: (d) Magnitudes and phase angles for Gc (s)G(s) =  20(s+8) . s(s+2)(s+4)  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  400  CHAPTER 8  Frequency Response Methods  Bode Diagram 60  Magnitude (dB)  40 20 0 −20 −40  Phase (deg)  −60 −90  −135  −180  −225 −1 10  0  1  10  2  10  10  Frequency (rad/sec)  FIGURE P8.2 CONTINUED: (d) Bode plot for Gc (s)G(s) =  P8.3  20(s+8) . s(s+2)(s+4)  (a) The bridged-T network we found has zeros at s = ±jωn and poles at s=−  q ωn ± ωn 1/Q2 − 1 . Q  The frequency response is shown in Figure P8.3 for Q = 10. (b) For the twin-T network, we evaluate the magnitude at ω = 1.1ωn or 10% from the center frequency (see Example 8.4 in Dorf & Bishop). This yields |G| ≈ 2.1 ×    0.1 3.9    × 1.1 = 0.05 .  Similarly, for the bridged-T network |G| = 2.1 ×    0.1 2.1    × 0.14 = 0.707 .  The bridged-T network possesses a narrower band than the twin-T network.  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  401  Problems  0  Gain dB  -10 -20 -30 -40 10-1  100  101  w/wn  Phase deg  100 50 0 -50 -100 10-1  100  101  w/wn  FIGURE P8.3 Bode plot for G(s) =  2 s2 +ωn 2 , s2 +(2ωn /Q)s+ωn  where ζ = 1/Q = 0.1.  The transfer function is  P8.4  1 s 30000(2s + 1) = . s(s + 10)(s + 20)(s2 + 15s + 150)  G(s) = Gc G1 H(s)     A summary of the magnitude and phase angles can be found in Table P8.4. The Bode plot is shown in Figure P8.4.  ω  1  3  5  8  10  15  24  |G(jω| dB  6.95  5.78  5.08  3.38  1.59  -5.01  -17.56  −40.89o  −52.39o  −77.28o  −118.41o  −145.99o  −203.52o  −258.57o  φ(deg) TABLE P8.4  Magnitudes and phase angles for GH(s) =  30000(2s+1) . s(s+10)(s+20)(s2 +15s+150)  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  402  CHAPTER 8  Frequency Response Methods Bode Diagram  Magnitude (dB)  50 0 -50 -100  Phase (deg)  -150 0 -90 -180 -270 -360 -2 10  FIGURE P8.4 Bode plot for GH(s) =  -1  0  1  10 10 Frequency (rad/sec)  10  2  10  30000(2s+1) . s(s+10)(s+20)(s2 +15s+150)  The Bode plot is shown in Figure P8.5.  Bode Diagram 50  Magnitude (dB)  0 −50 −100 −150 −200 −250 0 Phase (deg)  P8.5  10  −90 −180 −270 −360 −2 10  FIGURE P8.5 Bode plot for G(s) =  −1  10  0  10  1  2  10 10 Frequency (rad/sec)  10 . (s/4+1)(s+1)(s/20+1)(s/80+1)  3  10  4  10  3  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  403  Problems  (a) The transfer function is GH(s) =  3.98(1 + s/1) . s(1 + s/10)2  We have a zero at ω = 1 and two poles at ω = 10.0. The low frequency approximation is K/s and at ω = 1 we have K 20 log ω     = 12dB .  Therefore, K = 3.98 at ω = 1 (an approximation). The phase plot is shown in Figure P8.6a.  (a) -40 -60  Phase deg  -80 -100 -120 -140 -160 -180 -2 10  10  -1  10  0  10  1  10  2  (b) 100  50 Phase deg  P8.6  0  -50  -100 -1 10  FIGURE P8.6 Phase plots for (a) G(s) =  10  0  3.98(s/1+1) . s(s/10+1)2  1  10 Frequency (rad/sec)  (b) G(s) =  10  2  10  3  s . (s/10+1)(s/50+1)  (b) The transfer function is GH(s) =  s . (1 + s/10)(1 + s/50)  The poles are located by noting that the slope is ±20 dB/dec. The  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  404  CHAPTER 8  Frequency Response Methods  low frequency approximation is Ks, so 20 log Kω = 0dB . At ω = 1 we determine that K=1. The phase plot is shown in Figure P8.6b. The loop transfer function is L(s) =  Kv . s(s/π + 1)2  (a) Set Kv = 2π. The Bode plot is shown in Figure P8.7a.  Gain dB  40  20  0  -20 10-1  100 Frequency (rad/sec)  101  -80 -100  Phase deg  P8.7  -120 -140 -160 -180 10-1  100  101  Frequency (rad/sec)  FIGURE P8.7 (a) Bode plot for L(s) =  Kv , s(s/π+1)  where Kv = 2π.  (b) The logarithmic magnitude versus the phase angle is shown in Figure P8.7b.  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  405  Problems  40  30  Gain dB  20  10  0  -10  -20 -170  -160  -150  -140  -130  -120  -110  -100  -90  Phase deg  FIGURE P8.7 CONTINUED: (b) Log-magnitude-phase curve for L(jω).  P8.8  The transfer function is T (s) =  s2  K . + 10s + K  (a) When P.O. = 15%, we determine that ζ = 0.517 by solving √ 2 15 = 100e−πζ/ 1−ζ . So, 2ζωn = 10 implies that ωn = 9.67, hence K = ωn2 = 93.53. Also, q  Mpω = (2ζ 1 − ζ 2 )−1 = 1.13 . (b) For second-order systems we have q  ωr = ωn 1 − 2ζ 2 = 6.59 when ζ = 0.517 and ωn = 9.67. (c) We estimate ωB to be ωB ≈ (−1.19ζ + 1.85)ωn = 11.94 rad/s . P8.9  The log-magnitude phase curves are shown in Figure P8.9.  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  406  CHAPTER 8  Frequency Response Methods  (a)  (b)  0  40  -5  30  -10 20  -20  10  Gain dB  Gain dB  -15  -25  0  -30 -10 -35 -20  -40 -45 -200  -150  -100  -50  -30 -180  0  -160  Phase deg  -120  -100  Phase deg  FIGURE P8.9 Log-magnitude-phase curve for (a) Gc (s)G(s) = 1+0.5s . s2  P8.10  -140  1 (1+0.5s)(1+2s)  and (b) Gc (s)G(s) =  The governing equations of motion are F (s) = Kf If (s) and If (s) =  Vf (s) . Rf + L f s  Without loss of generality we can let Kf = 1.0. Also, we have F (s) = (M s2 + bs + K)Y (s) . Therefore, the transfer function is GH(s) =  KKf 50K = . 2 (Rf + Lf s)(M s + bs + K) (s + 0.5)(s2 + 2s + 4)  This is a type 0 system, therefore Kp = 25K. (a) If we allow a 1% error , we have ess = |R|/(1 + Kp ) = 0.01|R|. Thus Kp = 25K = 99. Select K=4. (b) The Bode plot is shown in Figure P8.10a. (c) The log-magnitude phase curve is shown in Figure P8.10b.  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  407  Problems  Gain dB  40  20  0 -20 10-2  10-1  100  101  100  101  Frequency (rad/sec)  Phase deg  0  -100  -200 -300 10-2  10-1 Frequency (rad/sec)  FIGURE P8.10 (a) Bode plot for GH(s) =  200 . (s2 +2s+4)(s+0.5)  40  30  Gain dB  20  10  0  -10  -20 -300  -250  -200  -150  -100  -50  0  Phase deg  FIGURE P8.10 CONTINUED: (b) Log-magnitude-phase curve for GH(s) =  200 . (s2 +2s+4)(s+0.5)  (d) The closed-loop transfer function Bode plot is shown in Figure P8.10c. We determine from the plot that Mpω = 1.6, ωr = 4.4 and ωB = 6.8.  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  408  CHAPTER 8  Frequency Response Methods  5  Gain dB  0 -5 -10 -15 10-1  100  101  Frequency (rad/sec)  Phase deg  100 0  -100 -200 10-1  100  101  Frequency (rad/sec)  FIGURE P8.10 CONTINUED: (c) Bode plot for closed-loop T (s) = Y (s)/R(s).  The Bode plot is shown in Figure P8.11.  Gain dB  200 100 0  -100 10-4  10-3  10-2  10-1  100  101  100  101  Frequency (rad/sec) 100  Phase deg  P8.11  0  -100 -200 10-4  10-3  10-2  10-1  Frequency (rad/sec)  FIGURE P8.11 Bode plot for G(s) =  0.164(s+0.2)(−s+0.32) . s2 (s+0.25)(s−0.009)  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  409  Problems  P8.12  The three transfer functions are G1 (s) = 10  G2 (s) =  1 s(s/0.6 + 1)  G3 (s) = 3s .  (a) When G3 (s) is out of the loop, the characteristic equation is 10 =0 s(s/0.6 + 1) √ or s2 + 0.6s + 6 = 0. Thus, ζ = 0.6/(2 6) = 0.12. (b) With G3 (s), the characteristic equation is 1 + G1 G2 (s) = 1 +  1 + G1 G2 (s) + G2 G3 (s) = 1 +  1.85 6 + =0, s(s + 0.6) s(s + 0.6)  or s2 + 2.4s + 6 = 0 . √ Thus, ζ = 2.4/(2 6) = 0.49. P8.13  By inspection of the frequency response, we determine L(s) = Gc (s)G(s)H(s) =  K . s(s/100 + 1)(s/1000 + 1)2  For small ω, we have 20 log K/ω = 40 dB at ω = 10. So, K = 1000. P8.14  The data we have are R1 = R2 = 1000Ω, c1 = 10−7 farad and c2 = 10−6 farad. The governing equations are 1  V2 (s) C1 s = , V1 (s) R1 + C11 s and Vo (s) KR2 = . V2 (s) R2 + C12 s So Vo (s) KR2 C2 s 109 s = = . V1 (s) (R1 C1 s + 1)(R2 C2 s + 1) (s + 107 )(s + 1000) (a) The Bode plot is shown in Figure P8.14. (b) The mid-band gain is = 40 dB. (c) The -3 dB points are (rad/sec): ωlow ≈ 7  and ωhigh ≈ 1.5 × 109 .  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  410  CHAPTER 8  Frequency Response Methods  Gain dB  40  20  0  -20 100  101  102  103  104  105  106  107  108  109  1010  107  108  109  1010  Frequency (rad/sec)  Phase deg  100  0  -100  -200 100  101  102  103  104  105  106  Frequency (rad/sec)  FIGURE P8.14 Bode plot for G(s) =  P8.15  109 s . (s+107 )(s+103 )  The data are plotted in Figure P8.15, denoted by an asterisk (*). 50 * *  0  *  *  *  *  *  *  * * *  -50 -100 10-1  100  101  102  -50 -100 *  *  *  *  *  *  * *  -150  *  -200 -250 -300 10-1  FIGURE P8.15 Bode plot for G(s) =  *  100  809.7 ; s(s2 +6.35s+161.3)  101  *  102  tabular data is indicated by an asterick (*).  The low frequency slope is -20 dB/dec and the initial low frequency φ is  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  411  Problems  −90o , so we have an integrator of the form K/s. The initial phase is −90o and the final phase −270o , so we have a minimum phase G(s). Now, |G| is 0.97 at ω = 8 and ω = 10 indicating two complex poles. We postulate a transfer function of the form G(s) =  K s    s2 2 ωn  +  2ζs ωn   .  +1  The phase angle φ = −180o at ω = ωn . Then, from Figure 8.10 in Dorf & Bishop, we determine that ωn = 12.7. At ω = 8, ωωn = 0.63 and φ, due to the complex poles is −30o (subtract −90o due to the integrator). Again, from Figure 8.10 in Dorf & Bishop, we estimate ζ = 0.25. To determine K, note that when ωωn ≤ 0.1, the effect of the complex poles on magnitude is negligible, so at ω = 1 we have K |G| = 5.02 ∼ . = j1 So K = 5.02. Therefore, G(s) =  s  s2 161.3  +  0.5s 12.7   =  +1  809.7 . s(s2 + 6.35s + 161.3)  (a) The unit step input response is shown in Figure P8.16. The step  Step Response 1.4  1.2  1 Amplitude  P8.16  5.02   0.8  0.6  0.4  0.2  0  0  0.1  0.2  0.3  FIGURE P8.16 Unit step input response for T (s) =  0.4 0.5 0.6 Time (sec)  60.2 s2 +12.1s+60.2 .  0.7  0.8  0.9  1  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  412  CHAPTER 8  Frequency Response Methods  response is given by y(t) = 1 − e−6.05t (cos 4.85t + 1.25 sin 4.85t) . (b) The system bandwidth is ωB = 4.95 rad/sec. P8.17  The transfer function is Gc (s)G(s) =  4(0.5s + 1) . s(2s + 1)(s2 /64 + s/20 + 1)  (a) The Bode plot is shown in Figure P8.17.  Gain dB  50  0  -50 -100 10-1  100  101  102  101  102  Frequency (rad/sec) -50  Phase deg  -100 -150 -200 -250 -300 10-1  100 Frequency (rad/sec)  FIGURE P8.17 Bode plot for Gs (s)G(s) =  4(0.5s+1) . s(2s+1)(s2 /64+s/20+1)  (b) When the magnitude is 0 dB, we have ω1 = 1.6 rad/sec and when φ = −180o , we have ω2 = 7.7 rad/sec . P8.18  The transfer function is Gc (s)G(s) =  12(s + 0.5) 0.2(2s + 1) = . (s + 3)(s + 10) (s/3 + 1)(s/10 + 1)  The Bode plot is shown in Figure P8.18. Near 0 dB, the frequency is  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  413  Problems  ω = 5.4 rad/sec. 0  Gain dB  -5 -10 -15 -20 10-1  100  101  102  101  102  Frequency (rad/sec) 50  Phase deg  0 -50 -100 -150 -200 10-1  100 Frequency (rad/sec)  FIGURE P8.18 Bode plot for Gc (s)G(s) =  P8.19  12(s+0.5) s2 +13s+30 .  Examining the frequency response, we postulate a second-order transfer function θ(s) ωn2 = 2 . I(s) s + 2ζωn s + ωn2 From the data we see that φ = −90o at ω = 2. Using Figure 8.10 in Dorf & Bishop, we determine that ωn = ω = 2. We also estimate ζ = 0.4 from Figure 8.10. Thus, θ(s) 4 = 2 . I(s) s + 1.6s + 4  P8.20  The transfer function is Gc (s)G(s) =  823(s + 9.8) . + 22s + 471  s2  The Bode plot is shown in Figure P8.20. The maximum value of 20 log10 |Gc (jω)G(jω)| = 32.3 dB occurs at ω = 20.6 rad/sec and the corresponding phase is φ = −19.6o .  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  414  CHAPTER 8  Frequency Response Methods  Bode Diagram  Magnitude (dB)  35 30 25 20  Phase (deg)  15 45  0  −45  −90 −1 10  0  1  10  10  2  10  Frequency (rad/sec)  FIGURE P8.20 Bode plot for Gc (s)G(s) =  The Bode plot is shown in Figure P8.21. The gain is 24 dB when φ = −180o 40  Gain dB  20 0 -20 -40 10-1  100  101  102  101  102  Frequency (rad/sec) 0  Phase deg  P8.21  832(s+9.8) s2 +22s+471 .  -100 -200 -300 10-1  100 Frequency (rad/sec)  FIGURE P8.21 Bode plot for Gc (s)G(s) =  −200s2 . s3 +14s2 +44s+40  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  415  Problems  P8.22  The transfer function is G(s) =  P8.23  10000(s + 1)(s + 80) . s(s + 300)(s + 9000)  The transfer function is G(s) =  100(s + 20)(s + 8000) . (s + 1)(s + 80)(s + 500)  The system is type 0 and the steady-state error to a unit step input is ess =  1 = 0.0025 1 + Kp  since Kp = lim G(s) = 400 . s→0  P8.24  (a) From the Bode plot we see that 20 log10 Mpω = 12 or Mpω = 3.981. For a second-order system we know that Mpω = (2ζ  q  1 − ζ 2 )−1 .  Solving for ζ (with Mpω = 3.981) yields ζ = 0.12. Also, from the Bode plot, ωr = 0.9rad/sec . So, ωr ωn = p = 0.91 . 1 − 2ζ 2  Therefore, the second-order approximate transfer function is T (s) =  ωn2 0.83 = 2 . s2 + 2ζωn s + ωn2 s + 0.22s + 0.83  (b) The predicted overshoot and settling time are P.O. = 68% and Ts = 37 sec. P8.25  The transfer function is G(s) =  100(s + 10) . s2 (s + 100)  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  416  CHAPTER 8  P8.26  Frequency Response Methods  The transfer function is T (s) =  Vo (s) 1 + R1 /R2 = . V (s) 1 + RCs  Substituting R = 10kΩ, C = 1µF , R1 = 9kΩ, and R2 = 1kΩ yields T (s) =  10 . 1 + 0.01s  The frequency response is shown in Figure P8.26. Bode Diagrams  20  15  Phase (deg); Magnitude (dB)  10  5  0 0 -20 -40 -60 -80 -100 1 10  10  2  10  Frequency (rad/sec)  FIGURE P8.26 Bode plot for T (s) =  P8.27  1+R1 /R2 1+RCs  The frequency response is shown in Figure P8.27.  TABLE P8.27  K  0.75  1  10  |L(jω)|jω=0 , dB  3.52  12.04  26.02  ωb , rad/s  8.3  14.0  33.4  ωc , rad/s  3.5  8.7  22.9  System performance as K varies.  3  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  417  Problems  Bode Diagram 10  Magnitude (dB)  0  K increases  −10 −20 −30 −40  K decreases  −50 −60 0  Phase (deg)  Phase plot remains unchanged as K varies −45  −90  −135 −1 10  FIGURE P8.27 Bode plot for K = 1  10  0  1  10 Frequency (rad/sec)  10  2  10  3  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  418  CHAPTER 8  Frequency Response Methods  Advanced Problems AP8.1  The spring-mass-damper system is described by mẍ + bẋ + kx = p . Taking the Laplace transform (with zero initial conditions) yields X(s) 1 = . 2 P (s) ms + bs + k From Figure AP8.1(b) in Dorf & Bishop, we determine that 20 log  1 X(j0) = 20 log = −26dB . P (j0) k  Solving for k yields k = 19.96 N/m . Also, ωn2 = k/m implies m = k/ωn2 , where ωn = corner frequency = 3.2 rad/sec. So, m = 1.949 kg . Comparing Figure AP8.1(b) in Dorf & Bishop to the known standard Bode plot of a second-order system, we estimate ζ ≈ 0.32. Therefore, b = 2mζωn = 2(1.949)(0.32)(3.2) = 3.992 N − s/m . AP8.2  The closed-loop transfer function is T (s) =  Y (s) Kb = . R(s) s + 1 + 0.2Kb  WIth K = 5, we have T (s) =  5b . s+1+b  The sensitivity is SbT =  ∂T b s+1 = . ∂b T s+1+b  With the nominal value of b = 4, we have SbT =  s+1 . s+5  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  419  Advanced Problems  The sensitivity plot is shown in Figure AP8.2.  0  -2  20*log(mag) (dB)  -4  -6  -8  -10  -12  -14 10-1  100  101  102  Frequency (rad/sec)  FIGURE AP8.2 Bode plot for SbT (s) =  AP8.3  s+1 s+5 .  The equation of motion is mẍ + bẋ + Kx = bṙ + Kr . Taking Laplace transforms yields X(s) bs + K = . 2 R(s) ms + bs + K Then, given the various system parameters m = 1 kg, b = 4 Ns/m, K = 18 N/m, we obtain the transfer function: X(s) 4s + 18 = 2 . R(s) s + 4s + 18 √ p Also, ωn = corner frequency = K/m = 18 = 4.243 rad/s and ζ = damping ratio =  b/m 4 = = 0.471 . 2ωn 2(4.243)  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  420  CHAPTER 8  Frequency Response Methods  The Bode plot is shown in Figure AP8.3. 10  Gain dB  0 -10 -20 -30 10-1  100  101  102  101  102  Frequency (rad/sec)  Phase deg  0 -50 -100 -150 -200 10-1  100 Frequency (rad/sec)  FIGURE AP8.3 Bode plot for G(s) =  The Bode plot is shown in Figure AP8.4.  Bode Diagram  Magnitude (dB)  −20 −40 −60 −80 −100 −120 0 −45 Phase (deg)  AP8.4  4s+18 . s2 +4s+18  −90 −135 −180 −225 −270 −1 10  0  10  Frequency (rad/sec)  FIGURE AP8.4 Bode plot for L(s) =  1 . (0.4s+1)(s2 +3.9s+15)  1  10  2  10  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  421  Advanced Problems  The closed-loop transfer function with unity feedback is given by T (s) =  10(s + 1) Gc (s)G(s) = 2 . 1 + Gc (s)G(s) s + 9s + 10  (a) Solving for Gc (s)G(s) yields Gc (s)G(s) =  10(s + 1) . s(s − 1)  (b) A summary of the plot data (see Figure AP8.5) is presented in Table AP8.5. (c) The open-loop system is unstable; the closed-loop system is stable.  40  30  20  20 log|GcG(j ω)|, dB  AP8.5  10  0  −10  −20  −30  −40 100  120  140  160  180 200 Phase, degrees  220  240  260  280  FIGURE AP8.5 Log-magnitude-phase curve for Gc G(jω).  ω  1  10  50  110  500  20 log |Gc G|  40  4.85  -13.33  -20.61  -33.94  phase (deg)  101.42  250.17  267.53  268.93  269.77  TABLE AP8.5  Summary of magnitude and phase for ω = 1, 10, 50, 110, 500.  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  422  CHAPTER 8  The transfer function is given by T (s) =  1/m . + (b/m)s + (k/m)  s2  Selecting k = 1 and b = 2 results in the Bode plot magnitude always √ less than 0 dB. Choosing b = 2/2 leads to a peak response with a sinusoidal input at ω = 0.66 rad/s. Figure AP8.6a shows the Bode plot and Figure AP8.6b shows the response to a sinusiodal input with frequency ω = 1 rad/s is less than 1 in the steady-state, as desired. Bode Diagram 10 System: sys Peak gain (dB): 6.3 At frequency (rad/sec): 0.661  0  Magnitude (dB)  −10  −20  −30  −40  −50 −2 10  −1  0  10  1  10  10  Frequency (rad/sec)  Impulse Response 1  0.5  Amplitude  AP8.6  Frequency Response Methods  0  −0.5  −1  −1.5  0  100  200  300  400 Time (sec)  500  600  700  800  FIGURE AP8.6 (a) Bode plot for b/m = 1 and k/m = 1. (b) Response to a sinusiodal input.  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  423  Advanced Problems  The transfer function is G(s) =  Vo (s) 1 + R2 C 2 s = . Vi (s) 1 + R1 C 1 s  Substituting C1 = 0.1 µF ,C2 = 1 mF , R1 = 10 kΩ, and R2 = 10 Ω yields G(s) =  0.01s + 1 . 0.001s + 1  The frequency response is shown in Figure AP8.7.  Bode Diagram  Magnitude (dB)  20 15 10 5 0 60 Phase (deg)  AP8.7  30  0 0 10  FIGURE AP8.7 Bode plot for G(s) =  1  10  0.01s+1 0.001s+1  2  3  10 10 Frequency (rad/sec)  4  10  5  10  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  424  CHAPTER 8  Frequency Response Methods  Design Problems CDP8.1  With the PI controller in the loop, the closed-loop transfer function from the input to the output is 26.035K(s + 2) θ(s) = 2 , R(s) s + (33.1415 + 26.035K)s + 52.07K where we switch off the tachometer feedback (see Figure CDP4.1 in Dorf and Bishop). The Bode plot is shown below for K = 40. From the step response we determine that P.O. = 0 and Ts = 0.19. With K = 40, the closed-loop poles are both real roots with values of s1 = −1072.6 and s2 = −1.9. 60  Gain dB  40  20  0 -1 10  10  0  10  1  10  2  Frequency (rad/sec)  Phase deg  0  -30  -60  -90 10  -1  10  0  10  1  10  2  Frequency (rad/sec)  DP8.1  The loop transfer function is L(s) = Gc (s)G(s) =  K(s + 2) . s2 (s + 12)  (a,b) Let K = 1. The Bode plot of the loop transfer function and the closed-loop transfer functions are shown in Figure DP8.1a and Figure DP8.1b, respectively. (c) Let K = 50. The Bode plot of the loop transfer function and the closed-loop transfer functions are shown in Figure DP8.1c and Figure DP8.1d, respectively.  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  425  Design Problems 50  Gain dB  0  -50  -100 -1 10  10  0  10  1  10  2  Frequency (rad/sec)  Phase deg  -120  -140  -160  -180 -1 10  10  0  10  1  10  2  Frequency (rad/sec)  FIGURE DP8.1 (a) Bode plot for the loop transfer function Gc (s)G(s) =  (s+2) . s2 (s+12)  50  Gain dB  0  -50  -100 -2 10  10  -1  0  10 Frequency (rad/sec)  10  1  10  2  Phase deg  0  -90  -180 10  -2  10  -1  0  10 Frequency (rad/sec)  FIGURE DP8.1 CONTINUED: (b) Bode plot for the closed-loop T (s) =  10  1  (s+2) . s3 +12s2 +s+2  10  2  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  CHAPTER 8  Frequency Response Methods 100  Gain dB  50  0  -50 -1 10  10  0  10  1  10  2  Frequency (rad/sec)  Phase deg  -120  -140  -160  -180 -1 10  10  0  10  1  10  2  Frequency (rad/sec)  FIGURE DP8.1 CONTINUED: (c) Bode plot for the loop transfer function Gc (s)G(s) =  50(s+2) . s2 (s+12)  20  Gain dB  0 -20 -40 -60 -1 10  10  0  10  1  10  2  Frequency (rad/sec) 0  Phase deg  426  -90  -180 10  -1  10  0  10  1  10  Frequency (rad/sec)  FIGURE DP8.1 CONTINUED: (d) Bode plot for the closed-loop T (s) =  50(s+2) . s3 +12s2 +50s+100  2  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  427  Design Problems  (d) The peak value of Mp ≤ 2 occurs for 14 ≤ K ≤ 350. The maximum bandwidth is achieved for the largest gain K. Thus, we select K = 350 and the corresponding bandwidth is ωB = 29 rad/sec. (e) The system is type 2—the steady-state error is zero for a ramp input. The open-loop transfer function is 20(s + 1) . s(s + 4)(s2 + 2s + 8)  Gc (s)G(s) =  (a) The phase angle is φ = −180o when ω = 3.54 rad/sec. The magnitude is 0 dB when ω = 0.87 rad/sec. (b) The closed-loop transfer function is T (s) =  s4  +  6s3  20(s + 1) . + 16s2 + 52s + 20  The closed-loop Bode plot is shown in Figure DP8.2. Bode Diagram Gm = 6.71 dB (at 3.54 rad/sec) , Pm = 105 deg (at 0.869 rad/sec) 20  Magnitude (dB)  0 −20 −40 −60 −80 −100 −45 −90 Phase (deg)  DP8.2  −135 −180 −225 −270 −1 10  0  1  10  10  2  10  Frequency (rad/sec)  FIGURE DP8.2 Bode plot for closed-loop T (s) =  20(s+1) . s4 +6s3 +16s2 +52s+20  (c) When K = 22, we have Mpω = 4.84dB ,  ωr = 3.11 ,  and ωB = 3.78 rad/sec .  When K = 25, we have Mpω = 7.18 dB ,  ωr = 3.18 rad/sec ,  and ωB = 3.94 rad/sec .  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  428  CHAPTER 8  Frequency Response Methods  (d) Select K = 22. DP8.3  The closed-loop transfer function is T (s) =  s3  +  7s2  K(s + 5) . + 12s + 10 + 5K  When K = 4.2, we have 10 log 10 Mpω = 3 dB. The system bandwidth is ωb = 3.7178 rad/sec. The steady-state tracking error to a unit step input is ess = lim sE(s) = lim 1 − T (s) . s→0  s→0  So, ess = 1 −  5K = 0.322 , 10 + 5K  when K = 4.2. Since the system is unstable when K > 14.8, the steadystate error does not exist after K = 14.8. The Bode plot is shown in Figure DP8.3. 20  Gain dB  0 -20 -40 -60 -80 10-1  100  101  102  101  102  Frequency (rad/sec)  Phase deg  0 -50 -100 -150 -200 10-1  100 Frequency (rad/sec)  FIGURE DP8.3 Bode plot for closed-loop T (s) =  DP8.4  K(s+5) , s3 +7s2 +12s+10+5K  where K = 4.2.  We have a second-order loop transfer function Gc (s)G(s) =  K . (0.3s + 1)(0.6s + 1)  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  429  Design Problems  With Mpω = 1.5, we determine that q  Mpω = (2ζ 1 − ζ 2 )−1  or  ζ = 0.3568 .  Now the characteristic equation is s2 + 5s + 5.56(1 + K) = 0 . So, solving 2ζωn = 5 yields ωn = 7. Therefore, K = 0.18ωn2 − 1 = 7.82 . The closed-loop transfer function is T (s) =  K 5.56(K + 1) . 2 K + 1 s + 5s + 5.56(K + 1)  So, the overall gain of the standard second-order system will be attenuated by the factor K/(K + 1). To compensate, we amplify the gain by a small factor. Thus we choose K = 10. The bandwidth is ωb = 11.25 rad/sec and the peak magnitude is Mpω =1.5. DP8.5  From the Bode plot of G(s) we find that there exists two pnoles, at approximately ω = 1 rad/sec and ω = 10 rad/sec. Then, by examining the Bode plot we estimate G(s) =  10 . (s + 1)(s + 10)  We use a scale factor of 10 because at low frequency the Bode plot has magnitude 0 dB (or a DC gain of 1). With G(s) as above, we can utilize the controller Gc (s) =  500 s + 20  yielding a crossover ωc = 12.9 rad/sec and a magnitude of at least 25 dB for ω < 0.1 rad/sec. Figure DP8.5 shows the compensator Bode plot of Gc (s)G(s).  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  430  CHAPTER 8  Frequency Response Methods Bode Diagram  Phase (deg)  Magnitude (dB)  50  -50 -100 -150 0 -45 -90 -135 -180 -225 10  -2  10  -1  FIGURE DP8.5 Bode Diagram for G(s)Gc (s) =  DP8.6  ωc=12.9  25 dB  0  0  1  10 10 Frequency (rad/sec)  10  2  10  3  5000 . (s+1)(s+10)(s+20)  Let K = −1 to meet the steady-state tracking error requirement and p = 2ζ, where ζ = 0.69 to obtain a 5% overshoot. The system is given by ẋ = Ax + Bu where   A=  0  1  −1 −1.38     ,    B=  The characteristic polynomial is  −1 0     , and  C=    0 1    .  s2 + 1.38s + 1 = 0 . The associated damping ratio is ζ = 0.69 and the natural frequency is ωn = 1 rad/s. Using the approximation ωb = (−1.19ζ + 1.85)ωn we obtain ωb ≈ 1.028 rad/s. The Bode plot is shown in Figure DP8.6. The bandwidth is ωb = 1.023 rad/s.  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  431  Design Problems  Bode Diagram 20  Magnitude (dB)  0 −20 −40 −60 −80 0  Phase (deg)  −45  −90  −135  −180 −2 10  −1  0  10  10 Frequency (rad/sec)  1  10  2  10  FIGURE DP8.6 Bode diagram for K = −1 and p = 1.38.  DP8.7  A viable controler is Gc (s) = KP +  3.33 KI + KD s = 5.5 + + 3.5s. s s  The loop transfer function is Gc (s)G(s) =  10.5s2 + 16.5s + 10 s2 (s2 + 4s + 5)  and computing Ka yields Ka = lim s2 Gc (s)G(s) = s→0  10 = 2, 5  as desired. The phase margin is P.O. = 44.35◦ and the bandwidth is ωb = 4.5 rad/sec. The step response is shown in Figure DP8.7.  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  CHAPTER 8  Frequency Response Methods  Step Response 1.4 System: sys_cl Peak amplitude: 1.32 Overshoot (%): 32.1 At time (sec): 1.11  1.2  System: sys_cl Settling Time (sec): 3.93  1 Amplitude  432  0.8  0.6  0.4  0.2  0  0  1  2  3 4 Time (sec)  FIGURE DP8.7 Step response for KP = 5.5, KI = 3.33, and KD = 3.5.  5  6  7  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  433  Computer Problems  Computer Problems CP8.1  The m-file script and Bode plot are shown in Figure CP8.1. The script automatically computes Mpω and ωr . num=[25]; den=[1 1 25]; sys = tf(num,den); w=logspace(0,1,400); [mag,phase]=bode(sys,w); [y,l]=max(mag); mp=20*log10(y), wr=w(l) bode(sys,w); mp = 14.0228 wr = 4.9458  Bode Diagrams From: U(1) 15  10  5  Phase (deg); Magnitude (dB)  0  -5  - 10 0  To: Y(1)  - 50  - 100  - 150  - 200 0 10  10  1  Frequency (rad/sec)  FIGURE CP8.1 Generating a Bode plot with the bode function.  CP8.2  The m-file script to generate the Bode plots is shown in Figure CP8.2a. The Bode plots are presented in Figures CP8.2b-CP8.2e. The transfer functions are (a) : G(s) =  1000 ; (s + 10)(s + 100)  (b) : G(s) =  s + 100 ; (s + 2)(s + 25)  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  CHAPTER 8  Frequency Response Methods  (c) : G(s) =  s2  100 ; + 2s + 50  (d) : G(s) =  s−6 . (s + 3)(s2 + 12s + 50)  % Part (a) num=[1000]; den=conv([1 10],[1 100]); sys1=tf(num,den); sys = tf(num,den); figure(1), bode(sys1), grid % Part (b) num=[1 100]; den=conv([1 2],[1 25]); sys2=tf(num,den); sys = tf(num,den); figure(2), bode(sys2), grid % Part (c) num=[100]; den=[1 2 50]; sys3=tf(num,den); sys = tf(num,den); figure(3), bode(sys3), grid % Part (d) num=[1 -6]; den=conv([1 3],[1 12 50]); sys4=tf(num,den); sys = tf(sys); figure(4), bode(sys4), grid  FIGURE CP8.2 (a) Script to generate the four Bode plots.  Bode Diagram  Magnitude (dB)  0 −20 −40 −60 −80 −100 0 Phase (deg)  434  −45 −90 −135 −180 −1 10  0  10  1  2  10 10 Frequency (rad/sec)  FIGURE CP8.2 CONTINUED: (b) Bode plot for G(s) =  1000 . (s+10)(s+100)  3  10  4  10  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  435  Computer Problems  Bode Diagram  Magnitude (dB)  20 0 −20 −40 −60  Phase (deg)  −80 0  −45  −90  −135 −1 10  0  10  1  2  10 10 Frequency (rad/sec)  FIGURE CP8.2 CONTINUED: (c) Bode plot for G(s) =  3  4  10  10  s+100 . (s+2)(s+25) Bode Diagram  20  Magnitude (dB)  10 0 −10 −20 −30  Phase (deg)  −40 0 −45 −90 −135 −180 0 10  1  10 Frequency (rad/sec)  FIGURE CP8.2 CONTINUED: (d) Bode plot for G(s) =  100 s2 +2s+50 .  2  10  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  436  CHAPTER 8  Frequency Response Methods  Bode Diagram −20  Magnitude (dB)  −30 −40 −50 −60 −70  Phase (deg)  −80 180 90 0 −90 −180 −1 10  0  1  10  2  10  10  Frequency (rad/sec)  FIGURE CP8.2 CONTINUED: (e) Bode plot for G(s) =  The Bode plots are shown in Figure CP8.3(a-d) with the transfer functions listed in the caption. The crossover frequency for (a) is 17 rad/sec.  Bode Diagram 20  Magnitude (dB)  0 −20 −40 −60 −80 −100 0 Phase (deg)  CP8.3  s−6 . (s+3)(s2 +12s+50)  −45 −90 −135 −180 −1 10  FIGURE CP8.3 (a) Bode plot for G(s) =  0  10  1  2  10 10 Frequency (rad/sec)  2000 . (s+10)(s+100)  3  10  4  10  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  437  Computer Problems  The crossover frequency for (b) is 0.99 rad/sec. Bode Diagram 20  Magnitude (dB)  0 −20 −40 −60 −80 0  Phase (deg)  −45 −90 −135 −180 −225 −270 −2 10  −1  10  0  10 Frequency (rad/sec)  FIGURE CP8.3 CONTINUED: (b) Bode plot for G(s) =  1  2  10  10  100 . (s+1)(s2 +10s+2)  The crossover frequency for (c) is 70.7 rad/sec. Bode Diagram 40  Magnitude (dB)  30 20 10 0 −10 −20  Phase (deg)  −30 0  −45  −90  −135 −2 10  −1  10  0  1  10 10 Frequency (rad/sec)  FIGURE CP8.3 CONTINUED: (c) Bode plot for G(s) =  50(s+100) . (s+1)(s+50)  The crossover frequency for (d) is 3.1 rad/sec.  2  10  3  10  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  438  CHAPTER 8  Frequency Response Methods  Bode Diagram 20  Magnitude (dB)  10 0 −10 −20 −30  Phase (deg)  −40 0  −45  −90 −1 10  0  1  10  FIGURE CP8.3 CONTINUED: (d) Bode plot for G(s) =  3  10  4  10  100(s2 +14s+50) . (s+1)(s+2)(s+500)  The m-file script and Bode plot are shown in Figure CP8.4a and b. The bandwidth is ωb = 10 rad/sec.  Bandwidth=10.0394 rad/sec 10  Magnitude (dB)  0 −10 −20 −30 −40 −50 0 Phase (deg)  CP8.4  2  10 10 Frequency (rad/sec)  −45 −90 −135 −180 −1 10  0  10  Frequency (rad/sec)  FIGURE CP8.4 (a) Bode plot for T (s) =  54 s2 +6s+54 .  1  10  2  10  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  439  Computer Problems numg=[54]; deng=[1 6 0]; sys_o = tf(numg,deng); sys_cl = feedback(sys_o,[1]) wb=bandwidth(sys_cl) bode(sys_cl), grid titlename=strcat('Bandwidth= ', num2str(wb), ' rad/sec') title(titlename)  FIGURE CP8.4 CONTINUED: (b) M-file script to obtain the closed-loop Bode plot.  The Bode plot of the closed-loop system is shown in Figure CP8.5. The closed-loop transfer function is T (s) =  s2  100 . + 6s + 100  (a) From the Bode plot we determine that Mpω ≈ 5 dB and  ωr ≈ 9 rad/sec .  Bode Diagrams From: U(1) 20  0  - 20  - 40  Phase (deg); Magnitude (dB)  - 60  - 80 0  - 50  To: Y(1)  CP8.5  - 100  - 150  - 200 -1 10  10  0  10  1  Frequency (rad/sec)  FIGURE CP8.5 Closed-loop system Bode plot.  10  2  10  3  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  440  CHAPTER 8  Frequency Response Methods  (b) From Equations (8.36) and (8.37) in Dorf & Bishop, we find that ζ ≈ 0.28  and  ωr /ωn ≈ 0.92  which implies that ωn = ωr /0.92 = 9.8 rad/sec . (c) From T (s) we find that ωn = 10 rad/sec  and ζ = 0.3 .  The actual values and the estimated values compare very well. The open-loop and closed-loop Bode plots are shown in Figure CP8.6a and b. The open-loop and closed-loop transfers functions are Gc (s)G(s) =  25 s3 + 3s2 + 27s + 25  and T (s) =  Gc (s)G(s) 25 = 3 . 1 + Gc (s)G(s) s + 3s2 + 27s + 50  Loop transfer function; bode(syso)  Magnitude (dB)  0 −20 −40 −60 −80 −100 0 −45 Phase (deg)  CP8.6  −90 −135 −180 −225 −270 −2 10  −1  10  0  10 Frequency (rad/sec)  FIGURE CP8.6 (a) Open-loop system Bode plot for Gc (s)G(s) =  1  10  25 . s3 +3s2 +27s+25  2  10  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  441  Computer Problems  Closed−loop system; bode(syscl)  Magnitude (dB)  0 −20 −40 −60 −80 −100 0 Phase (deg)  −45 −90 −135 −180 −225 −270 −1 10  0  10  1  2  10  10  Frequency (rad/sec)  FIGURE CP8.6 CONTINUED: (b) Closed-loop system Bode plot T (s) =  CP8.7  25 . s3 +3s2 +27s+50  The m-file script and plot of ωb versus p are shown in Figure CP8.7a and b. p=[0:0.001:1]; w=logspace(-1,1,1000); n=length(p); for i=1:n num=[1]; den=[1 2*p(i) 0]; sys = tf(num,den); sys_cl = feedback(sys,[1]); [mag,phase,w]=bode(sys_cl,w); a=find(mag<0.707); wb(i)=w(a(1)); end plot(p,wb) xlabel('p'), ylabel('Bandwidth (rad/sec)')  FIGURE CP8.7 (a) M-file script to generate plot of ωb versus p.  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  442  CHAPTER 8  Frequency Response Methods  1.6 1.5 1.4  Bandwidth (rad/sec)  1.3 1.2 1.1 1 0.9 0.8 0.7 0.6  0  0.1  0.2  0.3  0.4  0.5 p  0.6  0.7  0.8  0.9  1  FIGURE CP8.7 CONTINUED: (b) Plot of ωb versus p.  CP8.8  The transfer function from Td (s) to θ(s) is θ(s)/Td (s) =  s3  +  10s2  −0.01(s + 10) . + (0.01K − 10.791)s − 107.91 + 0.05K  Using the final value theorem and Td (s) = 1/s, we determine that lim sθ(s) =  s→0  −0.1 . −107.91 + 0.05K  The design specifications require that |ess | < 0.1o . So, solving for K yields K > 3300 . We can select K = 3300 as the initial value of K for the design. The m-file script is shown in Figure CP8.8a. For the design shown, the final selection for the gain is K = 6000. The disturbance response is shown in Figure CP8.8b.  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  443  Computer Problems  Mb=100; Ms=10; L=1; g=9.81; a=5; b=10; % K=6000; % Final design value of K % numg=[-1/Mb/L]; deng=[1 0 -(Mb+Ms)*g/Mb/L]; sysg = tf(numg,deng); numc=-K*[1 a]; denc=[1 b]; sysc = tf(numc,denc); % % Part (a) % sys = feedback(sysg,sysc); w=logspace(0,1,400); bode(sys,w) [mag,phase]=bode(sys,w); [M,l]=max(mag); MpDb=20*log10(M)-20*log10(mag(1)) % Mpw in decibels wr=w(l) % Mpw and peak frequency % % Part (b) % % From Eqs. (8.35) and (8.37) Mpw=10^(MpDb/20);zeta=sqrt((1-sqrt(1-(1/Mpw^2)))/2); wn=wr/sqrt(1-2*zeta^2); ts=4/zeta/wn po=100*exp(-zeta*pi/sqrt(1-zeta^2)) % % Part (c) % t=[0:0.1:10]; [y,x]=step(sys,t); plot(t,y*180/pi) xlabel('time [sec]') ylabel('theta [deg]') grid  MpDb = 4.0003 wr = 4.7226  meets specs  ts = 2.23 po = 32.75  0 -0.005 -0.01  theta [deg]  -0.015 -0.02 -0.025 -0.03 -0.035 -0.04  0  1  2  3  4  5  6  7  time [sec]  FIGURE CP8.8 (a) Design script. (b) Disturbance response - meets all specs!  8  9  10  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  444  CHAPTER 8  A viable filter is G(s) = 0.7  (s + 1000)(s + 1) . (s + 100)(s + 10)  The Bode plot is shown in Figure CP8.9 Bode Diagram  Magnitude (dB)  20 15 10 5 0 -5 90 Phase (deg)  CP8.9  Frequency Response Methods  45 0 -45 -90 -2 10  10  0  10  2  Frequency (rad/sec)  FIGURE CP8.9 (s+1000)(s+1) Bode plot for G(s) = 0.7 (s+100)(s+10)  .  10  4  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  C H A P T E R  9  Stability in the Frequency Domain  Exercises The Bode plot for the transfer function Gc (s)G(s) is shown in Figure E9.1, where Gc (s)G(s) =  2(1 + s/10) . s(1 + 5s)(1 + s/9 + s2 /81)  The gain and phase margins are P.M. = 17.5o .  G.M. = 26.2 dB and  Bode Diagram Gm = 26.2 dB (at 2.99 rad/sec) , Pm = 17.5 deg (at 0.618 rad/sec)  Magnitude (dB)  50 0 −50 −100 −150 −90 −135 Phase (deg)  E9.1  −180 −225 −270 −315 −2 10  FIGURE E9.1 Bode Diagram for Gc (s)G(s) =  −1  10  0  10 Frequency (rad/sec)  1  10  2  10  2(1+s/10) . s(1+5s)(1+s/9+s2 /81)  445  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  446  CHAPTER 9  E9.2  Stability in the Frequency Domain  The loop transfer function is Gc (s)G(s) =  10.5(1 + s/5) . s(1 + s/2)(1 + s/10)  The Bode plot is shown in Figure E9.2. The phase margin is P.M. = 40.4o at ωc = 4.96 rad/sec. Bode Diagram Gm = Inf dB (at Inf rad/sec) , Pm = 40.4 deg (at 4.96 rad/sec)  Magnitude (dB)  50  0  −50  Phase (deg)  −100 −90  −135  −180 −1 10  0  10  FIGURE E9.2 Bode Diagram for Gc (s)G(s) =  1  10 Frequency (rad/sec)  2  10  3  10  10.5(1+s/5) . s(1+s/2)(1+s/10)  E9.3  The phase margin P.M. ≈ 75o at 200 kHz. We estimate the −180o phase angle at 2 MHz, so the gain margin is G.M. ≈ 25 dB.  E9.4  The loop transfer function is Gc (s)G(s) =  100 . s(s + 10)  The Nichols diagram is shown in Figure E9.4. When the gain is raised by 4.6 dB, Mpω = 3 and the resonant frequency is ωR = 11 rad/sec.  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  447  Exercises  40  0 0.25  30 0.5 1  20 3 6  10  Gain dB  -1 -3 -6  0 -10  K=171 ------  -12  ------ K=100  -20  -20 -30 -40  -350  -300  -250  -200  -150  -100  -40 0  -50  Phase (deg)  FIGURE E9.4 Nichols Diagram for Gc (s)G(s) =  K) , s(s+10)  where K = 100 and K = 171.  E9.5  (a) The G.M. ≈ 5 dB and the P.M. ≈ 10o . (b) Lower the gain by 10 dB to obtain P.M. ≈ 60o .  E9.6  The Bode plot of the closed-loop transfer function is shown in Figure E9.6. The value of Mpω = 3 dB. The phase margin is P.M. = 40o when K = 50.  5 0 -5 -10  Gain dB  -15 -20 -25 -30 -35 -40 -45 10-1  100  101 Frequency (rad/sec)  FIGURE E9.6 Closed-loop Bode Diagram for T (s) =  50(s+100) s3 +50s2 +450s+5000 .  102  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  448  CHAPTER 9  E9.7  Stability in the Frequency Domain  The Nyquist plot is shown in Figure E9.7 for K = 5; the plot is a circle with diameter= K/5. For K > 5, we have P = 1 and N = −1 (ccw as Nyquist Diagram 0.5 0.4 0.3  Imaginary Axis  0.2 0.1 0 −0.1 −0.2 −0.3 −0.4 −0.5 −1  −0.8  −0.6  FIGURE E9.7 Nyquist Diagram for Gc (s)G(s) =  −0.4 −0.2 Real Axis  K s−5 ,  0  0.2  0.4  where K = 5.  shown). So Z = N + P = −1 + 1 = 0 and the system is stable for K > 5. (a) When K = 4, the G.M. = 3.5 dB. This is illustrated in Figure E9.8. Bode Diagram Gm = 3.52 dB (at 1.41 rad/sec) , Pm = 11.4 deg (at 1.14 rad/sec)  Magnitude (dB)  50  0  −50  −100  −150 −90  −135 Phase (deg)  E9.8  −180  −225  −270 −1 10  0  1  10  10 Frequency (rad/sec)  FIGURE E9.8 Bode Diagram for Gc (s)G(s) =  K , s(s+1)(s+2)  where K = 4.  2  10  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  449  Exercises  (b) The new gain should be K = 1 for a gain margin G.M. = 16 dB. E9.9  For K = 5, the phase margin P.M. = 5o as shown in Figure E9.9. Bode Diagram Gm = 1.58 dB (at 1.41 rad/sec) , Pm = 5.02 deg (at 1.29 rad/sec) 100 Magnitude (dB)  50 0 -50 -100 -150 -90  Phase (deg)  -135 -180 -225 -270 -2 10  10  FIGURE E9.9 Bode Diagram for Gc (s)G(s) =  0  10 Frequency (rad/sec)  K , s(s+1)(s+2)  10  1  10  2  where K = 5.  The Bode plot is shown in Figure E9.10a. The closed-loop frequency 100  Gain dB  50 0 GM=12.35 dB  -50 -100 10-2  10-1  100 Frequency (rad/sec)  101  102  0  Phase deg  E9.10  -1  -100 PM=23.14 deg -200 -300 10-2  10-1  FIGURE E9.10 (a) Bode Diagram for Gc (s)G(s) =  100 Frequency (rad/sec)  101  326s+1304 s4 +14.76s3 +151.3s2 +23.84s .  102  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  450  CHAPTER 9  Stability in the Frequency Domain  10 0 -10  Gain dB  -20 -30 -40 -50 -60 -70 10-1  100  101  102  Frequency (rad/sec)  FIGURE E9.10 CONTINUED: (b) Closed-loop frequency response: ωB = 6 rad/sec.  response is shown in Figure E9.10b. The bandwidth is ωB = 6 rad/sec. The Bode plot is shown in Figure E9.11. The system is stable.  Bode Diagram Gm = 3.91 dB (at 3.74 rad/sec) , Pm = 14.4 deg (at 2.76 rad/sec)  Magnitude (dB)  100  50  0  −50  −100 −90  −135 Phase (deg)  E9.11  −180  −225  −270 −2 10  −1  10  FIGURE E9.11 Bode Diagram for Gc (s)G(s) =  0  10 Frequency (rad/sec)  10(1+0.4s) . s(1+2s)(1+0.24s+0.04s2 )  1  10  2  10  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  451  Exercises  E9.12  We select the gain K = 10 to meet the 10% steady-state tracking error specification for a ramp input. The Bode plot and Nichols chart are shown in Figures E9.12a and E9.12b, respectively.  50  Gain dB  0 GM=14.82 dB -50 -100 -150 10-1  100  101 Frequency (rad/sec)  102  103  102  103  Phase deg  0 -100 -200  PM=31.79 deg  -300 10-1  100  101 Frequency (rad/sec)  40  0 0.25  30 0.5 1  20 10  Gain dB  -1  3 6 8  -3 -6  0  -12  -10  -20  -20 -30 -40  -350  -300  -250  -200  -150  -100  -50  -40 0  Phase (deg)  FIGURE E9.12 (a) Bode Diagram for Gc (s)G(s) = 10 . s(0.02s+1)(0.2s+1)  E9.13  10 . s(0.02s+1)(0.2s+1)  (b) Nichols chart for Gc (s)G(s) =  (a) The Nichols diagram is shown in Figure E9.13a and Mpω = 7.97 dB. (b) The closed-loop Bode plot is shown in Figure E9.13b. The bandwidth ωB = 18.65 rad/sec and the resonant frequency is ωr = 11.69 rad/sec.  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  452  CHAPTER 9  Stability in the Frequency Domain  40  0 0.25  30 0.5 1  20  Gain dB  10 0  -1  3 6  -3  8  -6 -12  -10  -20  -20 -30 -40  -350  -300  -250  -200  -150  -100  -50  -40 0  Phase (deg) 10  Gain dB  0 -10 -20 -30 -40 10-1  100  101  102  101  102  Frequency (rad/sec)  Phase deg  0 -50 -100 -150 -200 10-1  100 Frequency (rad/sec)  FIGURE E9.13 (a) Nichols Diagram for Gc (s)G(s) = 150 . s2 +5s+150  150 . s(s+5)  (b) Closed-loop Bode Diagram for T (s) =  (c) From Mpω = 8 dB, we estimate ζ = 0.2, so the expected P.O. = 52%. E9.14  (a) The peak resonance Mpω = 6 dB. (b) The resonant frequency is ωr = ω2 = 3 rad/sec. (c) The bandwidth is ωB = ω4 = 10 rad/sec. (d) The phase margin is P.M. = 30o .  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  453  Exercises  E9.15  The loop transfer function is Gc (s)G(s) =  100 , s(s + 20)  and the closed-loop transfer function is T (s) =  100 . s2 + 20s + 100  The magnitude plot for the closed-loop system is shown in Figure E9.15. With bandwidth defined as frequency at which the magnitude is reduced  Bode Diagram 0  −1  Magnitude (dB)  −2  −3  −4  −5  −6  −7 −1 10  0  10 Frequency (rad/sec)  FIGURE E9.15 Magnitude plot for the closed-loop T (s) =  1  10  100 . s2 +20s+100  -3 dB from the dc value, we determine that ωB = 6.4 rad/sec. E9.16  The transfer function of the approximation is G(jω) =  1 − jω/10 , 1 + jω/10  and the magnitude is |G(jω)| =  1 − jω/10 =1, 1 + jω/10  which is equivalent to the actual time delay magnitude. The phase ap-  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  454  CHAPTER 9  Stability in the Frequency Domain  proximation is φ = − tan−1 ω/10 + tan−1 (−ω/10) = −2 tan−1 ω/10 and the actual phase is φ = −0.2ω . The phase plots are shown in Figure E9.16. The approximation is accurate for ω < 3 rad/sec.  Actual _______ & Approximation −−−−−−− 0  −20  Phase deg  −40  −60  −80  −100  −120 −2 10  −1  0  10  1  10  10  Frequency (rad/sec)  FIGURE E9.16 Phase plots for time delay actual vs approximation.  E9.17  (a,b) The phase angle for P.M. = 30 is φ = −90o + tan−1  ω 2ω − tan−1 = −150o . 2 15 − ω 2  Solving for ω yields ω = 4.7. Then, at ω = 4.7, we have K = 10.82 when 1  |Gc G(jω)| =  K(ω 2 + 4) 2 1  ω((2ω 2 )2 + (15 − ω 2 )2 ) 2  The Bode plot is shown in Figure E9.17.  =1.  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  455  Exercises Bode Diagrams Gm=3.5545 dB (at 4.3301 rad/sec), Pm=40 deg. (at 3.5147 rad/sec) 50  Phase (deg); Magnitude (dB)  0  - 50  - 100 - 50  - 100  - 150  - 200  - 250 -1 10  10  0  10  1  10  2  Frequency (rad/sec)  FIGURE E9.17 Bode Diagram for Gc (s)G(s) =  K(s+2) , s3 +2s2 +15s  where K = 10.82.  (c) The steady-state error for a ramp is ess =  A A = 10K = 0.60A , Kv 15  where R(s) = A/s2 . E9.18  (a) The gain crossover is at ωc = 486 Hz, and the phase margin P.O. = 36.2o . So, ζ ≈ 0.36. Then, the expected percent overshoot to a step input is √ 2 P.O. = 100e−ζπ/ 1−ζ = 30% , where ζ = 0.36 . (b) The estimated bandwidth is ωB ≈ 2π(600). (c) Approximate  ωn ≈ ωr = 2π(480) . Then, Ts =  4 4 = ≈ 4 ms . ζωn (0.36)2π(480)  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  456  CHAPTER 9  E9.19  Stability in the Frequency Domain  The Bode plot is shown in Figure E9.19 for K = 16.75. The phase and gain margins are P M = 50.0o and GM = 2.72 dB. Gm=2.7233 dB (at 20.618 ad/sec), r Pm=50 deg . (at 13.434ad/sec) r 10  0  - 10  Phase (deg); Mag nitude (dB)  - 20  - 30  - 40 0  - 100  - 200  - 300  - 400  - 500 0 10  1  2  10  3  10  10  Frequency (rad/sec)  FIGURE E9.19 −0.1s Bode Diagram for Gc (s)G(s) = K es+10 , where K = 16.75.  The system response for both drivers is shown in Figure E9.20. T=1 sec (solid line) & T=1.5 sec (dashed line) 1 0 -1  Automobile velocity change  E9.20  -2 -3 -4 -5 -6 -7 -8  0  1  2  3  4  5  6  7  Time (sec)  FIGURE E9.20 Change in automobile velocity due to braking for two drivers.  8  9  10  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  457  Exercises  E9.21  The Bode plot is shown in Figure E9.21. 50  Gain dB  0 -50  GM=12.04 dB  -100 -150 10-1  100  101 Frequency (rad/sec)  102  103  102  103  Phase deg  0 -100 -200  PM=16.85 deg  -300 10-1  100  101 Frequency (rad/sec)  FIGURE E9.21 Bode Diagram for Gc (s)G(s) =  1300 . s(s+2)(s+50)  E9.22  When K = 10, the P.M. = 36.9o ; the system is stable. Decreasing the gain to K = 4 results in a P.M. = 60o .  E9.23  The Nichols chart is shown in Figure E9.23.  40  0 0.25  30 0.5 1  20 10  Gain dB  -1  3 6 8  -3 -6  0  -12  -10  -20  -20 -30 -40  -350  -300  -250  -200  -150  Phase (deg)  FIGURE E9.23 Nichols chart for Gc (s)G(s) =  438 . s(s+2)(s+50)  -100  -50  -40 0  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  458  CHAPTER 9  Stability in the Frequency Domain  The actual values are Mpω = 1.6598 (4.4 dB) ωr = 2.4228 rad/sec ωB = 4.5834 rad/sec . E9.24  Using the Nyquist criterion, we have P = 1 and N = 0 which implies Z = N +P = 1 . Therefore, the system has one root in the right half-plane.  E9.25  The Bode plot is shown in Figure E9.25. PM=27.73 deg at wc=8.29 rad/sec  Gain dB  50  0  -50 -100 10-1  100  101  102  101  102  Frequency (rad/sec)  Phase deg  0 -100 -200 -300 10-1  100 Frequency (rad/sec)  FIGURE E9.25 Bode plot for Gc (s)G(s) =  E9.26  11.7 . s(0.05s+1)(0.1s+1)  The Nichols chart for Gc (s)G(s) =  11.7 s(0.05s + 1)(0.1s + 1)  is shown in Figure E9.26, where we find that Mpω = 6.76 dB ωr = 8.96 rad/sec ωB = 13.73 rad/sec .  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  459  Exercises  40  0 0.25  30 0.5 1  20 3 6 8  10  Gain dB  -1 -3 -6  0  -12  -10  -20  -20 -30 -40  -350  -300  -250  -200  -150  -100  -50  -40 0  Phase (deg)  FIGURE E9.26 Nichols chart for Gc (s)G(s) =  The Bode plot for G(s) with K = 122.62 is shown in Figure E9.27.  K=122.63  Gm=10.938 dB (at 6 rad/sec), Pm=40 deg. (at 2.7978 rad/sec)  50  0  Phase (deg); Magnitude (dB)  E9.27  11.7 . s(0.05s+1)(0.1s+1)  - 50  - 100 - 50  - 100  - 150  - 200  - 250  - 300 -1 10  10  0  10  Frequency (rad/sec)  FIGURE E9.27 Bode plot for Gc (s)G(s) =  K , s(s+6)2  with K = 122.62.  1  10  2  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  460  CHAPTER 9  Stability in the Frequency Domain  The phase margin is P.M. = 40.0o and the gain margin is G.M. = 10.94 dB . E9.28  The phase margin is P.M. = 28o . The estimated damping is ζ=  P.M. = 0.28 . 100  The estimated percent overshoot is √ 2 P.O. = 100e−πζ/ 1−ζ = 40% . The actual overshoot is P.O. = 44.43%. E9.29  The F (s)-plane contour is shown in Figure E9.29, where F (s) = 1 + Gc (s)G(s) =  s+3 . s+2  F(s)-plane 0.6 *  0.4  0.2  *  Im  *  0  *  *  *  -0.2  *  -0.4 *  -0.6  1  1.1  1.2  1.3  1.4  1.5  1.6  1.7  Re  FIGURE E9.29 F (s)-plane contour, where F (s) = 1 + Gc (s)G(s) =  E9.30  The Bode plot is shown in Figure E9.30.  s+3 s+2 .  1.8  1.9  2  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  461  Exercises Bode Diagram  Magnitude (dB)  50 0  -50  Phase (deg)  -100 0 -45 -90 -135 -180 10  -2  10  0  10  2  10  4  Frequency (rad/sec)  FIGURE E9.30 Bode plot for G(s) = C [sI − A]−1 B + D =  The Bode plot is shown in Figure E9.31. The phase margin is P.M. = 50.6 deg. Bode Diagram Gm = Inf , Pm = 50.6 deg (at 0.341 rad/sec)  Magnitude (dB)  80 60 40 20 0 -20 -40 -90 Phase (deg)  E9.31  1000 . s2 +100s+10  -120  -150 -3 10  10  -2  10  -1  Frequency (rad/sec)  FIGURE E9.31 Bode plot for L(s) = G(s)H(s) =  2s+1 10s2 +s .  10  0  10  1  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  462  CHAPTER 9  E9.32  Stability in the Frequency Domain  The Bode plot is shown in Figure E9.32. The phase margin is P.M. = 29◦ . Bode Diagram Gm = Inf dB (at Inf rad/sec) , Pm = 29 deg (at 3.1 rad/sec)  Magnitude (dB)  20 0 −20 −40 −60  Phase (deg)  −80 0 −45 −90 −135 −180 −1 10  0  1  10  2  10  10  Frequency (rad/sec)  FIGURE E9.32 Bode plot for G(s) = C [sI − A]−1 B + D =  The Bode plot is shown in Figure E9.33. The phase margin is P.M. = 17.7◦ and the gain margin is G.M. = 5.45 dB. Bode Diagram Gm = 5.45 dB (at 5.68 rad/sec) , Pm = 17.7 deg (at 4.24 rad/sec)  Magnitude (dB)  50  0  −50  −100  −150 0 −45 Phase (deg)  E9.33  6.4 s2 +s+4 .  −90 −135 −180 −225 −270 −1 10  FIGURE E9.33 Bode plot for L(s) =  0  10  1  10 Frequency (rad/sec)  200 . (s2 +2.83s+4)(s+10)  2  10  3  10  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  463  Problems  Problems P9.1  (a) The loop transfer function is Gc (s)G(s) =  1 . (1 + 0.5s)(1 + 2s)  P = 0, N = 0; therefore Z = N +P = 0. The system is stable. (Note: See P8.1 for the polar plots.) (b) The loop transfer function is 1 + 0.5s . s2 P = 0, N = 0, therefore Z = N + P = 0. The system is stable. (c) The loop transfer function is s2  s+4 . + 5s + 25  P = 0, N = 0, Z = N + P = 0. Therefore, the system is stable. (d) The loop transfer function is 30(s + 8) . s(s + 2)(s + 4) P = 0, N = 2 therefore Z = P + N = 2. Therefore, the system has two roots in the right half-plane, and is unstable. P9.2  (a) The loop transfer function is Gc (s)G(s) =  s(s2  K , + s + 6)  and Gc (jω)G(jω) =  K K[−ω 2 − jω(6 − ω 2 )] == . jω(−ω 2 + jω + 6) [(6 − ω 2 )2 ω 2 + ω 4 ]  To determine the real axis crossing, we let Im{Gc (jω)G(jω)} = 0 = −Kω(6 − ω 2 ) or ω=  √  6.  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  464  CHAPTER 9  Stability in the Frequency Domain  Then, Re{Gc (jω)G(jω)}ω=√ 6 =  −Kω 2 ω4  √ ω= 6  =  −K . 6  So, −K/6 > −1 for stability. Thus K < 6 for a stable system.  (b) The loop transfer function is  Gc (s)G(s) =  K(s + 1) . s2 (s + 6)  The polar plot never encircles the -1 point, so the system is stable for all gains K (See Figure 10 in Table 9.6 in Dorf & Bishop). P9.3  (a,b) The suitable contours are shown in Figure P9.3. jw  jw  Gs  Gs  q =cos z q  r  r approaches infinity  r  s  -s 1  r approaches infinity s  (b)  (a)  FIGURE P9.3 Suitable contours Γs for (a) and (b).  (c) Rewrite the characteristic equation as 1+  96 =0. s(s2 + 11s + 56)  In this case, −σ1 = −1. Therefore, we have one pole inside the contour at s = 0, so P = 1. The polar plot yields N = −1, so Z = N + P = 0. Therefore, all three roots have real parts less than -1. In fact, the roots are s1 = −3, and s2,3 = −4 ± j4. P9.4  (a) P = 0, N = 2, therefore Z = 2. The system has two roots in the right hand s-plane. (b) In this case, N = +1 − 1 = 0, so Z = 0. Therefore the system is stable.  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  465  Problems  P9.5  (a) The loop transfer function is L(s) = Gc (s)G(s)H(s) =  K . (s + 1)(3s + 1)(0.4s + 1)  The steady-state error is ess =  |R| . 1+K  We require ess = 0.1|R|, so K > 9. (b) Use K = 9. The Nyquist plot is shown in Figure P9.5. We determine that P = 0 and N = 0. Therefore, Z = 0 and the system is stable.  8 6 4  Imag Axis  2 0 -2 -4 -6 -8 -2  0  2  4  6  8  10  Real Axis  FIGURE P9.5 Nyquist Diagram for L(s) = Gc (s)G(s)H(s) =  9 . (s+1)(3s+1)(0.4s+1)  (c) The phase and gain margins are P.M. = 18o and G.M. = 5 dB. P9.6  The rotational velocity transfer function is ω(s) = G(s) =  R(s) 1+  K s 3.7(2π)    s 68(2π)+1   .  At low frequency, we have the magnitude near 35 dB, so 20 log K = 35 dB  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  466  CHAPTER 9  Stability in the Frequency Domain  and K = 56. Since the frequency response plot is for rotational velocity ω(s), and we are interested in position control, we add an integrator. The characteristic equation is 1 56(23)(427) 1 + G(s) = 1 + =0. s s(s + 23)(s + 427) The roots are s1 = −430  and s2,3 = −10 ± j35 .  Thus, ωn = 36 and ζ = 0.28. The time constant of the closed-loop system is τ= The loop transfer function is L(s) = Gc (s)G(s)H(s) =  10K1 s(s + 7) . (s + 3)(s2 + 0.36)  (a) The Bode plot is shown in Figure P9.7 for K1 = 2.  Gain dB  100 50 0 -50 10-1  100  101  102  101  102  Frequency (rad/sec) 100 50  Phase deg  P9.7  1 = 99.6 msec . ζωn  0 -50 -100 -150 10-1  100 Frequency (rad/sec)  FIGURE P9.7 Bode Diagram for Gc (s)G(s)H(s) =  10K1 s(s+7) , (s+3)(s2 +0.36)  where K1 = 2.  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  467  Problems  (b) The phase margin P.M. = 80o and the gain margin G.M. = ∞, since φ never crosses = −180o . (c) The transfer function from Td (s) to θ(s) is θ(s) =  G(s) Td (s) . 1 + Gc (s)G(s)H(s)  Then, for a step disturbance θ(∞) = lims→0 sθ(s) = G(0) = 10/0.36 = 27.8, since H(0) = 0. (d) The system is so highly damped, there is very little resonant peak. (e) The estimated ζ = P.M./100 = 0.80. The actual ζ = 0.97. (a) The loop transfer function is Gc (s)G(s)H(s) =    s2 ω12  +  (0.02s + 1)  2ζ1 s ω1    s2 ω22    +1  +  2ζ2 s ω2   ,  +1  where ω1 = 20π = 62.8, ω2 = 14π = 43.9, ζ1 = 0.05 and ζ2 = 0.05. The Bode plot is shown in Figure P9.8a. The phase margin is P.M. = −9o . Therefore, the system is unstable.  Gain dB  20  0  -20 -40 100  101  102  103  102  103  Frequency (rad/sec) 0  Phase deg  P9.8  -50 -100 -150 -200 100  101 Frequency (rad/sec)  FIGURE P9.8 (a) Bode Diagram for Gc (s)G(s)H(s) = and ω2 = 14π.  s2 /ω12 +(0.1/ω1 )s+1 , (0.02s+1)(s2 /ω22 +(0.1/ω2 )s+1)  where ω1 = 20π  (b) In this case ζ2 = 0.25, with all other parameters the same as before.  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  468  CHAPTER 9  Stability in the Frequency Domain  10  Gain dB  0 -10 -20 -30 -40 100  101  102  103  102  103  Frequency (rad/sec)  Phase deg  0 -50 -100 -150 -200 100  101 Frequency (rad/sec)  FIGURE P9.8 CONTINUED: (b) Bode Diagram for Gc (s)G(s)H(s) = where ω1 = 20π and ω2 = 14π.  s2 /ω12 +(0.1/ω1 )s+1 , (0.02s+1)(s2 /ω22 +(0.5/ω2 )s+1)  The Bode plot is shown in Figure P9.8b. The phase margin is P.M. = 86o . Therefore, the system is now stable. P9.9  (a) The Bode plot is shown in Figure P9.9a The phase margin is P.M. = 83o and the gain margin is G.M. = ∞. (b) With the compensator, the loop transfer function is Gc (s)G(s)H(s) = K1  0.30(s + 0.05)(s2 + 1600)(s + 0.5) , s(s2 + 0.05s + 16)(s + 70)  where K2 /K1 = 0.5 . Let K1 = 1. The Bode plot is shown in Figure P9.9b. The phase margin is P.M. = 80o and the gain margin is G.M. = ∞, essentially the same as in (a). But the system in (b) is a type one, so that ess = 0 to a step input or disturbance. We cannot achieve a G.M. = 10 dB by increasing or decreasing K1 .  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  469  Problems  40  Gain dB  20 0 -20 -40 -60 10-3  10-2  10-1  100  101  102  103  101  102  103  102  103  102  103  Frequency (rad/sec)  Phase deg  0 -100 -200 -300 10-3  10-2  10-1  100 Frequency (rad/sec)  FIGURE P9.9 (a) Bode Diagram for Gc (s)G(s)H(s) =  0.3(s+0.05)(s2 +1600) . (s+70)(s2 +0.05s+16)  40  Gain dB  20 0 -20 -40 -60 10-3  10-2  10-1  10-2  10-1  100 101 Frequency (rad/sec)  Phase deg  0 -100 -200 -300 10-3  100  101  Frequency (rad/sec)  FIGURE P9.9 CONTINUED: (b) Bode Diagram for Gs (s)G(s)H(s) = K1 = 1.  0.15K1 (s+0.05)(s2 +1600)(s+0.5) , (s+70)(s2 +0.05s+16)  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  470  CHAPTER 9  The equations of motion are F (s) = 3I(s)  and I(s) =  Eo (s) Eo (s) = . R + Ls 0.1 + 0.2s  So, F (s) =  30 Eo (s) . (2s + 1)  The actuator without the spring (see Table 2.7, Number 9 in Dorf & Bishop) is modeled via X(s) 1 Ka = = . 2 Y (s) M s + Bs τa s 2 + s With the spring, we have Ka X(s) = 2 Y (s) τa s + s + Ks  or  GA (s) =  0.4s2  1 . + s + 1.5  Then, the loop transfer function is L(s) =  30K1 . (2s + 1)(0.4s2 + s + 1.5)  (a) The Bode plot for K1 = 0.2 in Fig. P9.10 shows the P.M. = 30o . 20  Gain dB  0 -20 -40 -60 10-2  10-1  100  101  100  101  Frequency (rad/sec) 0  Phase deg  P9.10  Stability in the Frequency Domain  -100 -200 -300 10-2  10-1 Frequency (rad/sec)  FIGURE P9.10 Bode Diagram for L(s) =  30K1 , (2s+1)(0.4s2 +s+1.5)  where K1 = 0.2.  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  471  Problems  (b) For K1 = 0.2, we determine that Mpω = 7.8 dB, ωr = 1.9 rad/sec, and ωB = 2.8 rad/sec. (c) The estimated percent overshoot is P.O. = 51% and the estimated settling time is Ts = 10 sec. This is based on ζ = 0.21 and ωn ≈ ωr = 1.9 rad/sec. The loop transfer function is Gc (s)G(s) =  5(K1 s + K2 )e−1.5s . s(5s + 1)  (a) Let K1 = K2 = 1. Then Gc (s)G(s) =  5(s + 1) −1.5s e . s(5s + 1)  The Bode plot is shown in Figure P9.11a. The phase margin is P.M. = −48o . The system is unstable. (b) Let K1 = 0.1 and K2 = 0.04. Then, the loop transfer function is Gc (s)G(s) =  5(0.1s + 0.04)e−1.5s . s(5s + 1)  The Bode plot shown in Figure P9.11b shows P.M. = 45o . Thus, the system is stable. 60  Gain dB  40 20 0 -20 10-2  10-1  100  101  100  101  Frequency (rad/sec) 0 -200  Phase deg  P9.11  -400 -600 -800  -1000 10-2  10-1 Frequency (rad/sec)  FIGURE P9.11 (a) Bode Diagram for Gc (s)G(s) =  5(s+1)e−sT s(5s+1)  , where T = 1.5.  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  472  CHAPTER 9  Stability in the Frequency Domain  40  Gain dB  20 0 -20 -40 10-2  10-1  100  101  100  101  Frequency (rad/sec) 0  Phase deg  -200 -400 -600 -800 -1000 10-2  10-1 Frequency (rad/sec)  FIGURE P9.11 CONTINUED: (b) Bode Diagram for Gc (s)G(s) =  5(0.1s+0.04)e−sT s(5s+1)  , where T = 1.5.  (c) When K2 = 0.1394, the phase margin is P.M. = 0o and G.M. = 0 dB. So, for stability we require K2 ≤ 0.1394 when K1 = 0. (a) The Bode plot is shown in Figure P9.12.  Bode Diagram Gm = 12 dB (at 3.46 rad/sec) , Pm = 67.6 deg (at 1.53 rad/sec) 20  Magnitude (dB)  0 −20 −40 −60 −80 −100 0 −45 Phase (deg)  P9.12  −90 −135 −180 −225 −270 −1 10  0  10  Frequency (rad/sec)  FIGURE P9.12 Bode Diagram for Gc (s)G(s) =  2 . (0.5s+1)3  1  10  2  10  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  473  Problems  The loop transfer function (without the time delay) is Gc (s)G(s) =  2 . (0.5s + 1)3  The phase margin is P.M. = 67.6o . (b) With the delay, the loop transfer function is Gc (s)G(s)H(s) =  2e−0.5s . (0.5s + 1)3  The phase margin is now P.M. = 23.7o . So the 0.5 sec time delay has reduced the phase margin by 43.9◦ . The loop transfer function is Gc (s)G(s) =  Ka (Ks + 1) −1.2s e . s  (a) Let Ka = K = 1. Without the time delay, the system has infinite phase and gain margin. However, with the time delay, the system has a negative gain margin, hence it is unstable. (b) A plot of phase margin versus Ka is shown in Figure P9.13.  100 80 60  Phase margin deg  P9.13  40 20 0 -20 -40 -60 0  0.1  0.2  0.3  0.4  0.5  0.6  0.7  0.8  Ka  FIGURE P9.13 Phase margin as a function of Ka for Gc (s)G(s) =  Ka (s+1)e−1.2s . s  0.9  1  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  474  CHAPTER 9  Stability in the Frequency Domain  Let K = 1, and find Ka for a stable system. Then, Gc (s)G(s) =  Ka (s + 1)e−1.2s . s  If Ka = 0.8, then the phase margin is P.M. = 50o . The loop transfer function is Gc (s)G(s) =  Ke−0.2s . s(0.1s + 1)  (a) The Nichols diagram is shown in Figure P9.14 for K = 2.5.  40  0 0.25  30 0.5 1  20  -1  2 3 6  10  Gain dB  P9.14  -3 -6  0  -12  -10  -20  -20 -30 -40  -350  -300  -250  -200  -150  -100  -50  -40 0  Phase (deg)  FIGURE P9.14 Nichols diagram for Gc (s)G(s) =  Ke−0.2s , s(0.1s+1)  for K = 2.5.  It can be seen that Mpω = 2.0 dB . The phase and gain margins are P.M. = 48.5o and G.M. = 7.77 dB. (b) We determine that ζ = 0.43 (based on Mpω = 2 dB) and ζ = 0.48 (based on the phase margin P.M. = 48.5o ). (c) The bandwidth is ωB = 5.4 rad/sec .  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  475  Problems  (a) The ship transfer function is G(s) =  −0.164(s + 0.2)(s − 0.32) . s2 (s + 0.25)(s − 0.009)  The closed-loop system is unstable; the roots are s1 = −0.5467 s2,3 = 0.2503 ± 0.1893j s4 = −0.1949 Therefore the ship will not track the straight track. (b) The system cannot be stabilized by lowering the gain; this is verified in the root locus in Figure P9.15, where it is seen that the locus has a branch in the right half-plane for all K > 0. (c) Yes, the system can be stabilized. (d) When the switch is closed, we have a derivative feedback, which adds 90o phase lead. This is not enough to stabilize the system. Additional lead networks are necessary.  0.6  0.4  0.2  Imag Axis  P9.15  0  x  o  xx  -0.2  0  o  -0.2  -0.4  -0.6 -0.6  -0.4  0.2  Real Axis  FIGURE P9.15 −0.164(s+0.2)(s−0.32) Root locus for 1 + GH(s) = 1 + K s2 (s+0.25)(s−0.009) = 0.  0.4  0.6  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  476  CHAPTER 9  P9.16  Stability in the Frequency Domain  The loop transfer function is Gc (s)G(s) =  K . (s/10 + 1)(s2 + s + 2)  When K = 3.2, the phase margin is P.M. ≈ 30o . The Bode plot is shown in Figure P9.16.  Gm=10.88 dB, (w= 3.464) Pm=29.91 deg. (w=2.083) 50  Gain dB  0  -50  -100 -1 10  10  0  10  1  10  2  Frequency (rad/sec) 0  Phase deg  -90 -180 -270 -360 -1 10  10  0  10  1  10  2  Frequency (rad/sec)  FIGURE P9.16 Bode plot for Gc (s)G(s) =  P9.17  K , (s/10+1)(s2 +s+2)  where K = 3.2.  (a) We require ess ≤ 0.05A, and we have ess =  A < 0.05A 1 + Kp  or Kp > 19. But 20K1 s→0 (0.5s + 1)  Kp = lim G1 (s)G2 (s)G3 (s)G4 (s) = lim s→0  So, Kp = 0.2K1 > 19, or K1 > 95.    0.1 1 + 4s  2  = 0.2K1 .  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  477  Problems  (b) Given 1 s+1 G1 (s) = K1 (1 + ) = K1 s s     ,  we require 1.05 < MPt < 1.30, or 0.70 > ζ > 0.36, or 70o > P.M. > 36o . Then, G1 (s)G2 (s)G3 (s)G4 (s) =  0.2K1 (s + 1) . s(0.5s + 1)(4s + 1)2  When K1 = 0.8, the P.M. = 40o . The Bode plot is shown in Figure P9.17a.  Gain dB  50  0  -50  -100 10-2  10-1  100  101  100  101  Frequency (rad/sec)  Phase deg  0 -100 -200 -300 10-2  10-1 Frequency (rad/sec)  FIGURE P9.17 (a) Bode plot for G1 (s)G2 (s)G3 (s)G4 (s) = P.M. = 40o .  0.2K1 (s+1) , s(0.5s+1)(4s+1)2  where K1 = 0.8 and  (c) For part (a), we had G1 (s)G2 (s)G3 (s)G4 (s) =  2.375 . (s + 2)(s + 0.25)2  The characteristic equation is s3 + 2.5s2 + 1.06s + 2.50 = (s + 2.48)(s2 + 0.02s + 1.013) . The dominant complex roots are lightly damped since ζ = 0.01 and ζωn = 0.01.  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  CHAPTER 9  Stability in the Frequency Domain  Thus, Ts =  4 = 400 sec . ζωn  For part (b), we had G1 (s)G2 (s)G3 (s)G4 (s) =  (0.2)(0.8)(s + 1) . s(0.5s + 1)(4s + 1)2  The characteristic equation is 8s4 + 20s3 + 8.5s2 + 1.16s + 0.16 = 0 . The roots are s1 = −2, s2 = −0.4 and s3,4 = −0.05 ± j0.15. Thus ζ = 0.16 and ζωn = 0.05. So, Ts =  4 4 = = 75 sec . ζωn 0.05  (d) Let U (s) be a unit step disturbance and R(s) = 0. Then Y (s) G3 (s)G4 (s) = = U (s) 1 + G1 (s)G2 (s)G3 (s)G4 (s) 1+  2 0.1 1+4s 20K1 (s+1) s(0.5s+1)(4s+1)2   The disturbance response is shown in Figure P9.17b. x10 -3 6 5 4 3  Amplitude  478  2 1 0 -1 -2  0  10  20  30  40  50  60  70  Time (secs)  FIGURE P9.17 CONTINUED: (b) System response to a unit disturbance U (s).  80  90  100  .  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  479  Problems  P9.18  The transfer function is Gc (s)G(s)H(s) =  5.3(s2 + 0.8s + 0.32)e−T s . s3  The Bode plot is shown in Figure P9.18. 80  Gain dB  60 40 20 0 -20 10-1  100  101  Frequency (rad/sec)  Phase deg  300  T=0 solid ___ & T=0.1 dashed ---- & T=0.2674 dotted ....  250 200 150 100 10-1  100  101  Frequency (rad/sec)  FIGURE P9.18 K(s2 +0.8s+0.32)e−sT , where T = 0 (solid line), Bode diagram for Gc (s)G(s)H(s) = s3 T = 0.1 (dashed line), and T = 0.2674 (dotted line).  The following results are verified in the figure. (a) The phase margin is P.M. = 81o at ω = 5.3 when T = 0. (b) For T = 0.1, the added phase is φ = −T ω (in radians). The phase margin is P.M. = 51o at ω = 5.3 when T = 0.1. (c) The system is borderline stable when T = 0.2674 sec. The phase margin is P.M. = 0o at ω = 5.3. P9.19  The transfer function is Gc (s)G(s) =  0.5 . s(1 + 2s)(4 + s)  (a) The Nichols diagram is shown in Figure P9.19. The gain margin is G.M. = 31.4 dB. (b) The phase margin is P.M. = 75o and Mpω = 0 dB. The bandwidth is 0.17 rad/sec.  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  480  CHAPTER 9  Stability in the Frequency Domain  Phase margin = 180-105=75o  Nichols Chart 40 0.25 dB 0.5 dB 1 dB 3 dB 6 dB  20 0  0 dB sys System: Gain (dB): Phase (deg): Frequency (rad/sec): 0.122  AM IW APR IW ARN IW  ARN AVN  Open  AMN  AMN IW  ATN  ATN IW  APNN  APNN IW  APRN  APRN IW  APVN  APVN IW  APMN ALMN  ALPQ  ARS0  ARRQ APTN APLQ O>?@ABCC> DEFG? HI?JK  FIGURE P9.19 Nichols diagram for Gc (s)G(s) =  AUN  Gain margin = 31.4 dB  APMN IW 0 AVQ  0.5 . s(2s+1)(s+4)  (a) Let K = 100. The Bode plot is shown in Figure P9.20a. The loop transfer function is Gc (s)G(s) =  K(s2 + 1.5s + 0.5) . s(20s + 1)(10s + 1)(0.5s + 1)  Gain dB  100 50 0 -50 10-3  10-2  10-1 Frequency (rad/sec)  100  101  100  101  0  Phase deg  P9.20  AVN IW  System: sys Gain (dB): 4 Phase (deg): Frequency (rad/sec): 1.44  -100 -200 -300 10-3  10-2  10-1 Frequency (rad/sec)  FIGURE P9.20 (a) Bode diagram for Gc (s)G(s) =  K(s2 +1.5s+0.5) , s(20s+1)(10s+1)(0.5s+1)  where K = 100.  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  481  Problems  (b) The phase margin is P.M. = −3.5o and the gain margin is G.M. = 2.7 dB. (c) You must decrease K below 100 to achieve a P.M. = 40o . For K = 0.1, the phase margin P.M. = 37.9o . (d) The step response is shown in Figure P9.20b for K = 0.1.  1.4  1.2  Amplitude  1  0.8  0.6  0.4  0.2  0 0  50  100  150  200  250  300  350  400  Time (secs)  FIGURE P9.20 CONTINUED: (b) Unit step response K = 0.1.  P9.21  The loop transfer function is Gc (s)G(s) =  K . s(s + 1)(s + 4)  (a) The Bode plot is shown in Figure P9.21 for K = 4. (b) The gain margin is G.M. = 14 dB . (c) When K = 5, the gain margin is G.M. = 12 dB . (d) We require Kv > 3, but Kv = K4 . So, we need K > 12. This gain can be utilized since K < 20 is required for stability.  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  482  CHAPTER 9  Stability in the Frequency Domain  50  Gain dB  0 -50 -100 -150 10-1  100  101  102  101  102  Frequency (rad/sec)  Phase deg  0 -100 -200 -300 10-1  100 Frequency (rad/sec)  FIGURE P9.21 Bode diagram for Gc (s)G(s) =  P9.22  K , s(s+1)(s+4)  where K = 4.  (a) The resonant frequency ωr = 5.2 rad/sec is point 6 on the Nichol's chart. (b) The bandwidth is between points 8 and 9. We estimate the bandwidth to be ωB = 7.5 rad/sec. (c) The phase margin P.M. = 30o . (d) The gain margin G.M. = 8 dB. (e) Since we have P.M. = 30o , then we estimate ζ = 0.3. We can also approximate ωn ≈ ωr = 5.2 .ap9.1 Thus, Ts =  P9.23  4 4 = = 2.5sec . ζωn 1.56  The phase margin is P.M. = 60 deg when K = 266. The gain margin is G.M. = 17.2 dB . The Bode plot is shown in Figure P9.23.  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  483  Problems Bode Diagram Gm = 17.2 dB (at 9.8 rad/sec) , Pm = 60 deg (at 2.58 rad/sec)  Magnitude (dB)  50 0 −50 −100  Phase (deg)  −150 −90 −135 −180 −225 −270 −1 10  0  10  FIGURE P9.23 Bode diagram for Gc (s)G(s) =  P9.24  1  2  10 Frequency (rad/sec)  K , s(s+8)(s+12)  10  3  10  where K = 266.  When K = 14.1, then P.M. = 45 deg, G.M. = ∞ dB and ωB = 29.3 rad/sec. Gm=356.59 dB (at 0 rad/sec), Pm=60 deg. (at 17.321 rad/sec) 100  Phase (deg); Magnitude (dB)  50  0  - 50 - 80  - 100  - 120  - 140  - 160  - 180 -1 10  10  0  10  1  10  2  Frequency (rad/sec)  FIGURE P9.24 Bode diagram for G(s) =  P9.25  K(s+20) , s2  where K = 14.1.  The phase margin is P.M. = 60 deg when K = 2.61 and T = 0.2 second. The Bode plot is shown in Figure P9.25.  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  484  CHAPTER 9  Stability in the Frequency Domain K=2.61; PM=60.09 at wc=2.61 rad/sec 40  Gain dB  20 0 -20 -40 -1 10  10  0  10  1  10  2  Frequency (rad/sec)  Phase deg  0  -500  -1000  -1500 -1 10  10  0  10  1  10  2  Frequency (rad/sec)  FIGURE P9.25 Bode diagram for Gc (s)G(s) =  P9.26  Ke−0.2s , s  where K = 2.61.  The loop transfer function is Gc (s)G(s) =  K . s(0.25s + 1)(0.1s + 1)  The Bode plot is shown in Figure P9.26a for K = 10. The Nichols chart is shown in Figure P9.26b. The phase and gain margins are P.M. = 9o  and  G.M. = 3 dB .  The system bandwidth is ωB = 8 rad/sec. From the P.M. = 9o , we estimate ζ = 0.09. Therefore, the predicted overshoot is √  −πζ/  P.O. = 100e  1−ζ 2  = 75% , where ζ = 0.09 .  The resonant peak occurs at ωr = 5.5 rad/sec. If we estimate ωn ≈ ωr = 5.5 rad/sec, then the settling time is Ts =  4 = 8 sec . ζωn  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  485  Problems  Gain dB  50  0  -50 -100 10-1  100  101  102  101  102  Frequency (rad/sec)  Phase deg  0 -100 -200 -300 10-1  100 Frequency (rad/sec)  40  0 0.25  30 0.5 1  20 10  Gain dB  -1  2 3  -3  16  0  -6 -12  -10  -20  -20 -30 -40  -350  -300  -250  -200  -150  -100  -50  -40 0  Phase (deg)  FIGURE P9.26 (a) Bode diagram for Gc (s)G(s) = for Gc (s)G(s) =  P9.27  K , s(0.25s+1)(0.1s+1)  K , s(0.25s+1)(0.1s+1)  where K = 10. (b) Nichols chart  where K = 10.  The loop transfer function is L(s) = Gc (s)G(s)H(s) =  4K . (s2 + 2s + 4)(s + 1)  The plot of the phase margin versus the gain K is shown in Figure P9.27. As the gain increases towards Kmax = 3.5, the phase margin decreases  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  486  CHAPTER 9  Stability in the Frequency Domain  towards zero.  180  160  140  Phase margin (deg)  120  100  80  60  40  20  0  1  1.5  2  2.5  3  3.5  K  FIGURE P9.27 Phase margin versus the gain K.  P9.28  The loop transfer function is Gc (s)G(s) =  KP . s(s + 1)  When KP = 1.414, we have P.M. ≈ 45◦ . Using the approximation that ζ ≈ P.M./100 we estimate that ζ = 0.45. Then using the design formula √ 2 P.O. = 100e−πζ/ 1−ζ = 20.5% . The actual overshoot is 23.4%. The step input response is shown in Figure P9.28. The actual damping ratio is ζ = 0.42. This shows that the approximation ζ ≈ P.M./100 is quite applicable and useful in predicting the percent overshoot.  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  487  Problems  Step Response 1.4  1.2  System: syscl Peak amplitude: 1.23 Overshoot (%): 23.3 At time (sec): 2.97  Amplitude  1  0.8  0.6  0.4  0.2  0  0  5  10 Time (sec)  FIGURE P9.28 Step response showing a 23.3% overshoot.  15  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  488  CHAPTER 9  Stability in the Frequency Domain  Advanced Problems The loop transfer function is L(s) = Gc (s)G(s)H(s) =  236607.5(s + 10)(s + 5) . s(s + 2)(s2 + 100s + ωn2 )(s + 1)  (a) The Bode plot for ωn2 = 15267 is shown in Figure AP9.1a. 150  Gain dB  100 50 0 -50 -100 10-3  10-2  10-1  10-2  10-1  100 101 Frequency (rad/sec)  102  103  102  103  0  Phase deg  AP9.1  -100 -200 -300 10-3  100  101  Frequency (rad/sec)  FIGURE AP9.1 (a) Bode Diagram for L(s) =  236607.5(s+10)(s+5) 2 )(s+1) , s(s+2)(s2 +100s+ωn  2 where ωn = 15267.  The phase and gain margins are P.M. = 48.6o  and  G.M. = 15.5 dB .  (b) The Bode plot for ωn2 = 9500 is shown in Figure AP9.1b. The gain and phase margins are P.M. = 48.5o  and  G.M. = 10.9 dB .  Reducing the natural frequency by 38% has the effect of reducing the gain margin by 30%.  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  489  Advanced Problems  150  Gain dB  100 50 0 -50 -100 10-3  10-2  10-1  100  101  102  103  101  102  103  Frequency (rad/sec)  Phase deg  0 -100 -200 -300 10-3  10-2  10-1  100 Frequency (rad/sec)  FIGURE AP9.1 CONTINUED: (b) Bode Diagram for L(s) =  2 where ωn = 9500.  (a) The Bode plot with T = 0.05 sec is shown in Figure AP9.2a. The phase margin is P.M. = 47.7o and the gain margin is G.M. = 11.2 dB. 40  Gain dB  20 0 -20 -40 100  101  102  Frequency (rad/s) -100  Phase deg  AP9.2  236607.5(s+10)(s+5) 2 )(s+1) , s(s+2)(s2 +100s+ωn  -200  -300 -400 100  101 Frequency (rad/s)  FIGURE AP9.2 (s+5) (a) Bode Diagram for Gc (s)G(s)H(s) = 8 s(s+2) e−sT , where T = 0.05s.  102  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  490  CHAPTER 9  Stability in the Frequency Domain  (b) The Bode plot with T = 0.1 sec is shown in Figure AP9.2b. The  40  Gain dB  20 0 -20 -40 100  101 Frequency (rad/s)  102  Phase deg  0 -200 -400 -600 -800 100  101  102  Frequency (rad/s)  FIGURE AP9.2 (s+5) CONTINUED: (b) Bode Diagram for Gc (s)G(s)H(s) = 8 s(s+2) e−sT , where T = 0.1s.  phase margin is P.M. = 22.1o and the gain margin is G.M. = 4.18 dB. A 100% increase in time delay T leads to a 50% decrease in phase and gain margins. (c) The damping ratio ζ ≈ P.M./100 and  √ 2 P.O. ≈ 100e−πζ/ 1−ζ .  So, for T = 0.05 sec, ζ ≈ 0.47 and P.O. ≈ 18.7%. Also, for T = 0.1 sec, ζ ≈ 0.22 and P.O. ≈ 49.2%. AP9.3  The loop transfer function is L(s) = Gc (s)G(s)H(s) =  66K(1 + 0.1s) . (1 + 0.01s)(1 + 0.01s)(1 + 1.5s)(1 + 0.2s)  (a) When K = 1, the gain and phase margins are G.M. = 18.4 dB and P.M. = 55o . (b) When K = 1.5, the gain and phase margins are G.M. = 14.9 dB and P.M. = 47.8o .  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  491  Advanced Problems  (c,d) The bandwidth and settling time with K = 1 are ωB = 233.6 rad/sec and Ts = 0.4 second. When K = 1.5, we determine that ωB = 294.20 rad/sec and Ts = 0.33 second. AP9.4  The loop transfer function is L(s) = Gc (s)G(s) =  K(s + 40) . s(s + 15)(s + 10)  The gain K = 28.8 satisfies the specifications. The actual gain and phase margins are G.M. = 18.8 dB and P.M. = 45o . The system bandwidth is ωB = 10.3 rad/sec. The step response is shown in Figure AP9.4.  Step Response  System: sys_cl Peak amplitude: 1.23 1.4 Overshoot (%): 23.4 At time (sec): 0.476  1.2  Amplitude  1 System: sys_cl Settling Time (sec): 1.1  0.8  0.6  0.4  0.2  0  0  0.2  0.4  0.6  0.8 1 1.2 Time (sec)  1.4  1.6  1.8  2  FIGURE AP9.4 Closed-loop system step response.  AP9.5  The loop transfer function is L(s) = Gc (s)G(s) = K  s + 0.4 . s4 + 9s3 + 18s2  The Bode plot for K = 1 is shown in Figure AP9.5. From the phase response, we determine that the maximum P.M. ≈ 41o . From the magnitude response (for K = 1), we find that the gain needs to be raised to K = 14 to achieve maximum phase margin at ω = 0.826 rad/sec. The  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  492  CHAPTER 9  Stability in the Frequency Domain  gain and phase margin with K = 14 are G.M. = 19.3 dB and P.M. = 40.9o . Also, the overshoot is P.O. = 38.3%. Bode Diagram Gm = 42.3 dB (at 3.79 rad/sec) , Pm = 16.7 deg (at 0.154 rad/sec)  Magnitude (dB)  50  0  −50  −100  Phase (deg)  −150 −135 System: sys Frequency (rad/sec): 0.865 Phase (deg): −139  −180  −225  −270 −2 10  −1  10  0  10 Frequency (rad/sec)  1  10  2  10  FIGURE AP9.5 s+0.4 Bode plot for L(s) = K s4 +9s 3 +18s2 with K = 1.  AP9.6  With D > 2m, the gain can be increased up to K = 100, while still retaining stability.  AP9.7  The loop transfer function is L(s) = Gc (s)G(s) =  K(s + 4) . s2  We select √ K=2 2 for P.M. = 45o . The system bandwidth is ωB = 5.88 rad/sec . The disturbance response is shown in Figure AP9.7. The maximum output due to a disturbance is y(t) = 0.11.  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  493  Advanced Problems  0.12  0.1  Amplitude  0.08  0.06  0.04  0.02  0 0  0.5  1  1.5  2  2.5  3  3.5  4  Time (secs)  FIGURE AP9.7 Closed-loop system disturbance response.  A reasonable choice for the gain is K = 2680. The phase margin is P.M. = 42.8◦ and the percent overshoot is P.O. = 18.9%. The Nichols chart is shown in Figure AP9.8.  Nichols Chart 60 40 0.25 dB 0.5 dB 1 dB 3 dB 6 dB  20 Open−Loop Gain (dB)  AP9.8  0 dB −1 dB  −20  −3 dB −6 dB −12 dB −20 dB  −40  −40 dB  −60  −60 dB  −80  −80 dB  −100  −100 dB  0  −120 −360  FIGURE AP9.8 Nichols chart.  −315  −270  −225 −180 −135 Open−Loop Phase (deg)  −90  −120 dB −45 0  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  494  CHAPTER 9  AP9.9  Stability in the Frequency Domain  The loop transfer function is L(s) = Gc (s)G(s) =  Kp (s + 0.2) 2 s (s2 + 7s + 10)  .  At the maximum phase margin, Kp = 4.9 for P.M. = 48.6o . The Bode diagram is shown in Figure AP9.9. Bode Diagrams Gm=21.788 dB (at 2.9326 rad/sec), Pm=48.457 deg. (at 0.50782 rad/sec) 100  50  0  Phase (deg); Magnitude (dB)  -50  -100  -150 -100  -150  -200  -250  -300 -3 10  10  -2  10  -1  10  0  10  1  10  2  Frequency (rad/sec)  FIGURE AP9.9 Phase and gain margin.  AP9.10  The closed-loop transfer function is T (s) =  s2  K . + 3s + 1  We require K = 1 a zero steady-state tracking error to a unit step. The step response is shown in Figure AP9.10. Computing T (jω) = 0.707 it follows that (jω)2  1 = 0.707 + 3jω + 1  or ω 4 + 7ω 2 − 1 = 0 .  Solving for ω yields ω = 0.37 rad/s. This is the bandwidth of the system.  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  495  Advanced Problems  Step Response 1 0.9 0.8  Amplitude  0.7 0.6 0.5 0.4 0.3 0.2 0.1 0  0  2  4  6  8 Time (sec)  10  12  14  16  FIGURE AP9.10 Unit step response.  The phase margin versus time delay is shown in Figure AP9.11a.  80 Time Delay=1  70  PM=58.5285  Time Delay=3.0455  60  Phase Margin (deg)  AP9.11  PM=0.001  50 40 30 20 10 0 −10 0.5  1  FIGURE AP9.11 Phase margin versus time delay.  1.5  2 Time Delay (s)  2.5  3  3.5  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  CHAPTER 9  Stability in the Frequency Domain  The maximum time delay is T = 3.04 s for stability. The step response is shown in Figure AP9.11b. The percent overshoot is P.O. = 7.6%.  1.4  1.2  1  Amplitude  496  0.8  0.6  0.4  0.2  0  0  FIGURE AP9.11 Unit step response.  1  2  3  4  5 6 Time (sec)  7  8  9  10  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  497  Design Problems  Design Problems CDP9.1  The plant model with parameters given in Table CDP2.1 in Dorf and Bishop is given by: 26.035 θ(s) = , Va (s) s(s + 33.142) where we neglect the motor inductance Lm and where we switch off the tachometer feedback (see Figure CDP4.1 in Dorf and Bishop). The closedloop system characteristic equation is 1+  26.035Ka =0. s(s + 33.142)  The phase margin is P.M. = 70.4◦ when Ka = 16. The step response with K = 16 is shown below. 1.2  1  Amplitude  0.8  0.6  0.4  0.2  0  DP9.1  0  0.05  0.1  0.15 Time (secs)  0.2  0.25  0.3  (a) The gain and phase margins are G.M. = 7 dB and P.M. = 60o . (b) The resonant peak and frequency are Mpω = 2 dB and ωr = 5 rad/sec. (c) We have ωB = 20 rad/sec. From Mpω = 2 dB we estimate ζ = 0.45 (Figure 8.11 in Dorf & Bishop). Also, ωr /ωn = 0.8, so ωn = 6.25. Thus, Ts = 1.4. (d) We need P.O. = 30o or ζ = 0.3 or P.M. ≈ 30o . So, we need to raise the gain by 10 dB or K = 3.2.  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  498  CHAPTER 9  The loop transfer function is L(s) = Gc (s)G(s) =  K(s + 0.5) . + 7.5s + 9)  s2 (s2  When K = 6.25, we have the maximum phase margin. The phase margin maximum is P.M. = 23o . The plot of P.M. versus K is shown in Figure DP9.2a.  24 22 20 18 Phase Margin deg  DP9.2  Stability in the Frequency Domain  16 14 12 10 8 6 4  0  1  2  3  4  5  6  7  8  9  K  FIGURE DP9.2 (a) Phase margin versus K for L(s) =  K(s+0.5) . s2 (s2 +7.5s+9)  The predicted damping is ζ = 0.23. It then follows that the predicted percent overshoot is √ 2 P.O. = 100e−πζ/ 1−ζ = 48% . The actual overshoot is 65%. The step input response is shown in Figure DP9.2b. The resonant peak occurs at ωr = 0.75 rad/sec. Approximating ωn ≈ ωr = 0.75 rad/sec, we can estimate the settling time as Ts =  4 = 23 sec . ζωn  The actual settling time is 20 sec.  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  499  Design Problems  1.8 1.6 1.4  Amplitude  1.2 1 0.8 0.6 0.4 0.2 0 0  5  10  15  20  25  30  Time (secs)  FIGURE DP9.2 CONTINUE: (b) Closed-loop unit step response.  We want to select the gain K as large as possible to reduce the steady-state error, but we want a minimum phase margin of P.M. = 45o to achieve good dynamic response. A suitable gain is K = 4.2, see Figure DP9.3. K=4.2; PM=45.34 at wc=0.102 rad/sec 20  Gain dB  0  -20  -40 -2 10  10  -1  10  0  10  1  Frequency (rad/sec) 0 -100 Phase deg  DP9.3  -200 -300 10  -2  10  -1  10 Frequency (rad/sec)  FIGURE DP9.3 Bode plot for G(s) =  Ke−10s 40s+1 .  0  10  1  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  500  CHAPTER 9  DP9.4  Stability in the Frequency Domain  We are given the loop transfer function L(s) = Gc (s)G(s) =  K s(s + 1)(s + 4)  which can be written as Gc (s)G(s) =  Kv . s(s + 1)(0.25s + 1)  The performance results are summarized in Table DP9.4.  Kv  TABLE DP9.4  G.M.  P.M.  ωB  P.O.  Ts  (dB)  (deg)  (rad/sec)  (%)  (sec)  0.40  21.9  64.2  0.62  4.4  9.8  0.75  16.5  49.0  1.09  19.0  10.1  Summary for Kv = 0.40 and Kv = 0.75.  When Kv = 0.40, we have ess 1 = = 2.5 , A 0.40 or 2 1/2 times the magnitude of the ramp. This system would be acceptable for step inputs, but unacceptable for ramp inputs. DP9.5  (a) With a time delay of T = 0.8 second, we determine that the proportional controller Gc (s) = K = 7 provides a suitable response with P.O. = 8.3 % ess = 12.5 %  Ts = 4.38 sec .  (b) A suitable proportional, integral controller is Gc (s) = K1 + K2 /s = 6 + 0.6/s .  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  501  Design Problems  The response to a unit step is P.O. = 5.14 %  ess = 0 % Ts = 6.37 sec .  The Nichols chart is shown in Figure DP9.5.  40  0 0.25  30 0.5 1  20 3 6 8  10  Gain dB  -1 -3 -6  0  -12  -10  -20  -20 -30 -40  -350  -300  -250  -200  -150  -100  -50  -40 0  Phase (deg)  FIGURE DP9.5 Nichols chart for Gc (s)G(s) =  DP9.6  (K1 s+K2 )e−0.8s , s(10s+1)  where K1 = 6 and K2 = 0.6.  With K = 170, at the two extreme values of b, we have b = 80 b = 300  P.M. = 91.62o P.M. = 75.23o  G.M. = 13.66 dB G.M. = 25.67 dB .  Since reducing the value of K only increases the P.M. and G.M., a value of K = 170 is suitable to meet P.M. = 40o and G.M. = 8 dB for the range of b. DP9.7  A suitable gain is K = 0.22 . This results in P.M. = 60.17o and G.M. = 13.39 dB. The step reponse is shown in Figure DP9.7.  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  502  CHAPTER 9  Stability in the Frequency Domain  1.2  1  Amplitude  0.8  0.6  0.4  0.2  0 0  2  4  6  8  10  12  14  16  18  20  Time (secs)  FIGURE DP9.7 Lunar vehicle step response.  A gain of K = 315000 will satisfy the P.O. specification, while giving the fastest response. The step response is shown in Figure DP9.8.  1.2  1  0.8  Amplitude  DP9.8  0.6  0.4  0.2  0 0  0.1  0.2  0.3  0.4  0.5 Time (secs)  FIGURE DP9.8 Steel rolling mill step response.  0.6  0.7  0.8  0.9  1  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  503  Design Problems  The closed-loop transfer function is Ts (2) =  Gc (s)G2 (s) G1 (s) To (s) + T2d (s) . 1 + Gc (s)G2 (s) 1 + Gc (s)G2 (s)  where 1 (10s + 1)(50s + 1)  G1 (s) = and G2 (s) =  0.01 . (10s + 1)(50s + 1)  The steady-state error (with Gc (s) = 500) to a unit step 2A (and after the system has settled out subsequent to a step of magnitude A) is ess = 2(0.167) = 0.33 . The step response is shown in Figure DP9.9. Gc=500 (solid); Gc=1/s (dashed); Gc=600+6/s (dotted) 2.5  2  1.5  T2/A  DP9.9  1  0.5  0 0  200  400  600  800  1000  Time (sec)  FIGURE DP9.9 Two tank temperature control step response.  A suitable integral controller is Gc (s) =  1 . s  1200  1400  1600  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  504  CHAPTER 9  Stability in the Frequency Domain  In this case, the steady-state tracking error is zero , since the system is a type 1. The system response is shown in Figure DP9.9. With the integral controller, the settling time is about Ts = 438 seconds and the P.O. = 7%. A suitable PI controller is Gc (s) = 600 +  6 . s  With the PI controller, the settling time is about Ts = 150 seconds and the P.O. = 10%. DP9.10  The system is given by ẋ = Ax + Br y = Cx where   A=  0  1  2 − K1 3 − K2       ,  B=  0 1  The associated transfer function is T (s) =  s2     , and  C=    1 0    .  1 . + (K2 − 3)s + K1 − 2  The characteristic polynomial is s2 + (K2 − 3)s + K1 − 2 = 0 . If we select K1 = 3, then we have a zero-steady error to a unit step response R(s) = 1/s, since s2 + (K2 − 3)s =0. s→0 s2 + (K2 − 3)s + K1 − 2  lim s [1 − T (s)] R(s) = lim  s→0  Let K=    3 4.3    .  The step response is shown in Figure DP9.10a. The bandwidth is ωb = 1.08 rad/s, as seen in Figure DP9.10b.  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  505  Design Problems  Step Response 1.4  1.2  Amplitude  1  0.8  0.6  0.4  0.2  0  0  1  2  3  4 5 Time (sec)  6  7  8  9  Bode Diagram 5  0 System: sys Frequency (rad/sec): 1.08 Magnitude (dB): −3  −5  Magnitude (dB)  −10  −15  −20  −25  −30  −35  −40 −1 10  0  10 Frequency (rad/sec)  1  10  FIGURE DP9.10 Step response with K = [3 4.3] and closed-loop Bode plot.  DP9.11  A time domain step response specification P.O. > 10% requires the dominant poles to have a damping ration of ζ = 0.6. This time domain specification can be transformed to a frequency response specification using the approximation P.M. ≈ 100ζ = 60◦ . To keep the problem tractable, we consider the controller with the form Gc (s) = KP +  KI 1 = KP + , s s  where we let KI = 1. The plot of the P.M. as a function of KP is shown  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  CHAPTER 9  Stability in the Frequency Domain  in Figure DP9.11a. If we select KP = 0.07 we expect a phase margin of approximately 60◦ , hence a percent overshoot P.O. ≤ 10%. The step response is shown in Figure DP9.11b. The actual phase margin is P.M. = 60.2◦ , the percent overshoot is P.O. = 5.9% and the settling time is Ts = 3.4 sec.  85  Phase Margin (deg)  80  75  70  65  60  55  0  0.05  0.1  0.15  0.2  0.25 KP  0.3  0.35  0.4  0.45  0.5  1.4  1.2  1  Amplitude  506  0.8  0.6  0.4  0.2  0  0  1  2  3 Time (sec)  4  5  6  FIGURE DP9.11 (a) Phase margin versus controller gain KP and KI = 1. (b) Step response with KP = 0.07 and KI = 1.  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  507  Computer Problems  Computer Problems The m-file script to generate the Bode plot (from which the gain and phase margin can be determined) is shown in Figure CP9.1. The transfer function is G(s) =  s2  141 . + 2s + 12  The gain margin is G.M. = ∞ and the phase margin is P.M. = 10o .  num=141; den=[1 2 12]; sys = tf(num,den); margin(sys); Bode Diagram Gm = Inf dB (at Inf rad/sec) , Pm = 10 deg (at 12.3 rad/sec)  Magnitude (dB)  40 20 0 −20 −40 0 Phase (deg)  CP9.1  −45 −90 −135 −180 −1 10  0  10  Frequency (rad/sec)  FIGURE CP9.1 Gain and phase margin with the margin function.  1  10  2  10  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  508  CHAPTER 9  The Nyquist plots are shown in Figures CP9.2a-c.  num=[5]; den=[1 5]; sys=tf(num,den); nyquist(sys) 0.5 0.4 0.3  Imaginary Axis  0.2 0.1 0 −0.1 −0.2 −0.3 −0.4 −0.5 −1  −0.8  −0.6  FIGURE CP9.2 (a) Nyquist plot for G(s) =  −0.4  −0.2  0 Real Axis  0.2  0.4  0.6  0.8  1  5 s+5 .  num=[50]; den=[1 10 25]; sys=tf(num,den); nyquist(sys) 1.5  1  Imaginary Axis  CP9.2  Stability in the Frequency Domain  0.5  0  −0.5  −1  −1.5 −1  −0.5  0  0.5 Real Axis  FIGURE CP9.2 CONTINUED: (b) Nyquist plot for G(s) =  50 s2 +10s+25 .  1  1.5  2  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  509  Computer Problems  num=[15]; den=[1 3 3 1]; sys=tf(num,den); nyquist(sys) 15  10  Imaginary Axis  5  0  −5  −10  −15 −5  0  5 Real Axis  FIGURE CP9.2 CONTINUED: (c) Nyquist plot for G(s) =  CP9.3  10  15  15 s3 +3s2 +3s+1 .  The m-file script to generate the Nichols chart for part (a) is shown in Figure CP9.3a. The Nichols charts for (b) and (c) are similiarly generated; all plots are in Figure CP9.3a-c.  Nichols Chart 40 0 dB 30  0.25 dB  num = [1]; den = [1 0.2]; sys = tf(num,den); nichols(sys) ngrid  Open−Loop Gain (dB)  0.5 dB 20  1 dB  −1 dB  3 dB  10  −3 dB  6 dB 0  −6 dB  −10  −12 dB  −20 dB −20 −360  FIGURE CP9.3 (a) M-file script and Nichols chart for G(s) =  −315  1 s+0.1 .  −270  −225 −180 −135 Open−Loop Phase (deg)  −90  −45  0  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  CHAPTER 9  Stability in the Frequency Domain  The gain and phase margin for each transfer function are as follows: (a) G.M. = ∞ and P.M. = 102o (b) G.M. = ∞ and P.M. = ∞  (c) G.M. = 20 dB and P.M. = ∞  Nichols Chart 40 0 dB 30  0.25 dB 0.5 dB  Open−Loop Gain (dB)  20  1 dB  −1 dB  3 dB 6 dB  10  −3 dB  0  −6 dB  −10  −12 dB  −20  −20 dB  −30 −40 dB  −40 −50 −60 −360  −315  −270  −225 −180 −135 Open−Loop Phase (deg)  FIGURE CP9.3 CONTINUED: (b) Nichols chart for G(s) =  −90  −45  −60 dB 0  1 s2 +2s+1 .  Nichols Chart 40 0 dB 0.25 dB 0.5 dB 1 dB  20  −1 dB  3 dB 6 dB  −3 dB −6 dB  0 Open−Loop Gain (dB)  510  −12 dB −20  −20 dB  −40  −40 dB  −60  −60 dB  −80  −80 dB  −100 −360  −315  −270  −225 −180 −135 Open−Loop Phase (deg)  FIGURE CP9.3 CONTINUED: (c) Nichols chart for G(s) =  −90  24 s3 +9s2 +26s+24 .  −45  −100 dB 0  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  511  Computer Problems  CP9.4  To obtain a phase margin P.M. = 40◦ we select K = 15 when T = 0.2 second. The variation in the phase margin for 0 ≤ T ≤ 0.3 is shown in Figure CP9.4. T=[0:0.01:0.3]; K=15; num=K;den=[1 12]; sys = tf(num,den); % w=logspace(-2,1,400); for i=1:length(T) [mag,phase,w]=bode(sys); ph(1:length(phase))=phase(1,1,:); ph=ph'; ph2=ph-w*T(i)*180/pi; [Gm,Pm,Wcg,Wcp]=margin(mag,ph2,w); clear ph ph2 PMo(i)=Pm; end plot(T,PMo), grid xlabel('Time delay (sec)') ylabel('Phase margin (deg)') K=15 160 140  Phase margin (deg)  120 100 80 60 40 20 0 −20  0  0.05  0.1  0.15 0.2 Time delay (sec)  0.25  0.3  FIGURE CP9.4 Variation in the phase margin for 0 ≤ T ≤ 0.3 with K = 15.  CP9.5  The loop transfer function is L(s) = Gc (s)G(s) =  K(s + 50) . s(s + 20)(s + 10)  0.35  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  512  CHAPTER 9  Stability in the Frequency Domain  The plot of system bandwidth versus the gain K is shown in Figure CP9.7.  K=[0.1:1:50]; w=logspace(-2,3,2000); den=[1 30 200 0]; for i=1:length(K) num=K(i)*[1 50]; sys = tf(num,den); sys_cl = feedback(sys,[1]); [mag,phase,w]=bode(sys_cl,w); L=find(mag<0.707); wb(i)=w(L(1)); end plot(K,wb), grid xlabel('Gain K') ylabel('Bandwidth (rad/sec)')  15  Bandwidth (rad/sec)  10  5  0  0  5  10  15  20  25 Gain K  30  35  40  45  50  FIGURE CP9.5 Variation in the system bandwidth for 0 ≤ K ≤ 50.  CP9.6  The m-file script and Bode plot are shown in Figure CP9.6. The gain and phase margin and ωc are determined to be G.M. = 2.23, P.M. = 26o and ωc = 12.6 rad/sec. So, the maximum value of bo is found to be bomax = 2.13bo = 1.11 . In this problem, there is also a minimum value of bo . Using the Routh-  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  513  Computer Problems gm = 2.2238  numg = -0.5*[1 0 -2500]; deng = [1 47 850 -3000]; sysg = tf(numg,deng); numc = 10*[1 3]; denc = [1 0]; sysc = tf(numc,denc); sys_o = series(sysc,sysg); bode(sys_o) [mag,phase,w] = bode(sys_o); [gm,pm,wg,wc] = margin(mag,phase,w)  pm = 26.3187 wg = 26.1155 wc = 12.6487  Gain dB  50  0  -50 10-1  100  101  102  103  102  103  Frequency (rad/sec)  Phase deg  250 200 150 100 50 10-1  100  101 Frequency (rad/sec)  FIGURE CP9.6 Using the margin function to compute stability margins.  Hurwitz method, we determine that (for stability) the range of bo is 0.14 < bo < 1.11 . CP9.7  The m-file script is shown in Figure CP9.7a. Since we do not have a value for J, we write the loop transfer function as Gc (s)G(s) =  K̄1 + K̄2 s s2  where K̄1 = K1 /J and K̄2 = K2 /J. We work with K̄1 and K̄2 , then we can always compute K1 and K2 whenever J is specified. A PD controller which meets the specs is given by Gc (s) = 0.04 + 0.3s . The step response is shown in Figure CP9.7b. The Bode plot is shown in  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  514  CHAPTER 9  Stability in the Frequency Domain  % % Part (a) % numc = [0.3 0.04]; denc = [1]; sysc = tf(numc,denc); numg = [1]; deng = [1 0 0]; sysg = tf(numg,deng); sys_o = series(sysc,sysg); sys_cl = feedback(sys_o,[1]); step(sys_cl), pause % % Part (b) % w = logspace(-1,1,400); [mag,phase] = bode(sys_o,w); [gm,pm,w1,w2] = margin(mag,phase,w); margin(mag,phase,w), pause % % Part (c) % T = [1:0.1:5]; for i = 1:length(T) [numd,dend] = pade(T(i),2); sysd = tf(numd,dend); sys_o1 = series(sysd,sys_o); sys_cl1 = feedback(sys_o1,sysd); p(:,i) = pole(sys_cl1); end plot(real(p),imag(p),'*');grid xlabel('Real Axis'); ylabel('Imag Axis')  FIGURE CP9.7 Script to assist in all three parts of the problem.  Figure CP9.7c. The phase margin is P.M. = 67.7o at ω = 0.32 rad/sec. The loop transfer function is Gc (s)G(s)H(s) =  K̄1 + K̄2 s −2T s e s2  where T is the one-way time delay. If the phase lag introduced by the delay is greater than 67.7o at ω = 0.32 rad/sec, then the system will become unstable. So, since the phase lag due to the time delay T̃ is φ(ω) = ω T̃ we have 67.7o π/180 = 0.32(2T ) where T̃ = 2T . Solving for T yields T = 1.82 seconds. This is the maximum allowable one-way time delay. Executing the third part of the m-file script in Figure CP9.7a generates the plot illustrating the movement of the closed-loop system roots as the time delay is varied. The plot is shown in Figure CP9.7d. Examining the root locations, we find that when T = 1.9, the closed-loop roots  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  515  Computer Problems  1.2  1  Amplitude  0.8  0.6  0.4  0.2  0 0  5  10  15  20  25  30  35  40  Time (secs)  FIGURE CP9.7 CONTINUED: (b) Step response without time delays meets specs.  Gain dB  20  0 -20  -40 10-1  100  101  Frequency (rad/sec)  Phase deg  0 -100 -200 -300 10-1  100  101  Frequency (rad/sec)  FIGURE CP9.7 CONTINUED: (c) System Bode plot shows P.M. = 67.7o .  are s1 = −4.56, s2,3 = −0.94 ± 2.02j, s4 = −0.19, and s5,6 = ±0.32j. Therefore, the system is marginally stable when T = 1.9, and is unstable as the time delay increases.  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  516  CHAPTER 9  Stability in the Frequency Domain  4 * *  3  *  *  *  *  *  2  Imag Axis  1 0  *  *  *  *  *  ** ** ** ** *** *** **** *** *** * *********** ** ****** ** **********  * * * * * * * * * * * * * * **********************  -1 -2 -3  *  *  *  *  *  *  * *** *** **** * * ** *** ** ** **  * *  -4 -8  -7  -6  -5  -4  -3  -2  -1  0  1  Real Axis  FIGURE CP9.7 CONTINUED: (d) Closed-loop root locations as the time delay varies.  CP9.8  The Nyquist plot and associated m-file code are shown in Figure CP9.8.  Nyquist Diagram 150  a=[0 1;-1 -15]; b=[0;30]; c=[8 0]; d=[0]; sys=ss(a,b,c,d); nyquist(sys)  Imaginary Axis  100  50  0  −50  −100  −150 −50  FIGURE CP9.8 Using the Nyquist function to obtain a Nyquist plot.  0  50  100 Real Axis  150  200  250  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  517  Computer Problems  CP9.9  The Nichols chart is shown in Figure CP9.9. The phase and gain margins are 37.1 degrees and ∞ dB, respectively.  a=[0 1;-1 -10]; b=[0;22]; c=[10 0]; d=[0]; sys=ss(a,b,c,d); nichols(sys) ngrid Nichols Chart 60  Open-Loop Gain (dB)  40 0.25 dB 0.5 dB 1 dB 3 dB 6 dB  20  0 dB ?-1 dB ?-3 dB ?-6 dB  0  ?-12 dB -20  ?-20 dB  -40  ?-40 dB  -60 -360  ?-60 dB -315  -270  -225  -180  -135  -90  -45  0  Open-Loop Phase (deg)  FIGURE CP9.9 The Nichols chart for the system in CP9.8.  CP9.10  (a) The Nyquist plot is shown in Figure CP9.10. The phase margin is P.M. = 18o . (b) When the time delay is T = 0.05 seconds, the phase margin is P.M. = 9o . (c) When the time delay is T = 0.1 seconds, the system is marginally stable. So, for T > 0.1 seconds, the system is unstable.  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  CHAPTER 9  Stability in the Frequency Domain  Nyquist Diagram 200 150 100 Imaginary Axi s  518  50  -1 point  0 -50 -100 -150 -200 -10  -8  -6  -4 Real Axi s  FIGURE CP9.10 Nyquist plot for G(s)H(s) =  10 . s(s+1)  -2  0  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  C H A P T E R  1 0  The Design of Feedback Control Systems  Exercises E10.1  From the design specifications, we determine that our desired ζ = 0.69 and ωn = 5.79. The characteristic equation is 1 + Gc (s)G(s) = 1 +  K(s + a) =0, s(s + 2)  or s2 + (2 + K)s + Ka = 0 . Our desired characteristic polynomial is s2 + 2ζωn s + ωn2 = s2 + 8s + 33.6 = 0 . Thus, K + 2 = 8, or K=6 and Ka = 33.6, so a = 5.6. The actual percent overshoot and settling time will be different from the predicted values due to the presence of the closed-loop system zero at s = −a. In fact, the actual percent overshoot and settling time are P.O. = 12.6% and Ts = 0.87s, respectively. E10.2  The characteristic equation is 400 1 1 + Gc (s)G(s) = 1 + K1 + s(s + 40) s     =1+  400(K1 s + 1) =0, s2 (s + 40)  or 1 + K1  400s =0. s3 + 40s2 + 400 519  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  520  CHAPTER 10  The Design of Feedback Control Systems  We desire ζ = 0.45 for an overshoot of 20%. The root locus is shown in Figure E10.2. We select a point slightly inside the performance region (defined by ζ = 0.45 ) to account for the zero. Thus, K1 = 0.5 and the closed-loop poles are s1 = −35 s2,3 = −2.7 ± j2 . The actual P.O. = 20.7% .  50 40 30  Imag Axis  20 10 0  x  * *  *  x o x  -10 -20 -30 -40 -50 -50  -40  -30  -20  -10  0  Real Axis  FIGURE E10.2 400s Root locus for 1 + K1 s3 +40s 2 +400 = 0.  E10.3  The step response is shown in Figure E10.3 for τ = 1 and K = 0.5. It can be seen that the P.O. = 4% , so this is a valid solution.  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  521  Exercises  1.2  1  Amplitude  0.8  0.6  0.4  0.2  0  -0.2  1  0  2  3  4  5  6  7  8  Time (secs)  FIGURE E10.3 Step response for K = 0.5 and τ = 1.  The Bode plot is shown in Figure E10.4. The phase and gain margins are marked on the plot, where it can be seen that P.M. = 75.4o and G.M. = 28.6 dB. Bode Diagram Gm = 28.6 dB (at 11.8 rad/sec) , Pm = 75.4 deg (at 0.247 rad/sec)  Magnitude (dB)  150 100 50 0 -50 -100 -150 -45 Phase (deg)  E10.4  -90 -135 -180 -225 -270 -4 10  10  -2  10  0  Frequency (rad/sec)  FIGURE E10.4 Bode plot for Gc (s)G(s) =  100(s+0.15)(s+0.7) . s(s+5)(s+10)(s+0.015)(s+7)  10  2  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  522  CHAPTER 10  E10.5  The Design of Feedback Control Systems  We require that Kv ≥ 2.7, ζ = 0.5 and ωn = 3 for the dominant roots. We want to place a zero to left of the pole at -2, so the complex roots will dominate. Set the zero at s = −2.2. Then for the desired roots find the location of pole p in compensator Gc (s) =  K1 (s + 2.2) (s + p)  to satisfy 180o phase at the desired roots. This yields p = 16.4. Using root locus methods, we find that KK1 = 165.7, so with K1 = 7.53, we determine that K = 22, and Gc (s) =  7.46(s + 2.2) . (s + 16.4)  Then Kv = 2.78 . E10.6  The closed-loop transfer function is T (s) =  326(s + 4) Gc (s)G(s) = 4 . 3 1 + Gc (s)G(s) s + 14.76s + 151.3s2 + 349.8s + 1304  The roots are s1,2 = −0.87 ± j3.2 s3,4 = −6.5 ± j8.7 . Assuming s1,2 dominates, then we expect overshoot P.O. = 43%  and Ts = 4.6 sec .  The discrepencies with the actual P.O. and Ts are due to the poles s3,4 and the zero at s = −4. E10.7  The loop transfer function is L(s) =  Ke−0.6s . s(s + 20)  A plot of P.M. as a function of K is shown in Figure E10.7. It can be seen that P.M. = 40o when K = 26.93.  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  523  Exercises  phase margin versus K (PM=40º, K=26.93) 90 80  Phase Margin deg  70 60 50 40 30 20 10 0  0  5  10  15 K  20  25  30  FIGURE E10.7 Plot of phase margin versus K.  E10.8  The open-loop transfer function is G(s) =  2257 806071.4 = , s(0.0028s + 1) s(s + 357.14)  and the compensator is Gc (s) =  K1 (s + z) , s  where z = K2 /K1 . The characteristic equation is s3 + 357.14s2 + K1 s + K2 = 0 . Using Routh-Hurwitz methods, the system is stable for 0 < K2 < 357.14 K1 or K2 /K1 < 357.14. Select the zero z at s = −10, then using root locus methods we determine that K1 = 0.08 and K2 = 0.8. The roots of the characteristic equation are s1 = −10.6  and s2,3 = −175 ± j175 ,  and ζ = 0.707, as desired. The step response is shown in Figure E10.8.  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  524  CHAPTER 10  The Design of Feedback Control Systems  1.2  1  Amplitude  0.8  0.6  0.4  0.2  0 0  0.05  0.1  0.15  0.2  0.25  0.3  0.35  0.4  0.45  0.5  Time (secs)  FIGURE E10.8 Step response with K1 = 0.08 and K2 = 0.8.  E10.9  The loop transfer function is L(s) = Gc (s)G(s) =  K1 (s + K2 /K1 ) , s(s + 1)  and Kv = lim sGc (s)G(s) = K2 . s→0  Select K2 = 5. The characteristic equation is s2 + (K1 + 1) + K2 = 0 , and we want s2 + 2ζωn s + ωn2 = 0 . √ √ Equating coefficients yields ωn = K2 = 5. Also, since we want P.O. = 5%, we require ζ = 0.69. Thus, 2ζωn = K1 + 1  implies K1 = 2.08 . √ The step response with K1 = 2.08 and K2 = 5 yields a P.O. > 5%. This  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  525  Exercises  is due to the zero at s = −1.08 . So, we raise the gain K1 = 3 and then the P.O. = 5%. The step response is shown in Figure E10.9.  1.2  1  Amplitude  0.8  0.6  0.4  0.2  0 0  0.5  1  1.5  2  2.5  3  Time (secs)  FIGURE E10.9 Step response with K1 = 3 and K2 = 5.  E10.10  The loop transfer function is L(s) = Gc (s)G(s) =  (KP s + KI ) . s(s + 1)(s + 2)  Let KI = 2. Then, the plot of the phase margin as a function of KP is shown in Figure E10.10, where it can be seen that P.M. = 71.6o is the maximum achievable phase margin. When KP = 1.54 and KI = 2 we have P.M. = 60o , as desired, and P.O. = 9% and Tp = 3.4 sec.  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  526  CHAPTER 10  The Design of Feedback Control Systems  75  Phase Margin (deg)  70  65  60  55  50  1  2  3  4  5  6  7  8  9  10  KP  FIGURE E10.10 Phase margin versus KP with KI = 2.  The Nichols diagram and the closed-loop Bode plot are shown in Figures E10.11a and E10.11b, respectively.  40  0 0.25  30 0.5 1  20  -1  2.3 10  Gain dB  E10.11  -3 -6  0  -12  -10  -20  -20 -30 -40  -350  -300  -250  -200  -150  -100  Phase (deg)  FIGURE E10.11 (a) Nichols diagram for Gc (s)G(s) =  1350(1+0.25s) . s(s+2)(s+30)(1+0.025s)  -50  -40 0  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  527  Exercises  Bode Diagram 20  Magnitude (dB)  0 −20 −40 −60 −80 −100 0  Phase (deg)  −45 −90 −135 −180 −225 −270 0 10  1  2  10  3  10  10  Frequency (rad/sec)  FIGURE E10.11 CONTINUED: (b) Closed-loop Bode plot.  E10.12  The loop transfer function is L(s) = Gc (s)G(s) =    KK1 s +  1 2  s2 (s + 5)    .  When KK1 = 5.12, the roots are s1,2 = −0.58 ± j0.58 s3 = −3.84 . The complex poles have ζ = 0.707 and the predicted settling time is Ts = 4/0.58 = 6.89 sec . The actual settling time is Ts = 6.22 s. E10.13  For the cascade compensator, we have T1 (s) =  Gc (s)G(s) 8.1(s + 1) = , 1 + Gc (s)G(s) (s + r1 )(s + r̂1 )(s + r2 )  where r1 = −1 + j2 and r2 = −1.67. For the feedback compensator, we  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  528  CHAPTER 10  The Design of Feedback Control Systems  have T2 (s) =  8.1(s + 3.6) G(s) = , 1 + Gc (s)G(s) (s + r1 )(s + r̂1 )(s + r2 )  where G(s) =  8.1 s2  and Gc (s) =  s+1 . s + 3.6  The response of the two systems differ due to different value of the zero of T1 and T2 , however, both systems have the same characteristic equation. E10.14  The Bode plot (with the lag network) is shown in Figure E10.14; the phase margin is P.M. = 46o . Bode Diagram Gm = 21.9 dB (at 1.84 rad/sec) , Pm = 46.4 deg (at 0.344 rad/sec) 100  Magnitude (dB)  50 0 −50 −100 −150 −90  Phase (deg)  −135  −180  −225  −270 −4 10  −3  10  FIGURE E10.14 Bode plot for Gc (s)G(s) =  E10.15  −2  10  −1  0  10 Frequency (rad/sec)  5(7.5s+1) s(s+1)(0.25s+1)(110s+1)  10  1  10  2  10  = 0.  At the desired crossover frequency ωc = 10 rad/sec, we have 20 log |Gc (j10)G(j10)| = −8.1 dB and 6  Gc (j10)G(j10) = −169o .  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  529  Exercises  Therefore, the phase margin is P.M. = 11o . So, φ = 30o − 11o = 19o  and  M = 8.1 dB .  Since φ > 0 and M > 0, a lead compensator is required. E10.16  At the desired crossover frequency ωc = 2 rad/sec, we have 20 log |Gc (j2)G(j2)| = 17 dB and 6  Gc (j2)G(j2) = −134o .  Therefore, the phase margin is P.M. = 46o . So, φ = 30o − 46o = −16o M = −17 dB . Since φ < 0 and M < 0, a lag compensator is required. E10.17  Using a prefilter Gp (s) =  KI KP s + KI  the closed-loop transfer function is T (s) =  s2  KI . + (KP + 1)s + KI  The required coefficients for a deadbeat system are α = 1.82 and Ts = 4.82. Therefore, KI = ωn2 KP = αωn − 1 . Since we desired a settling time less than 2 seconds, we determine that ωn = Ts /2 = 4.82/2 = 2.41 . Then, the gains are KP = 3.39 KI = 5.81 . The step response (with the prefilter) is shown in Figure E10.17. The percent overshoot is P.O. = 0.098% and the settling time is Ts = 1.99 seconds.  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  530  CHAPTER 10  The Design of Feedback Control Systems  1.2  1  Amplitude  0.8  0.6  0.4  0.2  0 0  0.5  1  1.5  2  2.5  3  3.5  4  4.5  5  Time (secs)  FIGURE E10.17 Step response for the deadbeat system.  E10.18  Consider the PI controller Gc (s) = Kp +  Kp s + KI 30s + 300 KI = = s s s  and the prefilter Gp (s) = 10 . Then, the closed-loop system is T (s) =  s2  300s + 3000 . + 280s + 3000  The percent overshoot is P.O. = 9.2% and the settling time Ts = 0.16 seconds. The steady-state tracking error to a unit step is zero, as desired. E10.19  Consider the PID controller Gc (s) = 29  s2 + 10s + 100 . s  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  531  Exercises  The closed-loop transfer function is 29(s2 + 10s + 100) . s3 + 24s2 + 290s + 2900  T (s) =  The settling time to a unit step is Ts = 0.94 seconds. E10.20  Consider the PD controller Gc (s) = KD s + Kp = 3s + 1 . The loop transfer function is L(s) = Gc (s)G(s) =  3s + 1 . s(s − 2)  The Bode plot is shown in Figure E10.20. The phase margin is P.M. = 40.4◦ . This is a situation where decreasing the gain leads to instability. The Bode plot shows a negative gain margin indicating that the system gain can be decreased up to -3.5 dB before the closed-loop becomes unstable. Bode Diagram Gm = −3.52 dB (at 0.816 rad/sec) , Pm = 40.4 deg (at 2.28 rad/sec)  Magnitude (dB)  40 20 0 −20  Phase (deg)  −40 −90 −135 −180 −225 −270 −2 10  −1  10  0  10 Frequency (rad/sec)  FIGURE E10.20 Bode plot for the loop transfer function L(s) =  E10.21  3s+1 . s(s−2)  The transfer function from Td (s) to Y (s) is T (s) =  s2  1 . + 4.4s + K  1  10  2  10  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  CHAPTER 10  The Design of Feedback Control Systems  The tracking error is E(s) = R(s) − Y (s). When R(s) = 0, then E(s) = −Y (s). The final value of the output to a unit step disturbance is ess = 1/K. If we want the tracking error to be less than 0.1, then we require K > 10. When K = 10, we have the disturbance response shown in Figure E10.21.  Step Response 0.12  0.1  0.08 Amplitude  532  0.06  0.04  0.02  0  0  0.5  FIGURE E10.21 Disturbance response for K = 10.  1  1.5 Time (sec)  2  2.5  3  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  533  Problems  Problems P10.1  (a) The loop transfer function is L(s) = Gc (s)G(s)H(s) =  (1 + ατ s)K1 K2 . α(1 + τ s)(Js2 )  We desire ζ = 0.6, Ts ≤ 2.5 or ζωn ≥ 1.6. The uncompensated closedloop system is T (s) =  K , s2 + K  where K = K1 K2 /J and K = ωn2 . We can select K = 20, and then ζωn > 1.6. First, plot the Bode diagram for G(s)H(s) =  20 s2  where K1 K2 /αJ = 20. The phase margin of the uncompensated system is 0o . We need to add phase at ωc . After several iterations, we choose to add 40o phase at ωc , so sin 40o =  α−1 = 0.64 . α+1  Therefore, α = 4.6. Then, 10 log α = 10 log 4.6 = 6.63dB . We determine the frequency where magnitude is -6.63 dB to be ωm = 6.6 rad/sec. Then, √ p = ωn α = 14.1 and z = p/α = 3.07 . The compensated loop transfer function (see Figure P10.1a) is 20 Gc (s)G(s)H(s) = 2 s  s 3.07 s 14.1    +1  . +1  (b) Since we desire ζωn ≥ 1.6, we place the compensator zero at z = 1.6. Then, we place the compensator pole far in the left half-plane; in this case, we selected p = 20. Thus, the compensator is Gc (s) =  s + 1.6 . s + 20  The root locus is shown in Figure P10.1b. To satisfy the ζ = 0.6 requirement, we find K = 250, and the compensated loop transfer  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  CHAPTER 10  The Design of Feedback Control Systems  Gain dB  100  50  0 -50 10-1  100  101  102  101  102  Frequency (rad/sec)  Phase deg  -140 -150 -160 -170 -180 10-1  100 Frequency (rad/sec)  FIGURE P10.1 (a) Compensated Bode plot for Gc (s)G(s)H(s) =  20(s/3.07+1) . s2 (s/14.1+1)  function is 20 250(s + 1.6) Gc (s)G(s)H(s) = 2 = 2 s (s + 20) s  s 1.6 + 1 s 20 + 1    .  20 15 *  10 5  Imag Axis  534  0  x  *o  x  -5 -10 *  -15 -20 -25  -20  -15  -10  -5  0  Real Axis  FIGURE P10.1 CONTINUED: (b) Root locus for Gc (s)G(s)H(s) = 1 + K s2s+1.6 . (s+20)  5  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  535  Problems  P10.2  The transfer function of the system is G(s) =  s3  1.0e + 14 , + 2000s2 + 1e + 11s  where we use the system parameters given in P7.11 with the following modifications: τ1 = τ1 = 0 and K1 = 1. Also we have scaled the transfer function so that the time units are seconds. The parameters in P7.11 are given for time in milliseconds. A suitable compensator is Gc (s) =  s + 500 . s+1  The closed-loop system response is shown in Figure P10.2. The percent overshoot is P.O. ≈ 20% and the time to settle is Ts < 0.01 second. 1.4  1.2  Amplitude  1  0.8  0.6  0.4  0.2  0  0  0.005  0.01  0.015  0.02  0.025 0.03 Time (secs)  0.035  0.04  0.045  0.05  FIGURE P10.2 Step response.  P10.3  The loop transfer function is Gc (s)G(s) =  16(s + 1) K(s + z) . s(s2 + 2s + 16) (s + p)  We desire dominant roots with Ts < 5 sec and P.O. < 5%, so use ζ = 0.69 and ζωn = 0.8. One solution is to select z = 1.1 (i.e. to the left of the  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  536  CHAPTER 10  The Design of Feedback Control Systems  existing zero at s = −1) and determine the pole p and gain K for dominant roots with ζ = 0.69. After iteration, we can select p = 100, so that the root locus has the form shown in Figure P10.3. Then, we select K = 320, 200  150  100  Imag Axis  50  0  -50  -100  -150  -200 -200  -150  -100  -50  0 Real Axis  50  100  150  200  FIGURE P10.3 16(s+1)(s+1.1) Root locus for 1 + K s(s2 +2s+16)(s+100) = 0.  so that ζ = 0.69. The final compensator is Gc (s) =  320(s + 1.1) . s + 100  The design specifications are satisfied with this compensator. P10.4  The uncompensated loop transfer function is G(s) =  1 1 s2 ( 40 s  + 1)  =  40 . + 40)  s2 (s  We desire 10% < P.O. < 20%, so 0.58 < ζ < 0.65, and Ts < 2 implies ζωn < 2. We will utilize a PD compensator Ka (s + a). We select a = 2, to obtain the root locus shown in Figure P10.4. Then with Ka = 23.5, we have the desired root location, and Gc (s) = 23.5(s + 2) . The design specifications are satisfied with the PD compensator.  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  537  Problems  30 + *  20  Imag Axis  10  0  + *o  x  x  -10  -20 + *  -30 -50  -40  -30  -20  -10  0  Real Axis  FIGURE P10.4 40(s+2) Root locus for 1 + Ka s2 (s+40) = 0.  P10.5  We desire P.O. < 10% and Ts < 1.5 sec. The compensator is a PI-type, given by Gc (s) = K2 +  K2 s + K3 K2 (s + a) K3 = = s s s  where a = K3 /K2 . So, ess = 0 for a step input and G(s) =  3.75Ka 25Ka = . (s + 0.15)(0.15s + 1) (s + 0.15)(s + 6.67)  The loop transfer function is Gc (s)G(s) =  25Ka K2 (s + a) . s(s + 0.15)(s + 6.67)  Using root locus methods, we select a = 0.2 (after several iterations) and determine Ka K2 to yield ζ = 0.65. This results in Ka K2 = 1. The root locus is shown in Figure P10.5. The design specifications are met. The actual percent overshoot and settling time are P.O. = 7.4% and Ts = 1.3 s. The controller is Gc (s) = 1 +  0.2 . s  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  538  CHAPTER 10  The Design of Feedback Control Systems  20 15 10 + *  Imag Axis  5 0  x  x+ x *o  -5  + *  -10 -15 -20 -20  -15  -10  -5  0  5  10  15  20  Real Axis  FIGURE P10.5 25(s+0.2) Root locus for 1 + Ka K2 s(s+0.15)(s+6.67) = 0.  As in P10.5, using root locus we find that placing z = 15 and p = 30 yields a root locus shape (see Figure P10.6) where the loop transfer function is  60  40  20  Imag Axis  P10.6  + *  0  x  + *  o  x  x + *  -20  -40  -60 -60  -40  -20  0 Real Axis  FIGURE P10.6 25(s+15) Root locus for 1 + Ka (s+0.15)(s+6.67)(s+30) = 0.  20  40  60  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  539  Problems  Gc (s)G(s) =  25Ka (s + z) . (s + p)(s + 0.15)(s + 6.67)  and where z, p and Ka are the parameters to be determined. Properly choosing the parameter values allows us to increase ζωn of the dominant roots (compared to the PI compensator of P10.5). Then, with Ka = 3.7, the dominant roots have ζ = 0.65. The design specifications are met with the compensator. P10.7  The plant transfer function is G(s) =  e−50s . (40s + 1)2  The steady-state error is ess =  A < 0.1A . 1 + Kp  Therefore, Kp > 9. Insert an amplifier with the compensator with a dc gain = 9, as follows Gc (s)G(s) =  9e−50s (s + 2) . (40s + 1)2 (s + p)  The system is unstable without compensation, and it is very difficult to compensate such a time delay system with a lead compensator. Consider a lag network Gc (s) =  s+z s+p  where z > p. Let z = 10p. Then, a plot of the P.M. versus p is shown in Figure P10.8a. Suitable system performance can be obtained with P.M. > 45o , so choose p = 0.0001. The Bode plot of the compensated and uncompensated systems is shown in Figure P10.7c, where we have selected z = 0.001 and p = 0.0001. The compensated system has P.M. = 62o  and  Ts = 9 minutes .  The step response is shown in Figure P10.7b.  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  540  CHAPTER 10  The Design of Feedback Control Systems (a)  Phase Margin (deg)  150 100 50 0 -50 0  0.5  1  1.5 p  2.5  3 x10 -3  (b)  1.5  Amplitude  2  1 0.5 0 0  100  200  300  400  500  600  700  800  Time (secs)  FIGURE P10.7 (a) Phase margin versus p. (b) Step response with p = 0.0001 and z = 0.001.  20  Gain dB  10 0 -10 -20 -30 10-4  10-3  10-2  10-1  10-2  10-1  Frequency (rad/sec) 0  Phase deg  -100 -200 -300 -400 -500 10-4  10-3 Frequency (rad/sec)  FIGURE P10.7 CONTINUED: (c) Bode plot for the compensated system (solid lines) and the uncompensated system (dashed line).  P10.8  The transfer function is G(s) =  5000 . s(s + 10)2  To meet the steady-state accuracy, we need Kv > 40. The uncompensated  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  541  Problems  Kv = 50, so the steady-state accuracy can be met. (a) Using the Bode method, we need P.M. = 70% (to meet P.O. < 5% specification). Let Gc (s) =  bs + 1 . as + 1  The plot of P.M. versus b is shown in Figure P10.8a, where we set a = 50b. Choosing b = 20 should satisfy the P.O. specification. The Bode plot is shown in Figure P10.8c. Thus, (a)  Phase Margin (deg)  80 70 60 50 40 30 0  5  10  15 b  25  30  (b)  1.5  Amplitude  20  1 0.5 0 0  5  10  15  20  25  30  35  40  45  50  Time (secs)  FIGURE P10.8 (a) Phase margin versus b; (b) Step response for lag compensator designed with Bode where a = 1000 and b = 20.  Gc (s)G(s) =  5000(20s + 1) . s(s + 10)2 (1000s + 1)  The step response is shown in Figure P10.8b. (b) We require that ζ = 0.7 to meet the P.O. specifications. Let Gc (s) =  K(bs + 1) . (as + 1)  Using root locus methods, we fix a and b, and then determine K for ζ = 0.7. Let a = 50b and select b = 10 (other values will work). The root locus is shown in Figure P10.8d. We find K = 2.5 when ζ = 0.7. Now, Kv = 125, so the steady-state accuracy requirement is satisfied for the step response as shown in Figure P10.8e.  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  CHAPTER 10  The Design of Feedback Control Systems  150  Gain dB  100 50 0 -50 -100 10-4  10-3  10-2  10-3  10-2  10-1 100 Frequency (rad/sec)  101  102  101  102  Phase deg  0 -100 -200 -300 10-4  10-1  100  Frequency (rad/sec)  FIGURE P10.8 CONTINUED: (c) Bode plot for the compensated system with Gc (s) =  20s+1 1000s+1 .  20 15 10 5  Imag Axis  542  *  0  *  x  *ox *  -5 -10 -15 -20 -20  -15  -10  -5  0  5  Real Axis  FIGURE P10.8 5000(10s+1) CONTINUED: (d) Root locus for 1 + K s(s+10)2 (500s+1) .  10  15  20  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  543  Problems  1.2  1  Amplitude  0.8  0.6  0.4  0.2  0 0  2  4  6  8  10  12  14  16  18  20  Time (secs)  FIGURE P10.8 CONTINUED: (e) Step response for lag compensator designed with root locus methods, where K = 2.5.  We desire a small response for a disturbance at 6 rad/sec. The Bode plot of Gc (s)G(s) is shown in Figure P10.9a where we consider a compensator  Gain dB  0 -50 -100 -150 10-1  100  101 Frequency (rad/sec)  102  103  100  101 Frequency (rad/sec)  102  103  0  Phase deg  P10.9  -100 -200 -300 10-1  FIGURE P10.9 (a) Bode plot for the compensated system with Gc (s) =  10(s2 +4s+10) . s2 +36  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  544  CHAPTER 10  The Design of Feedback Control Systems  of the form Gc (s) =  K(s2 + as + b) . s2 + 36  Notice that the magnitude is large at ω = 6, as desired. We select a = 4,  b = 10  and K = 10 .  The response to a sinusoidal disturbance at 6 rad/sec is shown in Figure P10.9b. Notice that the effect of the disturbance is virtually eliminated in steady-state.  0.02 0.015 0.01  Amplitude  0.005 0 -0.005 -0.01 -0.015 -0.02  0  10  20  30  40  50  60  70  80  90  100  Time (secs)  FIGURE P10.9 CONTINUED: (b) Disturbance response for a sinusoidal disturbance at 6 rad/sec.  P10.10  The step response with Gc (s) = 1 is shown in Figure P10.10. A suitable lag compensator is Gc (s) =  s + 0.05 . s + 0.005  The step response of the compensated system is also shown in Figure P10.10. The settling time of the compensated system is Ts = 28 seconds .  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  545  Problems  Compensated system (solid) & Uncompensated system (dashed) 30  25  Amplitude  20  15  Input (dotted line)  10  5  0 0  5  10  15  20  25  30  Time (sec)  FIGURE P10.10 Step response of uncompensated and compensated systems.  The root locus is shown in Figure P10.11 where a suitable lead-lag com-  300  200  100  Imag Axis  P10.11  +  0  x  +  oo +x +  -100  -200  -300 -300  -200  -100  0 Real Axis  FIGURE P10.11 160(s+17)(s+10) Root locus for 1 + K s2 (s+170)(s+1) = 0.  100  200  300  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  546  CHAPTER 10  The Design of Feedback Control Systems  pensator is Gc (s) = K  s + 10 s + 17 . s + 1 s + 170  The selected gain is K = 57, so that the damping of the complex roots is about ζ = 0.7. For this particular design, the closed-loop system zeros will affect the system response and the percent overshoot specification may not be satisfied. Some design iteration may be necessary or aprefilter can be utilized. A suitable prefilter is Gp (s) =  17 . s + 17  The acceleration constant is Ka = 9120. We choose K = 10. This yields a velocity constant Kv = 20K = 200, as desired. A suitable two-stage lead compensaator is Gc (s) =  (0.05s + 1)(0.05s + 1) . (0.0008s + 1)(0.0008s + 1)  The Bode plot is shown in Figure P10.12. The phase margin is P.M. = 75.06o . Phase margin=75.06 deg  100  Gain dB  50 0 -50 -100 10-1  100  101  102  103  104  103  104  Frequency (rad/sec) 0  Phase deg  P10.12  -100 -200 -300 10-1  100  101  102  Frequency (rad/sec)  FIGURE P10.12 200(0.05s+1)2 Bode plot for s(0.1s+1)(0.05s+1)(0.0008s+1)2 .  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  547  Problems  P10.13  (a) When Gc (s) = K = 0.288 , the phase margin is P.M. = 49.3o and the bandwidth is ωB = 0.95 rad/sec. (b) A suitable lag compensator is Gc (s) =  25s + 1 . 113.6s + 1  The compensated system phase margin is P.M. = 52.21o and Kv = 2, as desired. P10.14  A suitable lead compensator is Gc (s) =  1.155s + 1 . 0.032s + 1  The compensated system phase margin is P.M. = 50o and Kv = 2, as desired. The settling time is Ts = 3.82 seconds. One possible solution is Gc (s) = K  (s + 12)(s + 15) , (s + 120)(s + 150)  where K = 900. The disturbance response is shown in Figure P10.15. Step Response 0.1  0.08  Amplitude  P10.15  0.06  0.04  0.02  0  0  0.1  0.2  0.3 Time (sec )  FIGURE P10.15 Compensated system disturbance response.  0.4  0.5  0.6  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  548  CHAPTER 10  The PI controller is given by K(s + b) , s  Gc (s) =  where K and b are to be determined. To meet the design specifications, we need ζ = 0.6  and ωn = 6.67 rad/sec .  The closed-loop transfer function is T (s) =  K(s + b) . s2 + Ks + bK  Solving for the gains yields K = 2ζωn = 8 and b = ωn2 /K = 5.55. A suitable prefilter is Gp (s) =  5.55 . s + 5.55  The step response, with and without the prefilter, is shown in Figure P10.16.  Without prefilter (solid) & with prefilter (dashed) 1.4  1.2  1  Amplitude  P10.16  The Design of Feedback Control Systems  0.8  0.6  0.4  0.2  0 0  0.5  1  1.5  2  Time (sec)  FIGURE P10.16 Compensated system response with and without a prefilter.  2.5  3  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  549  Problems  P10.17  The plant transfer function is G(s) =  K . s(s + 10)(s + 50)  We desire ζωn > 10 to meet Ts < 0.4 sec and ζ = 0.65 to meet P.O. < 7.5%. Try a pole at s = −120. The root locus is shown in Figure P10.17. The gain K = 6000 for ζ = 0.65. Thus, Gc (s)G(s) =  6000(s/15 + 1) s(s + 10)(s + 50)(s/120 + 1)  and Kv =  6000 = 12 . 500  200 150 100  Imag Axis  50 0  *x  x  *  * ox x *  -50 -100 -150 -200 -200  -150  -100  -50  0  50  100  150  200  Real Axis  FIGURE P10.17 s/15+1 Root locus for 1 + K s(s+10)(s+50)(s/120+1) .  P10.18  (a) The loop transfer function is L(s) =  K1 e−2T s 0.25s + 1  where T = 1.28. The phase angle is φ = −2.56ω − tan 0.25ω . So, ω = 1.12 rad/sec when φ = −180o . However, the break frequency  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  CHAPTER 10  The Design of Feedback Control Systems  is 4 rad/sec. Therefore, you cannot achieve P.M. = 30o and have the system be stable for K1 < 1. The steady-state error is ess =  A A = 1 + Kp 1 + K1  since K1 = Kp . (b) Set K1 = 20, then Kp = 20 and this yields a 5% steady-state error. Without compensation, the system is now unstable. Let Gc (s) =  s/b + 1 s/a + 1  where b = 5 and a = 0.01. Then, the system is stable with P.M. = 63o . The system response is shown in Figure P10.18.  1.2  1  0.8  Amplitude  550  0.6  0.4  0.2  0  -0.2  0  2  4  6  8  10 Time (secs)  FIGURE P10.18 Unit step response with Gc (s) =  20(s/5+1) s/0.01+1 .  12  14  16  18  20  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  551  Problems  P10.19  (a) The open-loop transfer function is G(s) =  Ke−sT , (s + 1)(s + 3)  where T = 0.5 sec. We desire P.O. < 30%, thus ζ > 0.36. We will design for ζ = 0.4, which implies P.M. = 40o . Then φ = − tan−1 ω − tan−1  ω − 0.5ω(57.3o ) . 3  At ωc = 1.75, the phase margin is P.M. = 40o , and solving |G(jω)| =  K [(3 −  ω 2 )2  1  + (4ω)2 ] 2  =1  at ω = 1.75 yields K = 7. Then ess = 0.3. (b) We want ess < 0.12, so use ess = 0.10 as the goal. Then Gc (s)G(s) =  Ke−0.5s (s + 2) , (s + 1)(s + 3)(s + b)  and ess = where Kp =  2K 3b .  1 1 + Kp  If b = 0.1 then Kp = 6.7K and ess =  1 . 1 + 6.7K  So, we need 6.7K = 9, or K = 1.35. We need a lag compensator (i.e. b < 2) to meet ess < 12% and have stability. P10.20  We desire Kv = 20, P.M. = 45o and ωB > 4 rad/sec. Thus, we set K = 20, and G(s) =  s  s 2  20  +1  s 6   .  +1  Then, the Bode plot yields P.M. = −21o uncompensated at ωc = 5.2 rad/sec. The phase lead compensator must add 66 o plus phase lead to account for the shift of the crossover to a higher frequency with the phase lead compensator. Consider Gc (s) =    1 + ατ s 1 + τs  2  .  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  552  CHAPTER 10  The Design of Feedback Control Systems  One solution is to use α = 10 τ = 1/67 . Then Gc (s) =  100(s + 6.7)2 . (s + 67)2  The compensator has two zeros at ω = 6.7, two poles at ω = 67 yielding P.M. = 47o , ωc = 7.3 and ωB = 12 rad/sec. P10.21  We desire Kv = 20, P.M. = 45o and ωB ≥ 2. The lag compensator is Gc (jω) =  1 + jωτ 1 + jωατ  where α > 1. From the Bode plot, φ = −135o at ω ∼ = 1.3. So, at ω = 1.3, we need to lower the magnitude by 22 dB to cause ω = 1.3 to be ωc′ , the new crossover frequency. Thus, solving 22 = 20 log α yields α = 14. We select the zero one decade below ωc′ or Therefore,  1 τ  = 0.13.  1 0.13 = = 0.0093 . ατ 14 Then, the lag compensator is given by Gc (s) =  s 1 + 0.13 s + 0.13 . = s 1 + 0.0093 14(s + 0.0093)  The new crossover is ωc′ = 1.3, and ωB = 2.14 rad/sec. P10.22  We desire P.M. = 45o , Kv = 20 and 2 ≤ ωB ≤ 10. The lead-lag compensator is Gc (s) =  s 1 + sb 1 + 10a · . s 1 + 10b 1 + as  Since ωB ∼ = 1.5ωc , we design for a new crossover frequency ωc′ so that 1.4 < ωc′ < 7 . Try for ωc′ = 4. The phase φ = −190o at ω = 4, so we need to add phase lead of 55o plus phase to account for lag part of network at ωc′ . Use α = 10 and bracket ω = 4 with the lead network. Put the zero at ω = 0.8 = b  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  553  Problems  and the pole at ω = 8. For the lag compensator, put the zero at a lower frequency than ωc′ /10. So try a zero at ω = 0.2 = 10a and a pole at ω = 0.02 = a. Then, the lead-lag compensator is s s 1 + 0.8 1 + 0.2   . Gc (s) = s 1 + 8s 1 + 0.02      The compensated Bode plot yields  P.M. = 50o  ωc′ = 3.5 rad/sec, The steady-state error is  1 1 = = 0.05 . 1 + Kp 1 + K/25  ess =  So, we need K/25 ≥ 19 or K ≥ 475. One possible solution is Gc (s) =  4s + 1 12s + 1  and  K = 475 .  The compensated Bode plot is shown in Figure P10.23. The phase margin is P.M. = 46o . Bode Diagram Gm = Inf dB (at Inf rad/sec) , Pm = 46 deg (at 11.5 rad/sec)  Magnitude (dB)  40 20 0 −20 −40 0 Phase (deg)  P10.23  and ωB = 6.2 rad/sec .  −45 −90 −135 −180 −3 10  FIGURE P10.23 Bode plot for Gc (s)G(s) =  −2  10  −1  0  10 10 Frequency (rad/sec)  475(4s+1) . (s+5)2 (12s+1)  1  10  2  10  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  554  CHAPTER 10  P10.24  The Design of Feedback Control Systems  The arm-rotating dynamics are represented by G(s) =  80 s    s2 4900  +  s 70   .  +1  We desire Kv = 20, and P.O. < 10%. One possible solution is the lead-lag compensator Gc (s) =  (s + 50)(s + 0.48) . 4(s + 400)(s + 0.06)  With this compensator, we have P.O. = 9.5% P10.25  and  Kv = 20 .  Neglect the pole of the airgap feedback loop at s = 200. The characteristic equation is 1 + K̄  (s + 20)(s + c) =0, s3  where K K1 + K2 K2 b c= . K1 + K2  K̄ =  Choose c = 10 to attain the root locus structure shown in Figure P10.25. The gain K̄ = 38.87 insures the damping ratio of ζ = 0.5. Then, solving for K1 and b yields K1 =  K − K2 38.87  and b=  0.1K . 38.87K2  For given values of K and K2 (unspecified in the problem), we can compute K1 and b.  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  555  Problems  40 30 +  20  Imag Axis  10 0  o  o+  x  -10  0  -10 -20 +  -30 -40 -40  -30  -20  10  20  30  40  Real Axis  FIGURE P10.25 (s+20)(s+10) = 0. Root locus for 1 + K̄ s3  P10.26  The loop transfer function is Gc (s)G(s) =  0.15K(10as + 1) , s(s + 1)(5s + 1)(as + 1)  where K and a are to be selected to meet the design specifications. Suitable values are K = 6.25  and a = 0.15 .  Then, the phase margin is P.M. = 30.79o and the bandwidth is ωB = 0.746 rad/sec. The lead compensator is Gc (s) = 6.25 P10.27  1.5s + 1 . 0.15s + 1  (a) Let Gc (s) = K = 11. Then the phase margin is P.M. = 50o and the performance summary is shown in Table P10.27. (b) Let Gc (s) =  K(s + 12) , (s + 20)  where K = 32. Then, the phase margin is P.M. = 50o and the performance summary is given in Table P10.27.  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  556  CHAPTER 10  compensator  The Design of Feedback Control Systems  P.M.  P.O.  Tp  Ts  Mpω  ωB  Gc (s) = K = 11  50o  18%  0.34 sec  0.78 sec  1.5 dB  13.9 rad/sec  32(s+12) s+20  50o  18%  0.20 sec  0.47 sec  1.5 dB  26.3 rad/sec  TABLE P10.27  Performance Summary.  P10.28  The loop transfer function is Gc (s)G(s) =  K(as + 1) , s(s + 10)(s + 14)(10as + 1)  where K and a are to be selected to meet the design specifications, and we have set α = 10. The root locus is shown in Figure P10.28a. To satisfy  30  20  10  Imag Axis  Gc (s) =  *  0  *  x  x  *ox *  -10  -20  -30 -30  -20  -10  0 Real Axis  FIGURE P10.28 1400(s+1) (a) Root locus for 1 + K s(s+10)(s+14)(10s+1) = 0.  10  20  30  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  557  Problems  the steady-state tracking error we must select K > 1400 . Suitable values for the lag compensator are K = 4060  and a = 1 .  Then, the percent overshoot is P.O. = 31% and the settling time is Ts = 2.34 sec. The lag compensator is Gc (s) =  s+1 . 10s + 1  The step response is shown in Figure P10.28b.  1.4  1.2  Amplitude  1  0.8  0.6  0.4  0.2  0 0  0.5  1  1.5  2  Time (secs)  FIGURE P10.28 CONTINUED: (b) Step response.  P10.29  The plant transfer function is G(s) =  10e−0.05s . s2 (s + 10)  Gc (s) =  16(s + 0.7) (s + 9)  The lead network  2.5  3  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  558  CHAPTER 10  The Design of Feedback Control Systems  provides Mpω = 3.4 dB and ωr = 1.39 rad/sec. The step response is shown in Figure P10.29. The overshoot is P.O. = 37% and Ts = 3.5 sec.  1.4  1.2  Amplitude  1  0.8  0.6  0.4  0.2  0 0  1  2  3  4  5  6  Time (secs)  FIGURE P10.29 Unit step response with Gc (s) =  P10.30  16(s+0.7) . s+9  The vehicle is represented by G(s) =  K K ≈ . s(0.04s + 1)(0.001s + 1) s(0.04s + 1)  For a ramp input, we want ess 1 = 0.01 = . A Kv So, let G(s) =  100 . s(0.04s + 1)  The uncompensated P.M. = 28o at ωc = 47 rad/sec. We need to add 17o . Case (1) Phase lead compensation: Gc (s) =  1 + 0.021s . 1 + 0.01s  The phase margin is P.M. = 45o .  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  559  Problems  Case (2) Phase lead compensation: Gc (s) =  1 + 0.04s . 1 + 0.005s  The phase margin is P.M. = 65o . For Case 1, we have P.O. = 25% ,  Ts = 0.13 sec and Tp = 0.05 sec .  For Case 2, we have P.O. = 4% , P10.31  Ts = 0.04 sec and Tp = 0.03 sec .  As in P10.30, the plant is given by G(s) =  100 . s(0.04s + 1)  The uncompensated P.M. = 28o . We need P.M. = 50o . The phase lag compensator Gc (s) =  1 + 0.5s 1 + 2.5s  results in P.M. = 50o . The P.O. = 21%, Ts = 0.72 sec and Tp = 0.17 sec. P10.32  (a) To obtain Kv = 100, we have Gc (s)G(s) =  43.33(s + 500) . s(s + 0.0325)(s2 + 2.57s + 6667)  With K = 43.33, we have P.M. = 1.2o ,  Mpω = 26 dB ,  ωr = 1.8 rad/sec and ωB = 3.7 rad/sec .  The Bode plot is shown in Figure P10.32. (b) Let Gc (s) =  0.35s + 1 , 0.001s + 1  and K = 43.33 (as before). Then, P.M. = 36o ,  Mpω = 5.4 dB ,  ωr = 1.7 rad/sec and ωB = 3.0 rad/sec .  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  560  CHAPTER 10  The Design of Feedback Control Systems  100  Gain dB  50 0 -50 -100 -150 10-2  10-1  100  101  102  103  102  103  Frequency (rad/sec) -100  Phase deg  -150 -200 -250 -300 -350 10-2  10-1  100  101  Frequency (rad/sec)  FIGURE P10.32 Bode plot with Gc (s) = K = 43.33.  The step response is shown in Figure P10.33, where Gc (s) =  10(s + 0.71)(s + 0.02) . (s + 0.0017)(s + 10)  1.2  1  0.8  Amplitude  P10.33  0.6  0.4  0.2  0 0  0.5  1  1.5  2  2.5  3  3.5  4  4.5  Time (secs)  FIGURE P10.33 Step response with the lead-lag compensator Gc (s) =  10(s+0.71)(s+0.02) . (s+0.0017)(s+10)  5  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  561  Problems  Then, Kv = 80 and P.O. = 17%, Ts = 1.8 sec, and ζ = 0.54. The process model is G(s) =  s2 (s  1 , + 10)  and we consider the lead compensator Gc (s) = K  1 + sατ , 1 + sτ  where α = 100, τ = 0.4 and K = 0.5. Then, P.M. = 46.4o . The step response is shown in Figure P10.34. The system performance is P.O. = 22.7% Ts = 5.2 sec Tp = 1.72 sec .  1.4  1.2  1  Amplitude  P10.34  0.8  0.6  0.4  0.2  0  0  1  2  3  4  5 Time (secs)  6  7  FIGURE P10.34 40s+1 Step response with the lead compensator Gc (s) = 0.5 0.4s+1 .  8  9  10  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  562  CHAPTER 10  P10.35  The Design of Feedback Control Systems  The phase margin is shown in Figure P10.35. As the time delay increases, the phase margin decreases. The system is unstable when T > 2.1843 s. 140  120  100  Phase margin (deg)  80  60  40  Stability boundary  20  0  −20  0  0.5  1  1.5  2  Time delay (s)  FIGURE P10.35 Step response with Gc (s)G(s) =  2.5  where 0 ≤ T ≤ 2.5.  One possible solution is the integral controller Gc (s) = 2/s. The step response is shown in Figure P10.36. The steady-state tracking error to a  1.6 1.4 1.2 1  Amplitude  P10.36  2s+0.54 −T s e , s(s+1.76)  T=2.1843 s  0.8 0.6 0.4 0.2 0 -0.2  0  1  2  3  4  5  6  Time (secs)  FIGURE P10.36 Step response with the integral controller Gc (s) = 2/s.  7  8  9  10  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  563  Problems  step input is zero since the system is type-1. The phase margin is P.M. = 32.8◦ and the bandwidth is ωB = 4.3 rad/s . P10.37  One possible solution is Gc (s) =  1600(s + 1) . 25s + 1  The overshoot to a unit step is P.O. = 4.75% and the steady-state error to a step input is ess = 1%. The system bandwidth is ωB = 9.7 rad/sec. P10.38  The lead compensator is Gc (s) =  2.88(s + 2.04) . s + 5.88  The Bode plot is shown in Figure P10.38. The phase margin is P.M. = 30.4o at ωc = 9.95 rad/sec and the bandwidth is ωB = 17.43 rad/sec. 60  Gain dB  40 20 0 -20 -40 10-1  100  101  102  101  102  Frequency (rad/sec)  Phase deg  0 -100 -200 -300 10-1  100 Frequency (rad/sec)  FIGURE P10.38 Bode plot for Gc (s)G(s) =  P10.39  115.29(s+2.04) . s(s+2)(s+5.88)  The lag compensator is Gc (s) =  1 + 1.48s . 1 + 11.08s  The Bode plot is shown in Figure P10.39. The steady-state error specification is satisfied since Kv = 20. The phase margin is P.M. = 28.85o at  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  564  CHAPTER 10  The Design of Feedback Control Systems  Gain dB  100  50  0 -50 10-3  10-2  10-1  100  101  100  101  Frequency (rad/sec)  Phase deg  0 -100 -200 -300 10-3  10-2  10-1 Frequency (rad/sec)  FIGURE P10.39 Bode plot for Gc (s)G(s) =  40(1+1.48s) . s(s+2)(1+11.08s)  ωc = 2 rad/sec and the bandwidth is ωB = 3.57 rad/sec. P10.40  The lag compensator is Gc (s) =  2.5(1 + 1.64s) . 1 + 30.5s  The Bode plot is shown in Figure P10.40. The steady-state error specification is satisfied since Kv = 50 . The phase margin is P.M. = 28.93o at ωc = 1.98 rad/sec and the bandwidth is ωB = 3.59 rad/sec.  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  565  Problems  Gain dB  100  50  0  -50 10-3  10-2  10-1  100  101  100  101  Frequency (rad/sec)  Phase deg  0 -100 -200 -300 10-3  10-2  10-1 Frequency (rad/sec)  FIGURE P10.40 Bode plot for Gc (s)G(s) =  P10.41  100(1+1.64s) . s(s+2)(1+30.5s)  We use Table 10.2 in Dorf & Bishop to determine the required coefficients α = 1.9  and β = 2.2 .  Also, ωn Tr = 4.32  implies  ωn = 4.32,  since we require Tr = 1 second. The characteristic equation is s3 + 8.21s2 + 41.06s + 80.62 = s3 + (1 + p)s2 + (K + p)s + Kz = 0 . Equating coefficients and solving yields p = 7.21 P10.42  K = 33.85  z = 2.38 .  From Example 10.4 in Dorf & Bishop, we have the closed-loop transfer function T (s) =  (s2  96.5(s + 4) . + 8s + 80)(s + 4.83)  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  566  CHAPTER 10  The Design of Feedback Control Systems  A suitable prefilter is Gp (s) =  4 . s+4  The step response (with and without the prefilter) is shown in Figure P10.42. With prefilter (solid) & without prefilter (dashed) 1.4  1.2  Amplitude  1  0.8  0.6  0.4  0.2  0 0  0.1  0.2  0.3  0.4  0.5  0.6  0.7  0.8  0.9  1  Time (sec)  FIGURE P10.42 Step response with and without the prefilter.  P10.43  Let K = 100. The Bode plot is shown in Figure P10.43a and the response to a simusoidal noise input with ω = 100 rad/s is shown in Figure P10.43b.  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  567  Problems  Bode Diagram 60  40  Magnitude (dB)  20  0  −20 System: sysg Frequency (rad/sec): 100 Magnitude (dB): −40.1  −40  −60  −80 −1 10  0  10  1  2  10 Frequency (rad/sec)  3  10  10  0.07 0.06 0.05  Amplitude  0.04 0.03 0.02 0.01 0 −0.01  −0.02  0  1  2  3  4 Time (sec)  5  6  7  8  FIGURE P10.43 (a) Bode magnitude plot. (b) Response to a noise input.  P10.44  For 0.129 < K ≤ 69.87, the system is unstable. The percent overshoot is shown in Figure P10.44 .  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  CHAPTER 10  The Design of Feedback Control Systems  Percent Overshoot  100  50  0  −50  0  0.02  0.04  0.06  0.08  0.1  0.12  0.14  85  90  95  100  K 160 Percent Overshoot  568  140  120  100 65  70  75  80 K  FIGURE P10.44 Percent overshoot.  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  569  Advanced Problems  Advanced Problems AP10.1  (a) With Gc (s) = K , the closed-loop transfer function is T (s) =  s3  +  K . + 4s + K  5s2  When K = 2.05, the characteristic equation is s3 + 5s2 + 4s + 2.05 = 0 with poles at s = −4.1563 and s = −0.4219 ± j0.5615. Therefore ζ = 0.6, and the predicted overshoot is √  P.O. = 100e−π0.6/  1−0.62  = 9.5% < 13% .  The actual overshoot is P.O. = 9.3% and Ts = 8.7 seconds. (b) When Gc (s) =  82.3(s + 1.114) s + 11.46  the closed-loop transfer function is 82.3(s + 1.114) + + 61.3s2 + 128.14s + 91.6822 82.3(s + 1.114) = . (s + 1.196)(s + 12.26)(s + 1.5 ± j2)  T (s) =  s4  16.46s3  Therefore ζ = 0.6 and the predicted overshoot is P.O. = 9.5% < 13%. The actual overshoot is P.O. = 12% and Ts = 2.5 seconds. AP10.2  The lag network is given by Gc =  K(s + a1 ) . s + a2  The closed-loop transfer function is T (s) = K  s4  + (5 + a2  )s3  s + a1 . + (4 + 5a2 )s2 + (4a2 + K)s + Ka1  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  570  CHAPTER 10  The Design of Feedback Control Systems  Computing the steady-state tracking error yields s4 + (5 + a2 )s3 + (4 + 5a2 )s2 + 4a2 s s→0 s5 + (5 + a2 )s4 + (4 + 5a2 )s3 + (4a2 + K)s2 + Ka1 s 4a2 = < 0.125 . a1 K  ess = lim  If we select K = 2.05 (as in AP10.1), then a1 > 15.61a2 . So, take a2 = a1 /16. The lag compensator can now be written as Gc (s) = 2.05  s + a1 . s + a1 /16  Select a1 = 0.018. Then, the closed-loop transfer function is T (s) =  s4  +  5.0011s3  2.05(s + 0.018) . + 4.0056s2 + 2.0545s + 0.0369  The performance results are P.O. = 13% and Ts = 29.6 seconds for a step input, and ess = 0.12 for a ramp input. AP10.3  The plant transfer function is G(s) =  1 s(s + 1)(s + 4)  and the PI controller is given by Gc (s) =  Kp s + KI . s  The closed-loop transfer function is T (s) =  s4  +  5s3  Kp s + KI . + 4s2 + Kp s + KI  For a unit ramp, the steady-state tracking error is s4 + 5s3 + 4s2 =0. s→0 s5 + 5s4 + 4s3 + Kp s2 + KI s  ess = lim  Any KI > 0 and Kp > 0 (such that the system is stable) is suitable and will track a ramp with zero steady-state error. Since we want P.O. < 13%, the damping of the dominant roots should be ζ ≈ 0.6. One suitable  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  571  Advanced Problems  solution is to place the zero at s = −0.01 and select the PI controller 2.05(s + 0.01) . s  Gc (s) =  Therefore, Kp = 2.05 and KI = 0.0205. The closed-loop transfer function is T (s) =  s4  +  5s3  2.05(s + 0.01) . + 4s2 + 2.05s + 0.0205  The performance results are P.O. = 11.5% and Ts = 9.8 seconds for a step input, and ess = 0 for a unit ramp. AP10.4  The closed-loop transfer function is T (s) =  10K1 . s2 + 10(1 + K1 K2 )s + 10K1  From the performance specifications, we determine that the natural frequency and damping of the dominant poles should be ωn = 5.79 and ζ = 0.69. So, s2 + 8(1 + K1 K2 )s + 8K1 = s2 + 2ζωn s + ωn2 = s2 + 7.99s + 33.52 . Solving for the gains yields K1 = 4.19 and K2 = 0. The closed-loop transfer function is T (s) =  s2  33.52 . + 8s + 33.52  The performance results are P.O. = 5% and Ts = 1 second. AP10.5  (a) From the overshoot specification P.O. = 10%. The plant transfer function is G(s) =  1 . s(s + 1)(s + 10)  Let Gp = 1. A suitable compensator is Gc = K  s + 0.5 . s + 10  Using root locus methods, we determine that K = 45 yields P.O. ≈ 10%. The closed-loop poles are s1,2 = −2.5 ± j5.1, s3 = −15.48, and s4 = −0.45. (b) The closed-loop transfer function is T (s) =  s4  +  21s3  450(s + 0.5) . + 120s2 + 550s + 225  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  572  CHAPTER 10  The Design of Feedback Control Systems  The step response is shown in Figure AP10.5. The overshoot and settling time are P.O. = 9.5% and Ts = 3.4 seconds. (c) A suitable prefilter is Gp (s) =  0.5 . s + 0.5  The closed-loop transfer function is T (s) =  s4  +  21s3  225 . + 120s2 + 550s + 225  The step response is shown in Figure AP10.5. The overshoot and settling time are P.O. = 0% and Ts = 6.85 seconds.  1.4  1.2  Amplitude  1  0.8  0.6  0.4  0.2  0  0  1  2  3  4 Time (sec)  5  6  7  8  FIGURE AP10.5 Step response with prefilter (dashed line) and without prefilter (solid line).  AP10.6  From Example 10.12 in Dorf & Bishop, we have the relationship ωn Ts = 4.04 . Thereore, minimizing Ts implies maximizing ωn . Using Table 10.2 in Dorf & Bishop, we equate the desired and actual characteristic polynomials q(s) = s3 + 1.9ωn s2 + 2.2ωn2 s + ωn3 = s3 + (1 + p)s2 + (K + p)s + Kz .  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  573  Advanced Problems  Comparing coefficients yields (1 + p) = 1.9ωn ,  K + p = 2.2    1+p 1.9  2  Kz = ωn3 .  ,  So, from the first relationship we see that maximizing ωn implies maximizing p. Solving for p while maintaining K < 52 K=  2.2 2 (p + 2p + 1) − p < 52 3.61  we determine that −9.3643 < p < 9.005 . The largest p = 9. Therefore, K = 51.94 and z = 2.81. The step response is shown in Figure AP10.6. The settling time is Ts = 0.77 second.  1.2  1  Amplitude  0.8  0.6  0.4  0.2  0 0  0.2  0.4  0.6  0.8  1  1.2  1.4  1.6  1.8  Time (secs)  FIGURE AP10.6 Step response with minimum settling time.  AP10.7  Let Gp = 1. The closed-loop transfer function is T (s) =  K(s + 3) . s4 + 38s3 + 296s2 + (K + 448)s + 3K  When K = 311, the characteristic equation s4 + 38s3 + 296s2 + 759s + 933 = 0  2  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  574  CHAPTER 10  The Design of Feedback Control Systems  √ has poles at s = −1.619 ± j1.617 (ζ = 1/ 2), s = −6.25, and s = −28.51. (a) When Gp (s) = 1 and K = 311, the closed-loop transfer function is T (s) =  311(s + 3) . s4 + 38s3 + 296s2 + 759s + 933  The step input performance is P.O. = 6.5%, Ts = 2.5 seconds, and Tr = 1.6 seconds. With the prefilter Gp (s) =  3 s+3  and K = 311, the closed-loop transfer function is T (s) =  s4  +  38s3  933 . + 296s2 + 759s + 933  In this case, the step response is P.O. = 3.9%, Ts = 2.8 seconds, and Tr = 1.3 seconds. (b) Now, consider the prefilter Gp (s) =  1.8 s + 1.8  and K = 311. The closed-loop transfer function is (s) =  s5  +  39.8s4  559.8(s + 3) . + 364.4s3 + 1291.8s2 + 2299.2s + 1679.4  The step input response is P.O. = 0.7%, Ts = 2.14 seconds and Tr = 1.3 seconds. AP10.8  The plant transfer function is G(s) =  250 . s(s + 2)(s + 40)(s + 45)  The performance specifications are P.O. < 20%, Tr < 0.5 second, Ts < 1.2 seconds and Kv ≥ 10. A suitable lead compensator is Gc = 1483.7  s + 3.5 . s + 33.75  The closed-loop transfer function is T (s) =  250(1483.7)(s + 35) s(s + 2)(s + 40)(s + 45)(s + 33.75) + 250(1483.7)(s + 3.5)  The actual step input performance (see Figure AP10.8) is P.O. = 18%, Ts = 0.88 second, Tr = 0.18 second, and Kv = 10.7.  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  575  Advanced Problems  1.2  1  Amplitude  0.8  0.6  0.4  0.2  0 0  0.1  0.2  0.3  0.4  0.5  0.6  0.7  0.8  0.9  1  Time (secs)  FIGURE AP10.8 Step response with lead compensator.  The frequency response is shown in Figure AP10.9.  Bode Diagrams Gm=12.4 dB (Wcg=20.9); Pm=42.0 deg. (Wcp=9.0) 150 100 50 0 Phase (deg); Magnitude (dB)  AP10.9  -50 -100 -150 -50 -100 -150 -200 -250 -300 -3 10  10  -2  10  -1  10  0  Frequency (rad/sec)  FIGURE AP10.9 Bode plot with Gc (s) =  (s+2.5)(s+0.9871) (s+36.54)(s+0.0675)  10  1  10  2  10  3  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  576  CHAPTER 10  The Design of Feedback Control Systems  One lead-lag compensator that satisfies the specifications is Gc (s) =  (s + 2.5)(s + 0.9871) . (s + 36.54)(s + 0.0675)  The gain and phase margins are Gm = 12.35 dB and P m = 41.8◦ , respectively. The velocity error constant is Kv = 100. Therefore, all specifications are satisfied.  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  577  Design Problems  Design Problems The plant model with parameters given in Table CDP2.1 in Dorf and Bishop is given by: 26.035 θ(s) = , Va (s) s(s + 33.142) where we neglect the motor inductance Lm and where we switch off the tachometer feedback (see Figure CDP4.1 in Dorf and Bishop). With a PD controller the closed-loop system characteristic equation is s2 + (33.142 + 26.035KD )s + 26.035Kp = 0 . Using Table 10.2 in Dorf and Bishop we determine that for a second-order system with a deadbeat response we have α = 1.82 and ωn Ts = 4.82. Since we desire Ts < 0.25 seconds, we choose ωn = 19.28. Equating the actual characteristic equation with the desired characteristic equation we obtain s2 + ωn αs + ωn2 = s2 + (33.142 + 26.035KD )s + 26.035Kp . Solving for Kp and KD yields the PD controller: Gc (s) = 14.28 + 0.075s . The step response is shown below. The settling time is Ts = 0.24 second. 1 0.9 0.8 0.7 0.6 Amplitude  CDP10.1  0.5 0.4 0.3 0.2 0.1 0  0  0.05  0.1  0.15 Time (secs)  0.2  0.25  0.3  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  578  CHAPTER 10  The plant is given as G(s) =  20 . s (s + 2)  One possible lead compensator is Gclead (s) =  50(s + 1) . s + 20  Similarly, a suitable lag compensator is Gclag (s) =  s + 0.1 . s + 0.022  The loop transfer function with the lead-lag compensator is Gc (s)G(s) =  1000(s + 1)(s + 0.1) . s (s + 2) (s + 0.022)(s + 20)  The step response and ramp response are shown in Figure DP10.1. The velocity constant is Kv = 50, so the steady-state error specification is satisfied.  Step response  1.5  1  0.5  0  0  0.1  0.2  0.3  0.4  0.5 0.6 Time (sec)  0.7  0.8  0.9  1  0  0.1  0.2  0.3  0.4  0.5 0.6 Time (sec)  0.7  0.8  0.9  1  1 Ramp response  DP10.1  The Design of Feedback Control Systems  0.8 0.6 0.4 0.2 0  FIGURE DP10.1 Step response and ramp response.  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  579  Design Problems  (a) When Gc (s) = K, we require K > 20 to meet the steady-state tracking specification of less than 5%. (b) The system is unstable for K > 20. (c) A single stage lead compensator is Gc1 (s) =  1 + 0.49s . 1 + 0.0035s  With this compensator, the bandwidth is ωB = 68.9 rad/sec and the phase margin is P.M. = 28.57o . (d) A two stage lead compensator is Gc2 (s) =  (1 + 0.0185s)(1 + 0.49s) . (1 + 0.00263s)(1 + 0.0035s)  With the two stage compensator, the bandwidth is ωB = 83.6 rad/sec and the phase margin is P.M. = 56.79o . The step response for the two compensators is shown in Figure DP10.2. Single stage (solid) & two stage (dashed) 1.4  1.2  1  Amplitude  DP10.2  0.8  0.6  0.4  0.2  0 0  0.1  0.2  0.3  0.4  0.5  0.6  Time (sec)  FIGURE DP10.2 Step response for one- and two-stage lead compensators.  0.7  0.8  0.9  1  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  580  CHAPTER 10  The mast flight system is modeled as 6 . s(s + 1.5)(s + 3.9)  G(s) =  Consider the proportional controller Gc (s) = K = 0.85 . The system step response is shown in Figure DP10.3. The percent overshoot is P.O. = 15.9%, the rise time is Tr = 3.63 seconds, and the phase margin is P.M. = 52o .  Step Response 1.4  System: syscl Peak amplitude: 1.16 Overshoot (%): 15.9 At time (sec): 3.63  1.2  1 Amplitude  DP10.3  The Design of Feedback Control Systems  0.8  0.6  0.4  0.2  0  0  2  4  FIGURE DP10.3 Step response for the mast flight system.  6 Time (sec)  8  10  12  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  581  Design Problems  DP10.4  One possible compensator is Gc (s) = 5682  s + 12.6 . s + 87.3  The step response is shown in Figure DP10.4. The performance results  1.2  1  Amplitude  0.8  0.6  0.4  0.2  0 0  0.1  0.2  0.3  0.4  0.5  0.6  0.7  0.8  0.9  1  Time (sec)  FIGURE DP10.4 Step response for the high speed train system.  are P.O. = 4.44% DP10.5  Ts = 0.36 sec Kv = 14.1 .  The design specifications are Kv > 200; Ts < 12 ms and percent overshoot P.O. < 10%. The step response is shown in Figure DP10.5. A suitable compensator is Gc (s) = K  s + 403 , s + 2336  where K = 1.9476e + 13. Then, P.O. = 9.5%  Ts = 10 ms  Kv = 560 .  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  582  CHAPTER 10  The Design of Feedback Control Systems  1.2  1  Amplitude  0.8  0.6  0.4  0.2  0 0  2  4  6  8  10  12  14  16  18  20  Time (ms)  FIGURE DP10.5 Step response for the tape transport system.  A solution to the problem is the PI controller Gc (s) =  4.21s + 1.2 . s  The step response is shown in Figure DP10.6. 1.2  1  0.8  Amplitude  DP10.6  0.6  0.4  0.2  0 0  1  2  3 Time (sec)  FIGURE DP10.6 Step response for the engine control system.  4  5  6  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  583  Design Problems  The performance results are P.O. = 8.8%  and Ts = 2.14 .  The system is a type-1, so the steady-state error for a step input is zero, as desired. The jet aircraft roll angle motion is represented by the transfer function G(s) =  10 . (s + 10)(s2 + 2s + 20)  A good controls solution is obtained with a PID controller Gc (s) =  10s2 + 20s + 150 . s  The system is type-1, so the steady-state tracking error is zero for a step input. The performance results are P.O. = 9.13%  and  Ts = 1.56 .  Step Response 1.4  1.2  1 Amplitude  DP10.7  0.8  0.6  0.4  0.2  0  0  0.5  1  1.5 2 Time (sec)  FIGURE DP10.7 Step response for the jet aircraft roll control system.  2.5  3  3.5  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  584  CHAPTER 10  DP10.8  The Design of Feedback Control Systems  One good solution is obtained with the following PI controller 27.35(s + 2) . s  Gc (s) =  The system is type-1, so the steady-state tracking error is zero for a step input. The step response is shown in Figure DP10.8. Step Response From: U(1) 1.4  1.2  0.8 To: Y(1)  Amplitude  1  0.6  0.4  0.2  0  0  0.2  0.4  0.6  0.8  1  1.2  Time (sec.)  FIGURE DP10.8 Step response for the windmill radiometer.  DP10.9  Consider the PID controller Gc (s) = Kp + KD s +  KI 1.554s2 + 1.08s + 1 = s s  and the lead-lag controller Gc (s) = K    s+a s+b    s+c s+d    = 6.04  (s + 10)(s + 2) . (s + 1)(s + 5)  Both are stabilizing in the presence of a T = 0.1 second time delay. For the PID controller the phase margin is P.M. = 40o . For the lead-lag controller the phase margin is P.M. = 45o . We find (for these particular designs)  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  585  Design Problems  that the lead-lag controller is more able to remain stable in the process of increasing time delay. For a time-delay of T = 0.2 seconds, the lead-lag compensator has a phase margin of P.M. = 22o , while the PID controller is unstable. DP10.10  One solution is Gc (s) =  50(s + 0.01) . s+2  The Bode magnitude is shown in Figure DP10.10. You want high gain at  Bode Diagram 80  60 System: sys Frequency (rad/sec): 0.101 Magnitude (dB): 26.9  40  Magnitude (dB)  20  0  −20 System: sys Frequency (rad/sec): 10 Magnitude (dB): −26.9  −40  −60  −80  −100 −4 10  −3  10  −2  10  −1  10 Frequency (rad/sec)  0  10  1  10  2  10  FIGURE DP10.10 Step response for the windmill radiometer.  low frequency to improve disturbance rejection and decrease sensitivity to plant changes and low gain at high frequency to attenuate measurement noise. DP10.11  One solution is the PD controller Gc (s) = 0008(s + 10) . The step response is shown in Figure DP10.11. The closed-loop transfer  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  CHAPTER 10  The Design of Feedback Control Systems  function is T (s) =  s2  4 , + 3.4s + 4  where we use the prefilter Gp (s) =  4 . 0.36s + 3.6  Step Response 1.4 System: sys_cl Peak amplitude: 1.01 Overshoot (%): 0.637 At time (sec): 2.97  1.2  1 Amplitude  586  0.8  0.6  0.4  0.2  0  0  0.5  1  1.5  2  2.5 Time (sec)  3  FIGURE DP10.11 Step response for the polymerase chain reaction system.  3.5  4  4.5  5  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  587  Computer Problems  Computer Problems The m-file script and step response is shown in Figure CP10.1. The phase margin and percent overshoot are P.M. = 50o P.O. ≈ 18% , respectively.  nnumc=[110]; denc=[1 0]; sysc = tf(numc,denc); numg=[1]; deng=[1 10]; sysg = tf(numg,deng); syss = series(sysc,sysg); [Gm,Pm]=margin(syss); Pm % sys_cl = feedback(syss,1); [y,t]=step(sys_cl); step(sys_cl); grid S=stepinfo(y,t); PO=S.Overshoot  Pm = 49.9158 PO = 17.5724  Step Response 1.4  1.2  1 Amplitude  CP10.1  0.8  0.6  0.4  0.2  0  0  0.2  0.4  0.6 Time (sec)  0.8  FIGURE CP10.1 Phase margin and step response for the closed-loop system.  1  1.2  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  588  CHAPTER 10  CP10.2  The Design of Feedback Control Systems  Using a proportional controller the closed-loop characteristic equation is 1+K  s2  24.2 . + 8s + 24.2  A simple m-file script which computes the P.M. as a function of the gain K yields the proportional controller gain K = 6. Checking the phase margin of the system reveals that P.M. ≈ 40◦ , as desired. n=24.2; d=[1 8 24.2]; sys = tf(n,d); K=6; margin(K*sys), grid Bode Diagram Gm = Inf dB (at Inf rad/sec) , Pm = 39.9 deg (at 11.6 rad/sec)  Magnitude (dB)  20 0 −20 −40 −60  Phase (deg)  −80 0 −45 −90 −135 −180 −1 10  0  10  1  10 Frequency (rad/sec)  2  10  3  10  FIGURE CP10.2 Bode plot with a proportional controller K = 6 in the loop.  CP10.3  The uncompensated system is type-1. To realize a zero steady-state error to a ramp input we need to increase the system type by one. One controller that does this is the PI controller: Gc (s) =  KP s + KD . s  The step response is shown in Figure CP10.3 where it can be seen in the tracking error plot that the settling time is Ts < 5 seconds. The actual settling time is Ts = 3.6 seconds .  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  589  Computer Problems  KP=20; KD=10; nc=[KP KD]; dc=[1 0]; sysc = tf(nc,dc); n=1; d=[1 2 0]; sys = tf(n,d); sys_o = series(sysc,sys); sys_cl = feedback(sys_o,[1]); t=[0:0.001:10]; sys1 = tf([1],[1 0]); sys_cl1 = series(sys_cl,sys1); subplot(121) y=step(sys_cl1,t); plot(t,y,t,t,'--'), grid xlabel(' Time (sec)'), ylabel('Ramp response') e=y-t'; L=find(abs(e)>0.02); Ts=t(L(length(L))) subplot(122) plot(t,e,[0 10],[0.02 0.02],':',[0 10],[-0.02 -0.02],':') xlabel(' Time (sec)'), ylabel(' Track ing error') grid 10  0.1  9 0.05 8 0  6  Tracking error  Ramp response  7  5 4 3  -0.05  -0.1  -0.15  2 -0.2 1 0  0  5 Time (sec)  10  FIGURE CP10.3 Ramp response with a PI controller Gc (s) =  CP10.4  -0.25  0  20s+10 s  5 Time (sec)  10  in the loop.  From the percent overshoot spec we determine that P.O. < 10% implies ζ > 0.6. So, we target a phase margin P.M. = 100ζ = 60o . The m-file script which generates the uncompensated Bode plot is shown in Figure CP10.4a.  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  590  CHAPTER 10  The Design of Feedback Control Systems  numg = 100*conv([1 1],[1 0.01]); deng = conv([1 10],conv([1 2 2],[1 0.02 0.0101])); sysg = tf(numg,deng) w=logspace(-1,2,200); [mag,phase,w]=bode(sysg,w); [Gm,Pm,Wcg,Wcp]=margin(mag,phase,w); % Phi=60-Pm Pm Phi=(60-Pm)*pi/180; alpha=(1+sin(Phi))/(1-sin(Phi)) M=-10*log10(alpha)*ones(length(w),1); [mag,phase,w]=bode(sysg,w); for i = 1:length(w), magdB(i) = 20*log10(mag(1,1,i)); end semilogx(w,magdB,w,M), grid xlabel('Frequenc y (rad/sec)'), ylabel('mag [dB]') title('Uncompensated Bode Plot') hold on semilogx([.56072 5.6072 56.072 560.72],[20 0 -20 -40],'--')  È Phi = 56.2111 Pm = 3.7889 alpha = 10.8408  Uncompensated Bode Plot 60  40  mag [dB]  20  0  -20  -40  -60  -80 10-1  100  101  102  Frequency (rad/sec)  FIGURE CP10.4 (a) Uncompensated Bode plot.  We assume that K = 1 and raise the gain at a later step to meet settling time requirement. The uncompensated phase margin is P.M. = 3.7o , so that the lead compensator needs to add φ = 56.2o . The script also calculates α = 10.84. Following the design procedure outlined in Dorf & Bishop, we locate the compensator zero at ω = 2 rad/sec (see dashed line in Figure CP10.4a). Then, p = αz implies p = 21.68. After several iter-  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  591  Computer Problems  ations, we converge on K = 4 as a "good" value. The lead compensator is Gc (s) = 4  s+2 . s + 22  The step response is shown in Figure CP10.4b. The compensated Bode is shown in Figure CP10.4c.  K=4; numg = 100*conv([1 1],[1 0.01]); deng = conv([1 10],conv([1 2 2],[1 0.02 0.0101])); sysg = tf(numg,deng) numc=K*[1 2]; denc=[1 22]; sysc = tf(numc,denc); sys_o = series(sysc,sysg); sys_cl = feedback(sys_o,[1]); t=[0:0.01:5]; f=10*pi/180; [y,t,x]=step(f*sys_cl,t); plot(t,y*180/pi), grid xlabel(' Time (sec)') ylabel('Attitude rate (deg/sec)'), pause w=logspace(-1,2,200); [mag,phase,w]=bode(sys_o,w); [Gm,Pm,Wcg,Wcp]=margin(mag,phase,w); bode(sys_o) title(['Gain Margin = ',num2str(Gm),' Phase Margin = ',num2str(Pm)])  12  Attitude rate (deg/sec)  10  8  6  4  2  0 0  0.5  1  1.5  2  2.5 Time (sec)  FIGURE CP10.4 CONTINUED: (b) Step response.  3  3.5  4  4.5  5  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  592  CHAPTER 10  The Design of Feedback Control Systems Gain Margin = 14.96 Phase Margin = 60.49  Gain dB  50 0 -50 -100 10-3  10-2  10-1 100 Frequency (rad/sec)  101  102  10-2  10-1  101  102  Phase deg  100 0 -100 -200 -300 10-3  100  Frequency (rad/sec)  FIGURE CP10.4 CONTINUED: (c) Bode plot with lead compensator.  CP10.5  The closed-loop transfer function is θ(s)/θd (s) =  s2  K̄1 + K̄2 s + K̄2 s + K̄1  where K̄1 = K1 /J and K̄2 = K2 /J. A percent overshoot P.O. ≤ 20% requires ζ > 0.45. Select as the initial damping ζ = 0.7  (initial selection) .  For a second-order system with ζ = 0.7, we find that ω/ωn ≈ 0.9 when |θ(s)/θd (s)| = 0.7. So, we select ωn = ωB /0.9 as a starting choice. Therefore, since ωB = 10, we have ωn = 11 . The m-file script is shown in Figure CP10.5a. After several iterations, we find a set of "good" values for ζ = 0.8  and ωn = 4.5  (final selection) .  The step response and closed-loop Bode plot are shown in Figures CP10.5b and CP10.5c.  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  593  Computer Problems  % Par t (a) wn=4.5; zeta=0.8; K2=2*zeta*wn; K1=wn^2; % Par t (b) num=[K2 K1]; den=[1 0 0]; sys = tf(num,den); sys_cl = feedback(sys,[1]); f=10*pi/180; % set-up for 10 deg step input t=[0:.05:3]; [y,t,x]=step(f*sys_cl,t); plot(t,y*180/pi), xlabel(' time [sec]'), ylabel(' theta [deg]'), grid, pause % Par t (c) w=logspace(-1,2,400); [mag,phase,w]=bode(sys_cl,w); for i = 1:length(w), magdB(i) = 20*log10(mag(1,1,i)); end semilogx(w,magdB,[w(1) w(length(w))],[-3 -3]), grid xlabel('Frequenc y (rad/sec)') ylabel('Gain dB')  FIGURE CP10.5 (a) Script to generate the step response and the closed-loop Bode plot.  12  10  theta [deg]  8  6  4  2  0 0  0.5  1  1.5 time [sec]  FIGURE CP10.5 CONTINUED: (b) Step response.  2  2.5  3  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  594  CHAPTER 10  The Design of Feedback Control Systems  5  0  Gain dB  -5  -10  -15  -20  -25 10-1  100  101  102  Frequency (rad/sec)  FIGURE CP10.5 CONTINUED: (c) Closed-loop Bode plot.  CP10.6  The settling time and phase margin specifications require that the dominant closed-loop poles have natural frequency and damping of ζ ≥ 0.45 and ωn ≥ 1.78. The uncompensated roots locus is shown in Figure CP10.6a.  10 +  K=10  8 6  x  4  Imag Axis  numg=[1 10]; deng=[1 2 20]; sysg = tf(numg,deng); axis([-15,1,-10,10]); rlocus(sysg); hold on % zeta=0.45; wn=1.7778; x=[-10:0.1:-zeta*wn]; y=-(sqr t(1-zeta^2)/zeta)*x; xc=[-10:0.1:-zeta*wn]; c=sqr t(wn^2-xc.^2); plot(x,y,':',x,-y,':',xc,c,':',xc,- c,':') rlocfind(sysg), hold off  2 0  o  -2 -4  x  -6 -8 +  -10  -14  -12  -10  -8 Real Axis  FIGURE CP10.6 (a) Uncompensated root locus.  -6  -4  -2  0  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  595  Computer Problems  From the final value theorem, we determine that lim = sE(s) ≤ 0.1A  s→0  implies  A = 0.1A . 1 + GGc (s)  Therefore, the compensated Kpcomp ≥ 9. With the compensator Gc (s) = K  s+z s+p  we find that Kpcomp =  Kz Kpuncomp . p  But Kpuncomp = 0.5 and (from the uncompensated root locus) a gain of K = 10 results in roots of the characteristic equation in the desired region. Solving for z 1 Kpcomp = ≈2. p K Kpuncomp Select z = 0.5 to minimize changing the root locus. Then, p = 0.25, and the compensator is Gc (s) = 10  s + 0.5 . s + 0.25  The compensated root locus is shown in Figure CP10.6b and the step response is shown in Figure CP10.6c. The phase margin of the compensated  FIGURE CP10.6 CONTINUED: (b) Compensated root locus.  10 +  8 6 x  4  Imag Axis  numg=[1 10]; deng=[1 2 20]; sysg = tf(numg,deng); numc=[1 0.5]; denc=[1 0.25]; sysc = tf(numc,denc); sys_o = series(sysc,sysg); axis([-15,1,-10,10]); rlocus(sys_o); hold on % zeta=0.45; wn=1.7778; x=[-10:0.1:-zeta*wn]; y=-(sqr t(1-zeta^2)/zeta)*x; xc=[-10:0.1:-zeta*wn]; c=sqr t(wn^2-xc.^2); plot(x,y,':',x,-y,':',xc,c,':',xc,- c,':') rlocfind(sys_o) hold off  2 0  o  o+ x  -2 -4  x  -6 -8 +  -10  -14  -12  -10  -8 Real Axis  -6  -4  -2  0  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  596  CHAPTER 10  The Design of Feedback Control Systems  system is P.M. = 62.3o and the settling time Ts < 5 seconds. numg=[1 10]; deng=[1 2 20]; sysg = tf(numg,deng); numgc=10*[1 0.5]; dengc=[1 0.25]; sysc = tf(numgc,dengc); sys_o = series(sysc,sysg); sys_cl = feedback(sys_o,[1]); t=[0:0.1:5]; step(sys_cl,t) [mag,phase,w]=bode(sys_o); [gm,pm,w1,w2]=margin(mag,phase,w); pm  >> pm = 62.3201  1.2  1  Amplitude  0.8  0.6  0.4  0.2  0 0  0.5  1  1.5  2  2.5  3  3.5  4  4.5  5  Time (secs)  FIGURE CP10.6 CONTINUED: (c) Step response and phase margin verification.  CP10.7  Both design specifications can be satisfied with an integral controller Gc (s) = K1 +  K2 10 = . s s  The simulation results and m-file script are shown in Figures CP10.7a and b.  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  597  Computer Problems Unit Step Response  Phi dot  1.5  1  0.5 0 0  0.1  0.2  0.3  0.4  0.5  0.6  0.7  0.8  0.9  1  0.7  0.8  0.9  1  Time (sec) Unit Ramp Response  Tracking error  0  -0.05  -0.1 -0.15  0  0.1  0.2  0.3  0.4  0.5  0.6  Time (sec)  FIGURE CP10.7 (a) Simulation results.  K1=0; K2=10; numc=[K1 K2]; denc=[1 0]; sysc = tf(numc,denc); numg=[23]; deng=[1 23]; sysg = tf(numg,deng); sys_o = series(sysc,sysg); sys_cl = feedback(sys_o,[1]); t=[0:0.01:1]; ys=step(sys_cl,t); subplot(211) plot(t,ys), xlabel(' Time (sec)'), ylabel('Phi dot') title('Unit Step Response'), grid u=t; yr=lsim(sys_cl,u,t); subplot(212) plot(t,yr-u','--') xlabel(' Time (sec)'), ylabel(' Track ing error') title('Unit Ramp Response'), grid  FIGURE CP10.7 CONTINUED: (b) M-file design script.  CP10.8  From Example 10.3, we have that the loop transfer function is Gc (s)G(s) =  8.1(s + z) , s2 (s + 3.6)  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  598  CHAPTER 10  The Design of Feedback Control Systems  where z = 1. We want to determine a value of z so the the percent overshoot is reduced from 46% to less than 32%. A valid design is Gc (s)G(s) =  8.1(s + 0.45) . s2 (s + 3.6)  The m-file script and step response are shown in Figure CP10.8. The percent overshoot is P.O.=27.7 %. Step Response From: U(1) 1.4  1.2  1  0.8 To: Y(1)  Amplitude  K1 = 8.1; numg = [K1]; deng = [1 0 0]; sysg = tf(numg,deng); numc = [1 0.45]; denc = [1 3.6]; sysc = tf(numc,denc); sys_o = series(sysc,sysg); sys_cl = feedback(sys_o,[1]); step(sys_cl) y=step(sys_cl); po=100*(max(y)-1)  0.6  0.4  0.2  0  0  1.6  3.2  4.8  6.4  8  Time (sec.)  FIGURE CP10.8 Response of system with new lead compensator design.  CP10.9  From AP10.10, we have the transfer function is Vo (s) Vi (s) 1 + R2 C 2 s = . 1 + R1 C 1 s  T (s) =  Substituting C1 = 0.1 µF ,C2 = 1 mF , R1 = 10 kΩ, and R2 = 10 Ω yields T (s) =  1 + 0.01s . 1 + 0.001s  The frequency response is shown in Figure CP10.9.  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  599  Computer Problems Bode Diagrams  20  15  c1=0.0000001; c2=0.001; r1=10000; r2=10; n=[c2*r2 1]; d=[c1*r1 1]; sys=tf(n,d) bode(sys)  Phase (deg); Magnitude (dB)  10  5  0 60 50 40 30 20 10 0 1 10  10  2  10  3  10  4  Frequency (rad/sec)  FIGURE CP10.9 Op-amp circuit frequency response.  CP10.10  The plot of K versus phase margin is shown in Figure CP10.10. The value of K that maximizes the phase margin is K = 4.15.  60  FIGURE CP10.10 Plot of K versus phase margin.  55  50  45  40  P.M.  K=[0.1:0.01:10]; T=0.2; [np,dp]=pade(T,6); sysp=tf(np,dp); for i=1:length(K) ng=K(i)*[1 0.2]; dg=[1 6 0 0]; sysg=tf(ng,dg); [gm,pm]=margin(sysg*sysp); PM(i)=pm; end plot(K,PM), grid [P,n]=max(PM); K(n) xlabel('K'), ylabel('P.M.')  35  30  25  20  15  0  1  2  3  4  5 K  6  7  8  9  10  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  C H A P T E R  1 1  The Design of State Variable Feedback Systems  Exercises E11.1  The system is given by ẋ = Ax + Bu u = Kx where   A=  0  1  −1 0       B=  1 0 0 1     and    K=  −k  0  0 −2k     .  Then, with u = Kx, we have   ẋ =   −k  1  −1 −2k    x .  The characteristic equation is   det[sI − A] = det   s+k  −1  1  s + 2k     = s2 + 3ks + 2k 2 + 1  = s2 + 2ζωn s + ωn2 = 0 .  Solving for k where ωn2 = 2k 2 + 1 and ζ = 1 (critical damping) yields k = 2. 600  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  601  Exercises  E11.2  Let u = −k1 x1 − k2 x2 + r . Then,   ẋ =   0    1    x+  9 − k1 −k2  0 1    r ,  and det(sI − A) = s2 + k2 s + k1 − 9 = 0 . We want ζ = 1, so the desired characteristic equation is pd (s) = (s + co )2 , where co is to be determined to meet Ts = 4 and where k2 = 2co and k1 = c2o + 9. Solving for the state response of x1 (t) to a unit step input we find x1 (t) = 1 − e−co t − co te−co t . When t ≥ Ts = 4 sec we want x1 (t) ≥ 0.98. Solving for co at t = Ts yields co = 1.459, E11.3  k1 = 11.13,  and  k2 = 2.92 .  The controllability matrix is Pc =  h  B AB  i    =  0  1  1 −3     ,  and det Pc 6= 0, therefore the system is controllable. The observability matrix is   Po =   C CA      =  0  2  0 −6     ,  and det Po = 0; therefore the system is unobservable. E11.4  The controllability matrix is Pc =  h  B AB  i    =  0  0  2 −4     ,  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  602  CHAPTER 11  The Design of State Variable Feedback Systems  and the det Pc = 0; therefore the system is uncontrollable. The observability matrix is   C  Po =   CA      =    1 0   ,  −10 0  and det Po = 0; therefore the system is also unobservable. E11.5  The controllability matrix is Pc =  h  B AB  i    =    1 −2 −2   ,  3  and det Pc = −1 6= 0; therefore the system is controllable. The observability matrix is   Po =   C CA      =  1 0 0 1     ,  and det Po = 1 6= 0; therefore the system is observable. E11.6  The controllability matrix is Pc =  h  B AB  i    =  0  1  1 −2     ,  and det Pc 6= 0; therefore the system is controllable. The observability matrix is   Po =   C CA      =  1 0 0 1     ,  and det Po 6= 0; therefore the system is observable.  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  603  Exercises  E11.7  The block diagram is shown in Fig. E11.7.  2 U(s)  12  1 s  1 s  + - -  -  +  2  Y(s)  5 3 FIGURE E11.7 The block diagram for E11.7.  E11.8  The block diagram is shown in Fig. E11.8.  10 8  U(s)  4  1 s  + - - -  +  -  1 s  1 3 9 FIGURE E11.8 The block diagram for E11.8.  1 s  2  +  ++ Y(s)  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  604  CHAPTER 11  E11.9  The Design of State Variable Feedback Systems  The controllability matrix is Pc =  h  B AB  i    =  k1    k1 − k2   ,  k2 −k1 + k2  and det Pc = −k12 + k22 . So, the condition for complete controllability is k12 6= k22 . E11.10  A matrix differential equation representation is     ẋ =     0  1  0  0  −10 −6 −3  y = [−3 4 E11.11            0   0       1  x +  0 u 1    2]x + [0]u .  The system is given by ẋ = Ax + Bu y = Cx + Du where     A=              0   0  h i       , B = , C = 1  1 2 0 , and D = [1] .  0   0 1 0 0  −2 0 −7  The controllability matrix is  Pc =  h  1    2  B AB A B  i     0  =  0     0  1   1 −7   ,  1 −7  49    and det Pc = −1 6= 0; therefore the system is controllable. The observability matrix is      C     Po =   CA  =    CA2         1 2 0 1    0   2   ,  −4 0 −13    © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  605  Exercises  and det Po = −29 6= 0; therefore the system is observable. The transfer function is G(s) =  s2  6 . + 5s + 6  The response of the system to a unit step is y(t) = 1 − 3e−2t + 2e−3t . The step response is shown in Figure E11.12  1 0.9 0.8 0.7 Step Response  E11.12  0.6 0.5 0.4 0.3 0.2 0.1 0  0  FIGURE E11.12 Unit step response.  0.5  1  1.5  2  2.5 Time (s)  3  3.5  4  4.5  5  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  606  CHAPTER 11  The Design of State Variable Feedback Systems  Problems P11.1  Consider the system ẋ = x + u u = −kx . So, ẋ = x − kx = (1 − k)x and x(t) = e(1−k)t x(0) . The system is stable if k > 1. Computing the value of J (assuming k > 1) yields J=  Z  ∞  e2(1−k)t x2 (0)dt =  0  1 . k−1  Thus, J is minimum when k → ∞. This is not physically realizable. Select k = 35. Then, the value of the performance index J is J=  1 . 34  The system is not stable without feedback. P11.2  (a) The performance index is given J=  ∞  Z  (x2 + λu2 )dt .  0  The system is ẋ = x + u u = −kx . So, J=  Z  0  ∞  2  2 2  (x + λk x )dt =  Z  ∞  2  2  2  (1 + λk )x dt = (1 + λk )  0  Z  ∞  x2 dt .  0  Carrying out the integration (assuming k > 1) yields J = (1 + λk 2 )  1 . k−1  We want to determine k that minimizes J. Taking the partial of J  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  607  Problems  with respect to k and setting the result to zero yields ∂J λk 2 − 2λk − 1 = =0, ∂k (k − 1)2 or λk 2 − 2λk − 1 = 0 . Solving for k yields k =1+  r  1 , λ  1+  where we reject the solution k = 1 −  q  1 + λ1 , since we require k > 1.  (b) For λ = 2, we determine that k = 2.2 and Jmin = 8.9. P11.3  The system is given by   ẋ =   1 0 −1 2      x +   1 1    u  u = −k(x1 + x2 ) = −k[1 1]x .  Then, with feedback applied, the system is   ẋ =   (1 − k)  −k  −(1 + k) (2 − k)    x .  Solving HT P + PH = −I yields 2p11 (1 − k) − 2p12 (k + 1) = −1 p12 (3 − 2k) − p11 k − p22 (k + 1) = 0 −2kp12 + 2p22 (2 − k) = −1 . Solving for p11 , p12 and p22 yields −(2k 2 − 6k + 7) 4(4k 2 − 8k + 3) 2k 2 − 2k − 1 = 4(4k 2 − 8k + 3)  p11 = p12  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  608  CHAPTER 11  The Design of State Variable Feedback Systems  p22 =  −(2k 2 − 6k + 3) . 4(4k 2 − 8k + 3)  The performance index is computed to be J = xT (0)Px(0) = p11 + 2p12 + p22 =  1 , 2k − 1  when x(0) = [1 1]T . So as k → ∞, J → 0. The system is unstable without feedback. The performance index is J = xT (0)Px(0) = p11 − 2p12 + p22 . From Example 11.12 in Dorf and Bishop, we determine that J=  2k 2 + 1 . 2k 2  So, when k → ∞, the performance index J → 1. The plot of J versus k is shown in Figure P11.4.  60  50  40  J  P11.4  30  20  10  0  0  1  2  3  FIGURE P11.4 The performance index J versus k.  4  5 K  6  7  8  9  10  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  609  Problems  P11.5  The system is given by   ẋ =   0 1 0 0  u = −kx .      x +   0 0 1 1    u  The performance index is J=  Z  ∞  T  T  (x x + u u)dt =  0  Z  ∞  (1 + k 2 )(xT x)dt .  0  First, we solve HT P + PH = −(1 + k 2 )I , yielding, (1 + k 2 ) 2k 3 k + k2 + k + 1 = 2k 2 3 2k + k 2 + 2k + 1 = . 2k  p12 = p22 p11  The performance index is then given by J = p11 + 2p12 + p22 =  2k 4 + 4k 3 + 3k 2 + 4k + 1 . 2k 2  Taking the partial of J with respect to k, setting the result to zero and solving for k yields ∂J 2k 4 + 2k 3 − 2k − 1 = =0 ∂k k3 or 2k 4 + 2k 3 − 2k − 1 = 0 . Solving for k yields k = 0.90. The plot of J versus k is shown in Figure P11.5. The value of the performance index is J = 6.95 when k = 0.90.  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  610  CHAPTER 11  The Design of State Variable Feedback Systems  8.8 8.6 8.4 8.2  J  8 7.8 7.6 7.4 7.2 7 6.8 0.5  0.6  0.7  0.8  0.9  1  1.1  1.2  1.3  1.4  k  FIGURE P11.5 The performance index J versus k.  P11.6  (a) For P11.3, we have J=  1 . 2k − 1  So, as k → ∞, then J → 0. But k = ∞ is not a practical solution, so select k = 10. Then, J = 1/19, and   ẋ =   −9 −10 −11  −8     x = Ax .  The closed-loop system roots are determined by solving det[sI − A] = s2 + 17s − 38 = 0 , which yields s = −19 and s2 = 2. The system is unstable. The original system was unstable, and it remains unstable with the feedback. In general,   ẋ =   (1 − k)  −k  −(1 + k) (2 − k)     x = Ax  and det[sI − A] = s2 + s(2k − 3) + (2 − 4k) = 0. A Routh-Hurwitz analysis reveals that the system is unstable for all k.  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  611  Problems  (b) For P11.4, we have   ẋ =   0  1  −k −k     x = Ax ,  and det[sI − A] = s2 + ks + k = 0 . The performance index was found to be J =1+  4k + 1 . 2k 2  As k → ∞, we have J → 0. But k = ∞ is not a practical choice for k. Select k = 10. Then, det[sI − A] = s2 + 10s + 10 = (s + 1.13)(s + 8.87) . The closed-loop system is stable. (c) In P11.5, we found that k = 0.90 for Jmin . We are given   ẋ =   0  1  −k −k    x  and det[sI − A] = s2 + ks + k = s2 + 0.9s + 0.9 = (s + 0.45 + j0.835)(s + 0.45 − j0.835) . P11.7  The closed-loop system is   ẋ =   0  1  −k1 −k2     x = Hx ,  and det[sI − H] = s2 + k2 s + k1 = s2 + 2ζωn s + ωn2 = 0 . We desire ωn = 2, so set k1 = 4. With xT (0) = [1, 0], we have J = p11 , and solving HT P + PH = −I  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  612  CHAPTER 11  The Design of State Variable Feedback Systems  yields    0  −4  1 −k2       p11 p12    p11 p12    −1  0  0  −1  +  p12 p22  p12 p22  =  and p11 =     0  1  −4 −k2        ,  20 k 2 + 20 k2 + = 2 . 8 8k2 8k2  Select k2 = for Jmin , where Jmin =  √  5 2 .  √  20  Then  det[sI − H] = s2 +  √  20s + 4 = 0 ,  and ωn = 2 and ζ = 1.12. The system is overdamped. P11.8  From Example 11.11 in Dorf and Bishop, we have   So,  P=  k22 +2 2k2 1 2  1 2 1 k2  J = xT (0)Px(0) =     .  k22 + 2 2k2  when xT (0) = [1 0]. Taking the partial of J with respect to k2 and setting the result to zero yields ∂J k2 + 2 =1− 2 2 =0 . ∂k2 2k2 Solving for the optimum value of k2 yields √ k2 = 2 . P11.9  Let x1 = φ and x2 = ω. We have that ω=  dφ . dt  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  613  Problems  The state equations are x˙1 = x2 x˙2 = Ku . Select a feedback such that u = −x1 − K1 x2 + r when r(t) is the reference input. Then,   ẋ =   and  0  1  −K −KK1      x+  0 K    r ,  det[sI − A] = s2 + K1 Ks + K . so that the overshoot is 4%. Since Ts = 1 = ζω4n , we √ require ζωn = 4 or ωn = 4 2. Then, s2 + 8s + 32 = s2 + K1 Ks + K, or 8 K = 32 and K1 = 32 = 41 . We desire ζ =  P11.10  √1 , 2  The system with feedback is given by   ẋ = Ax =   −10 −25 1  0    x ,  where x1 (0) = 1, and x2 (0) = −1. The characteristic equation is   det[sI − A] = det   s + 10 25 −1  s     = s(s + 10) + 25 = s2 + 10s + 25 = 0 .  The roots are s1,2 = −5. The solution is   x(t) =   φ11 φ12 φ21 φ22       x(0) =   φ11 − φ12 φ21 − φ22     since x1 (0) = 1 and x2 (0) = −1. We compute the elements of the state transition matrix as follows: φ22 (t) = (1 + 5t)e−5t  and  φ21 (t) = te−5t ,  therefore x2 (t) = −(1 + 4t)e−5t .  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  614  CHAPTER 11  The Design of State Variable Feedback Systems  Similarly, φ11 (t) = (1 − 5t)e−5t  φ12 = −25e−5t .  and  Therefore, x1 (t) = (1 + 20t)e−5t . P11.11  Let u = −k1 x1 − k2 x2 + αr where r(t) is the command input. A state variable representation of the plant is   −5 −2  h  0 1  ẋ =  y=  2  0 i      x+  x+  h  0  i  0.5 0  u.    u  The closed-loop transfer function is T (s) =  α . s2 + (k1 /2 + 5)s + 4 + k2  To meet the performance specifications we need ωn = 4.8 and ζ = 0.826. Therefore, the desired characteristic polynomial is q(s) = s2 + 2(0.826)4.8s + 23 = s2 + 8s + 23 . Equating coefficients and solving for k1 and k2 yields k2 = 19 and k1 = 6. Select α = 23 to obtain zero steady-state error to a step input. P11.12  A state variable representation of the dc motor is        ẋ =        −3 −2 −0.75 0 0      1             0   0       0  x +  0 u       0   0      3  0  0 0  0  2  0 0  0  0  1 0  0  0  0 2 0  y = [0 0 0 0 2.75]x .  0  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  615  Problems  The controllability matrix is   1 −3     0   Pc =   0    0   0  3  4.5 −18  3 −9    13.5    18     −18    9  0  6 −18  0  0  6  0  0  0    12  and the det Pc 6= 0, so the system is controllable. The observability matrix is   0     0   Po =   0    0   33  0  0  0  2.75  0  0  5.5  0  0  5.5  0  0  11  0  0  0  0  0  0  0  and the det Po 6= 0, so the system is observable. P11.13          ,       To meet the Kv = 35 specification, we need K = 2450. A state variable representation is   ẋ =   0  1  0 −70      x +   0 2450    u  y = [1 0]x . Let u = −k1 x1 − k2 x2 . Then, the closed-loop characteristic equation is q(s) = s2 + (2450k2 + 70)s + 2450k1 = 0 . The desired characteristic polynomial is s2 + 72.73s + 2644.63 = 0 where we select ζ = 0.707 and ωn = 51.42 to meet the performance specifications. Equating coefficients and solving for the gains yields k1 =  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  616  CHAPTER 11  The Design of State Variable Feedback Systems  1.08 and k2 = 0.0011. P11.14  Let u = −k1 x1 − k2 x2 − k3 r where r(t) is the command input. Then, the closed-loop system in state variable form is   ẋ =   −10 − k1 −k2 1  0      x+  1 0    r  y = [0 1]x . To meet the performance specifications, we want the closed-loop characteristic polynomial to be q(s) = s2 + 8s + 45.96 = 0 where ζ = 0.59 and ωn = 6.78. The actual characteristic polynomial is det(sI − A) = s2 + (10 + k1 )s + k2 = 0 . Equating coefficients and solving for the gains yields k2 = 45.96 and k1 = −2. Select k3 = k2 = 45.96 to obtain a zero steady-state error to a step input. This results in a settling time of Ts = 0.87 s and a percent overshoot of P.O. = 10%. P11.15  The transfer function is G(s) = C(sI − A)−1 B =  1 . s+1  The system is not controllable and not observable. P11.16  Let u = −Kx . Then, Ackermann's formula is K = [0, 0, ..., 1]P −1 c q(A) where q(s) is the desired characteristic polynomial, which in this case is q(s) = s2 + 2s + 10 .  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  617  Problems  A state-space representation of the limb motion dynamics is   −4  ẋ =   0  1 −1      x +     1  u .  0  The controllability matrix is   Pc = [B AB] =   1 −4 0  1     and    P−1 c =  1 4 0 1     .  Also, we have   q(A) = A2 + 2A + 10I =   18 0 −3 9     .  Using Ackermann's formula, we have K = [−3  9] .  P11.17  The system is either uncontrollable or unobservable if a = 5 or a = 8. Both of these values correspond to system real poles. So, if a takes on either value, a pole-zero cancellation occurs in the transfer function.  P11.18  A matrix differential equation representation is   ẋ =   0  1  −1 −2 y = [1      x+  0 1    u  0]x .  Let u(t) = −k1 x1 − k2 x2 . Then, the closed-loop characteristic equation is q(s) = s2 + (2 + k2 )s + 1 + k1 = 0 . We desire the characteristic equation √ s2 + 2 2s + 2 = 0 .  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  618  CHAPTER 11  The Design of State Variable Feedback Systems  Equating coefficients and solving for the gains yields k1 = 1 and k2 = √ 2 2 − 2 = 0.828. P11.19  A state space representation is   ẋ =   0    1    0  x+  3 −2 y = [3  1]x .    1  1    r  The controllability matrix is 0  Pc =      ,  1 −2  and det Pc 6= 0, so the system is controllable. The observability matrix is   Po =   3 1 3 1     ,  and the det Po = 0, so the system is not observable. P11.20  The characteristic equation associated with A is s2 (s2 + 0.2s + 0.0015) = 0 . There are two roots at the origin, so the system is unstable. The system can be stabilized with δ = −k1 x1 − k3 x3 = 20x1 − 10x3 .  P11.21  (a) Let x1 = i1 , x2 = i2 and u = v. Then, the state equation is   ẋ =   −(R1 +R3 ) L1 R3 L2  R3 L1 −(R3 +R2 ) L2    x +   Also, y = vo , but y = [R3    − R3 ]x .  1 L1  0    u .  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  619  Problems  (b) The observability matrix is   Po =   C CA      =  R3   3 − RL1 R − R32 1  and det Po =    1 L1  +  1 L2    R2 R3 L2  −R3  + R32    1 L1  +  1 L2  R2 R1 − R32 . L2 L1   So, when R2 R1 = , L1 L2 det Po = 0 and the system is not observable. (c) Let a=  R1 + R3 , L1  b=  R3 + R2 . L2  and  Then   det[sI − A] = det  "  (s + a) 3 −R L2  3 −R L1  (s + b)     #  R32 = (s + a)(s + b) + = (s + r)2 L1 L2 = s2 + (a + b)s + ab +  R32 . L1 L2  The system has two equal roots when R32 (a + b) − 4 ab + L1 L2 2  !  or   R1 + R3 R3 + R2 + L1 L2  2  −4  (R1 + R3 )(R3 + R2 ) + R32 =0. L1 L2       © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  620  CHAPTER 11  (a) Without state feedback the state differential equation is given by   ẋ =   y=  h  −0.4 −1 1 0 1  0 i  x.      x+  1 0    u  The step response is shown in Figure P11.22a. (a) Without state feedback  2  x2  1.5 1 0.5 0 0  2  4  6  8  10  12  14  16  18  20  1.4  1.6  1.8  2  Time (sec) (b) With state feedback  1.5  1  x2  P11.22  The Design of State Variable Feedback Systems  0.5 0 0  0.2  0.4  0.6  0.8  1  1.2  Time (sec)  FIGURE P11.22 Step response (a) without state feedback, and (b) with state feedback.  (b) Consider state feedback u = −K(ax2 + bx1 ) + cr where r is the reference input and K, a, b and c are to be determined. Then, the state differential equation is   −0.4 − Kb −1 − Ka  h  0 1  ẋ =   y=  1 i  0 x,      x+  c 0    r  and det(sI − A) = s2 + (0.4 + Kb)s + (1 + Ka) = 0. Our specifications 4 are P.O. = 5% and Ts = 1.35 sec. So, ζ = 0.69 and ωn = ζ1.35 = 4.3.  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  621  Problems  Solving for K, a and b yields Ka = ωn2 − 1 and Kb = 2ζωn − 0.4 . Select K = 1. Then, a = 17.49 and b = 5.53. Select c = 1 + Ka to achieve a zero steady-state tracking error. (c) The step response is shown in Figure P11.22b for the system with state feedback. P11.23  Using the internal model design method for step inputs, we have          0 1 0   0   e  ė   =  0 0 1  + 0      z  ż 0 0 0 1     where we choose      w ,    w = −K1 e − K2 z . To place the poles at s = −10 and s = −2±j we use Ackermann's formula to compute K1 = 50 K2 = [45  14] .  The compensator has the form shown in Figure 11.14 in Dorf and Bishop. P11.24  Using the internal model design method for ramp inputs, we have       0 1 0 0   e  ė      0 0 1 0     ë  =       ė    0 0 0 1     z ż 0 0 0 0   where we choose           0     0       + w     0      1  w = −K1 e − K2 ė − K3 z . To place the poles at s = −20 and s = −2 ± 2j we can use Ackermann's formula. We also need an additional pole (must be a stable pole); select  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  622  CHAPTER 11  The Design of State Variable Feedback Systems  s = −20 as the fourth pole. Then, K1 = 3200 K2 = 1920 K3 = [568 44] . The compensator has the form shown in Figure 11.16 in Dorf and Bishop. P11.25  The observability matrix is   Po =   C CA      =  1  −4  21 −36     ,  and det Po = 48 6= 0; therefore the system is completely observable. The desired poles of the observer are s1,2 = −1. This implies that the desired characteristic polynomial is pd (s) = s2 + 2s + 1 . The actual characteristic polynomial is det |λI − (A − LC)| = det  λ − 1 + L1  −4 − 4L1  5 + L2  λ − 10 − 4L2  = λ2 + (L1 − 4L2 − 11)λ + 10L1 + 8L2 + 30 = 0 . Solving for L1 and L2 yields   L=  L1 L2      =  −0.25 −3.3125     .  Checking we find that det(λI − (A − LC)) = s2 + 2s + 1. The response of the estimation error is shown in Figure P11.25, where e(0) = [ 1 1 ]T .  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  623  Problems  Response to Initial Conditions 2.5 To: Out(1)  2 1.5 1  Amplitude  0.5 0 1 To: Out(2)  0.5 0 -0.5 ?-1 -1.5  0  1  2  3  4  5  6  Time (sec )  FIGURE P11.25 Estimation error response to an initial condition.  P11.26  The observability matrix is        C   2      Po =   CA  =  0     CA  2  −4 2  32  20    0   −4   .  14    The det Po = 728 6= 0, hence the system is observable. The gain matrix      0.14     L=  −0.93     0.79  results in the observer poles at s1,2 = −1 ± j and s3 = −5, as desired. P11.27  The observability matrix is   Po =   C CA      =  1  0  1 0     .  The det Po = 0, hence the system is not completely observable. So, we cannot find an observer gain matrix that places the observer poles at the desired locations.  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  624  CHAPTER 11  P11.28  The Design of State Variable Feedback Systems  Selecting K = 16 yields a zero steady-state error to a unit step input. The step response is shown in Figure P11.28.  Step Response 1 0.9 0.8  Amplitude  0.7 0.6 0.5 0.4 0.3 0.2 0.1 0  0  0.5  1  1.5 Time (sec)  FIGURE P11.28 Estimation error response to an initial condition.  P11.29  The system transfer function is Y (s) =  2 U (s) . s+3  The associated state variable model is ẋ = −3x + 2u y=x.  2  2.5  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  625  Advanced Problems  Advanced Problems AP11.1  The closed loop system in state-space form is given by              ẋ1       ẋ  =   2    ẋ3  0  1  0  0  −1  2    −2KK1 −2KK2 −4 − 2KK3               x1   0       x  +  0 u  2      x3  2K    x1  h i    y= 1 0 0   x2  .   The closed-loop transfer function is T (s) =  s3  + (2KK3 +  5)s2  x3    4K . + (4KK2 + 2KK3 + 4)s + 4KK1  Setting the steady-state error to zero, we determine that ess = 1 − T (0) = 1 −  1 . K1  Solving for K1 yields K1 = 0.5 . Choosing K2 = 0.5  and K3 = 1.5  results in a percent overshoot of P.O. = 2.82%. AP11.2  A state variable representation is given by ẋ = Ax + Bu where      −3 −1 −1    A= 0 0   4  ,    0  1  0       3     B=  0  .    0    © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  626  CHAPTER 11  The Design of State Variable Feedback Systems  Let u = −Kx . Then, with K=  h  4.00 24.33 39.67  i  ,    b1  the closed-loop system poles are s = −4, −5, and −6. AP11.3  Given   A=  0  1  −1 −2    ,  and B =   b2     ,  we compute the determinant of the controllability matrix as det Pc = det[B AB] = − (b1 + b2 ) . The system is controllable if and only if the determinant is non-zero. So, for the system to be controllable, we require that b2 6= −b1 . AP11.4  Consider the state variable feedback law u = −Kx . Using Ackermann's formula, we determine that K = [−14.2045 − 17.0455 − 94.0045 − 31.0455] results in the closed-loop system characteristic roots at s = −2±j, s = −5 and s = −5.  AP11.5  The closed-loop transfer function for the system is T (s) =  s3  + (9 + 2K3  )s2  2Kp . + (26 + 2K2 + 10K3 )s + (26 + 6K2 + 12K3 )  Setting the steady-state error for a step input to zero yields ess = 1 −  2Kp =0. 26 + 6K2 + 12K3  Solving for Kp in terms of K2 and K3 yields Kp = 13 + 3K2 + 12K3 . Now, choosing K2 = 5  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  627  Advanced Problems  K3 = 2 results in the closed-loop characteristic roots at s1 = −4 s2 = −4 s3 = −5 . Also, the prefilter gain is Kp = 52 . AP11.6  (a) A state variable representation is given by   A= C=  h  0  1  −1 −2 1 0  i  .     ,    B=  0 1     ,  Since the determinant of the controllability matrix det[B AB] 6= 0, the system is controllable. (b) The state variable representation is ẋ = Ax + Bu , or    ẋ1 ẋ2      =  0  1  −1 −2     x1 x2      +  The determinant of the controllability matrix  1 −1    u .  det Pc = det[B AB] = 0 . Therefore, the system is uncontrollable. AP11.7  The closed-loop transfer function is T (s) =  s3  + (10 + 60K3  )s2  120 . + (16 + 120(K3 + K2 ))s + 120  The state feedback gains K2 = 0.283  and  K3 = 0.15  place the poles at the desired locations. The plot of the roll output for a unit step disturbance is shown in Figure AP11.7.  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  628  CHAPTER 11  The Design of State Variable Feedback Systems  0.35  0.3  Amplitude  0.25  0.2  0.15  0.1  0.05  0 0  0.5  1  1.5  2  2.5  3  3.5  4  Time (secs)  FIGURE AP11.7 Roll angle response to a step disturbance.  AP11.8  The state equations are (using the parameters of P3.36 in Dorf and Bishop) 8 1 [80θ − 50h] = −x1 + x2 50 5 θ̇ = ẋ2 = ω = x3 Km Km Kb Km Ka 353 25000 ω̇ = ẋ3 = ia = − ω+ vi = − x3 + vi . J JRa JRa 30 3 ḣ = ẋ1 =  In state variable form we have (without feedback)   8  −1 5  ẋ =   0 0    0 1  0 − 353 30  0          x +         0 0 25000 3     vi .    (a) In this case we have vi = −kh + ar = −kx1 + ar, where k and a are the parameters to be determined and r is the reference input. With the feedback of h(t) we have     ẋ =     −1  8 5  0  0  0  1  353 − 25000 3 k 0 − 30          x+      0 0 a 25000 3      r .    © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  629  Advanced Problems  Since we only have one parameter to adjust, namely k, we will probably not be able to simultaneously meet both design specifications, In fact with k = 0.00056 we obtain the percent overshoot P.O. = 9.89%. The settling time criterion cannot simultaneously be met—the best that can be obtained is Ts ≈ 7.5 seconds. In this case, we choose a = 0.00056 to make the steady-state value of h(t) = 1. (b) In this case we have vi = −k1 h − k2 θ + ar = −k1 x1 − k2 x2 + ar, where k1 , k2 , and a are the parameters to be determined and r is the reference input. Since we have two parameter to adjust, namely k1 and k2 we will probably be able to simultaneously meet both design specifications. In fact with k1 = 0.00056  and k2 = 0.001  we obtain the percent overshoot P.O. = 4.35%. The settling time criterion is easily met— Ts ≈ 5 seconds. In this case, we choose a = 0.0012 to make the steady-state value of h(t) = 1. AP11.9  (a) The state vector differential equation is    0 1   −2 0  ẋ =   0 0         0     0    u , x +     1   0      0 0   1 0   0    1 0 −1 0  1  where x1 = z, x2 = ż, x3 = y and x4 = ẏ. (b) The characteristic equation is  s4 + 3s2 + 1 = (s + j0.618)(s − j0.618)(s + j1.618)(s − j1.618) = 0 . So, the system is oscillatory. (c) Let u = −kx4 . Then characteristic equation is s4 + ks3 + 3s2 + 2ks + 1 = 0 which is stable if k > 0. (d) Rewrite the characteristic equation as 1+  ks(s2 + 2) =0. s4 + 3s2 + 1  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  630  CHAPTER 11  The Design of State Variable Feedback Systems  The root locus is shown in Figure AP11.9. A reasonable solution for k is k = 1.35.  3  2 x o  1  Imag Axis  x  0  o  x  -1 o x  -2  -3 -3  -2  -1  0  1  2  3  Real Axis  FIGURE AP11.9 s(s2 +2) Root locus for 1 + k s4 +3s2 +1 = 0.  AP11.10  The state differential equation is ÿ = ky + αu where k and α depend on the system parameters, such as mass and length. The transfer function is y α = 2 u s −k which is unstable at the top of the arc. Since we can only use ẏ for feedback, we have ẏ sα = 2 . u s −k Let Gc (s) =  K1 s + K2 . s  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  631  Advanced Problems  Then GGc (s) =  α(K1 s + K2 ) (s2 − k)  and the closed-loop characteristic equation is αK1 s + αK2 + s2 − k = 0 or s2 + αK1 s + αK2 − k = 0 . Select αK2 − k > 0 and αK1 > 0 for stability. AP11.11  The state-space representation of the plant is ẋ = Ax + Bu y = Cx where   A=  0  1  −2 −3     ,    B=  0 1     , and  C=  h      1 0  i  .  With the intermediate variables defined as z = ẋ  and w = u̇  we have  where    1 0  0  ė  = 0 0 1   ż 0 −2 −3        0      e  + 0 w       z 1     e=y−r . To meet the design specifications, we require the closed-loop poles to lie to the left of the line in the complex plane defined by s = −0.8. We choose K2 = [10 3] and Gc (s) =  8 . s  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  632  CHAPTER 11  The Design of State Variable Feedback Systems  This places the closed-loop poles at s = −2, −2 and −2. The closed-loop transfer function with the internal model controller is T (s) =  8 . s3 + 6s2 + 12s + 8  The step response is shown on Figure AP11.11.  1 0.9 0.8  Amplitude  0.7 0.6 0.5 0.4 0.3 0.2 0.1 0 0  2  4  6  8  10  12  14  Time (secs)  FIGURE AP11.11 Internal model controller step response.  AP11.12  The state-space representation of the plant is ẋ = Ax + Bu y = Cx where   A=  0  1  −2 −3     ,    B=  0 1     , and  With the intermediate variables defined as z = ẍ  and w = ü  C=  h  1 0  i  .  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  633  Advanced Problems  we have      0 1   0 0      ė     ë  =         ż  where e = y − r.    0    0     e     1 0   +  ė     0 0 0 1       z 0 0 −2 −3    0   0    w  0   1    6  5  Amplitude  4  3  2  1  0 0  1  2  3  4  5  6  Time (secs)  FIGURE AP11.12 Internal model controller ramp response.  To meet the design specifications, we require the closed-loop poles to lie to the left of the line in the complex plane defined by s = −0.67. We choose  w = −[K1 K2         e   e         K3 ]   ė  = −[16 32 22 5]  ė  .      z  Then,    Gc (s) =  K1 + K2 s 16 + 32s = . 2 s s2  z  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  634  CHAPTER 11  The Design of State Variable Feedback Systems  The closed-loop transfer function with the internal model controller is T (s) =  s4  +  8s3  32s + 16 . + 24s2 + 32s + 16  This places the closed-loop poles at s = −2, −2, −2 and −2. The ramp response is shown in Figure AP11.12. AP11.13  The controllability matrix is   −5 −3    4  −3  22  44  Pc =   1     18  and the observability matrix is  Po =      .  Computing the determinants yields det Pc = −87 6= 0  and  det P0 = 242 6= 0 ,  hence the system is controllable and observable. The controller gain matrix K=  h  3.02 6.11  i  places the closed-loop poles at the desired locations. Similarly, the observer gain matrix   L=  2.38 −1.16     places the observer poles at the desired locations. AP11.14  The controllability matrix is    0  Pc =   0   0 4  4 −12  4      −12     24  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  635  Advanced Problems  and the observability matrix is   2    Po =   −16   −9  2  29  41      −4 −15   .   120  Computing the determinants yields det Pc = −64 6= 0 and  det P0 = 10870 6= 0 ,  hence the system is controllable and observable. The controller gain matrix K=  h  −0.5 1.25 0.5  and the observer gain matrix    i     57.43     L=  −16.11     −104.43  yields the desired closed-loop system poles and observer poles, respectively. AP11.15  The state-variable representation of the system is   ẋ =   0  1  −7 −2      x+  0 1  y = [ 1 4 ]x + [0]u .    u  The observability matrix is   P0 =   1  4  −28 −7     ,  and det P0 = 105 6= 0, hence the system is observable. The observer gain matrix   L=  −7.18 6.29     places the observer poles at s1,2 = −10 ± 10, as desired.  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  636  CHAPTER 11  The Design of State Variable Feedback Systems  Design Problems CDP11.1  A state variable representation is   0  h  1 0  ẋ =  y=  1  0 −33.14 i  x      0  x+  0.827     va  where x1 = x and x2 = ẋ. Note that we are neglecting the motor inductance and assuming that the position x(t) is the output. Assume that we have available for feedback the angle θ and angle rate θ̇ (see CDP4.1), so that va = −  k1 k2 x1 − x2 + au r r  where u(t) is the reference input (that is, the desired position x(t)), the gains k1 and k2 and the scaling parameter a are to be determined. Recall that x = rθ = 0.03175θ . With the feedback in the loop we have   ẋ =  y=  h  0  1  −26.03k1 −33.14 − 26.03k2 1 0  i  x      x+  0 0.827a  Choosing k1 = 50, k2 = 1 and a = 1574.1 results in P.O. = 1.1%  and  Ts = 0.11 second .  The closed-loop poles are s1,2 = −29.59 ± 20.65j. DP11.1  The governing differential equation is ÿ − 2000y = −20i . In state variable form, the system is described by   ẋ =   0  1  2000 0      x +   0 −20    i .    u  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  637  Design Problems  Consider the state feedback i = −k1 x1 − k2 x2 + βr where r(t) is the reference input and k1 , k2 and β are to be determined. Then, the closed-loop system is   ẋ =   0  1  2000 − 20k1 −20k2      x+  0 −20β    r .  The characteristic equation is s2 + 20k2 s − 2000 + 20k1 = 0 . For stability, let 20k1 − 2000 > 0. Select k1 = 125. Then, ωn = 22.36 rad/sec, and k2 =  2ζωn . 20  Let ζ = 0.59 to meet 10% overshoot specification. Thus, k2 =  2(0.59)(22.36) = 1.32 . 20  The closed-loop transfer function is T (s) =  s2  −20β . + 26.4s + 500  s2  500 . + 26.4s + 500  Choose β = −25 so that T (s) = The feedback law is i = 125x1 + 1.32x2 − 25r . DP11.2  The automobile engine control system (see DP10.8 in Dorf and Bishop) is modeled as G(s) =  2e−sT . (0.21s + 1)(4s + 1)  In this case, we will assume the delay is negligible. Therefore, T = 0. A  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  CHAPTER 11  The Design of State Variable Feedback Systems  state variable representation of the system is   ẋ =   0  1  −1.19 −5.01  y = [1 0]x .      x +   0 1.19    r  Let r(t) = −k1 x1 − k2 x2 + k3 u where u(t) is the command input. Using ITAE methods, our desired characteristic polynomial is q(s) = s2 + 1.4ωn s + ωn2 = 0 . Select ωn = 11.315 to obtain a settling time Ts < 0.5 seconds. The characteristic polynomial of the closed-loop system is s2 + (5.01 + 1.19k2 )s + (1.19 + 1.19k1 ) = 0 . Equating coefficients and solving for the gains yields k1 = 106.59  and k2 = 9.235 .  Select k3 = 107.59 to yield a zero steady-state error to a step input.  1.2  1  0.8  Amplitude  638  0.6  0.4  0.2  0 0  0.1  0.2  0.3  0.4  Time (secs)  FIGURE DP11.2 The step response of the engine control system.  0.5  0.6  0.7  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  639  Design Problems  DP11.3  The compensator is x̂˙ = [A − BK − LC] x̂ + Ly + Mr ũ = −Kx̂ where   A − BK − LC =  N = 363.64 ,  K=  h  −28.7  1  −365.19 −20  344.55 15.82  i       ,  M=   , and L =   0 200     ,  28.7 165.19     .  We selected the desired eigenvalues of A − BK at p = −10 ± 10j and the desired eigenvalues of A − LC at q = −20 ± 10j. For initial conditions we let x(0) = [1 1] and x̂(0) = [0 0]. 1.5  Actual x1  x1  1  Estimated x1  0.5  0  0  0.1  0.2  0.3  0.4  0.5 Time (s)  0.6  0.7  0.8  0.9  1  0.4  0.5 Time (s)  0.6  0.7  0.8  0.9  1  6  x2  4  Estimated x2  2 0  −2  Actual x2 0  0.1  0.2  0.3  FIGURE DP11.3 The step response showing the actual and estimated states.  DP11.4  The design specifications are (a) Percent overshoot < 20% (b) Ts < 1.5s, and  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  640  CHAPTER 11  The Design of State Variable Feedback Systems  (c) steady-state error less than 20% of the input magnitude. The state differential equation is ẋ = Ax + Bu y = Cx where             1 0   0  0      A=  0 −σ1 −α1  =  0 g −α2 −σ2              0   0      B=  n  =  6.27  g  The transfer function is  9.8         9.8  1 −0.415 −1.43  and C =    0  h    −0.0111   ,   −0.0198  1 0 0  i  .  ns + nσ2 − α1 g θ(s) = 3 δ(s) s + (σ1 + σ2 )s2 + (σ1 σ2 − α1 α2 )s + α1 g 6.27s + 0.0154 = 3 . s + 0.435s2 − 0.0077 + 0.109 Let u = −K1 x1 − K2 x2 − K3 x3 . Then the closed-loop system matrix is   0   A − BK =   −nK1   1 −σ1 − nK2  g − gK1 −α2 − gK2  0      −α1 − nK3   ,   −σ2 − gK3  where K = [K1 K2 K3 ]. From the design specifications, we have the desired roots at s3 +a2 s2 +a1 s+ao = s3 +36s2 +225s+1350 = (s+30)(s+3+j6)(s+3−j6) = 0 . The actual characteristic equation is s3 + (gK3 + K2 n + σ1 + σ2 )s2 + (−α1 α2 − α1 gK2 + K1 n − α2 nK3 + gK3 σ1 + K2 nσ2 + σ1 σ2 )s + α1 g − α1 gK1 + gK3 n + σ2 nK1 = 0 .  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  641  Design Problems  Comparing coefficients yields        0  n  n −α1 g + σ2 n    g          a2 − σ 1 − σ 2   K1         nσ2 − α1 g −α2 n + gσ1    K2  =  a1 + α1 α2 − σ1 σ2 0    gn  where  K3  a0 − α1 g  a2 = 36 a1 = 225 a0 = 1350 . The solution for K is K = [53.11 − 28.64 21.96] . DP11.5  The controllability and observability matrices are   Pc =     P0 =   0.05  −0.04  0.001 −0.001 1  0  −0.8 0.02  Computing the determinants yields det Pc = −1.002e − 05 6= 0       and   , respectively.  and Po = 0.02 6= 0 ,  hence the system is controllable and observable. The feedback gain matrix K = [ 3820 −179620 ] yields the desired closed-loop poles. The observer gain matrix   L=  120 180000     yields the desired observer poles. The integrated system is shown in Figure DP11.5.         © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  642  CHAPTER 11  The Design of State Variable Feedback Systems  A=  -0.8 0.02 -0.02 0  B=  0.05 0.001  System Model  u  . x=Ax+Bu  y  C  Observer  Control Law  -K  x  ^ x  . ^ ^ ~ x=Ax+Bu+Ly  ~ ^ y=y-Cx  +  C= 1 0  C  K= 3820 -179620 L=  120 180000  FIGURE DP11.5 Integrated controller and observer.  DP11.6  (a) The characteristic equation associated with the system matrix is q(s) = s2 + (12 + K2 )s + (36 + K1 ) = 0 , where we have assumed state feedback of the form u = −K1 x1 − K2 x2 . The deadbeat control characteristic equation is s2 + αωn s + ωn2 = 0 , where α = 1.82 and we use ωn = 9.64 to meet the settling time specification. Then, equating coefficients and solving for the gains yields K1 = 56.93  and K2 = 5.54 .  (b) Since the closed-loop poles are located at s1,2 = −8.77 ± 4, we can select the observer poles to be about ten times farther in the left-half plane, or s1,2 = −88, −88 .  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  643  Design Problems  Then the observer gains are   L=    164   .  5740  (c) The block diagram is shown in Figure DP11.6.  0 1 -36 -12  A=  B=  0 1  System Model  u  y  C  Observer  Control Law  -K  x  . x=Ax+Bu  ^ x  . ^ ^ ~ x=Ax+Bu+Ly  ~ ^ y=y-Cx  +  C= 1 0  C  K= 56.93 5.54 L=  164 5740  FIGURE DP11.6 Block diagram for integrated controller and observer.  DP11.7  The compensator is x̂˙ = [A − LC] x̂ + Ly + Bu u = −Kx̂ where   −60   A − LC =   −1095   1 0    0   1   ,  −3750 −5 −10    © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  644  CHAPTER 11  The Design of State Variable Feedback Systems  N = 4000 ,       60  h i    K = 3998 595 30 , and L =   1095  .     3748  We selected the desired eigenvalues of A − BK at p1,2 = −10 ± 10j, p3 = −20 and the desired eigenvalues of A − LC at q1,2 = −20 ± 10j, q3 = −30. For initial conditions we let x(0) = [1 1 1] and x̂(0) = [0 0 0]. The transfer function from r to y is T (s) =  4000s3 + 2.8e05s2 + 6.8e06s + 6e07 . s6 + 110s5 + 5100s4 + 1.29e05s3 + 1.9e06s2 + 1.58e07s + 6e07  The bandwidth is 11.7 rad/s.  1.5 1 0.5 0  0  0.1  0.2  0.3  0.4  0.5  0.6  0.7  0.8  0.9  1  0  0.1  0.2  0.3  0.4  0.5  0.6  0.7  0.8  0.9  1  0  0.1  0.2  0.3  0.4  0.5  0.6  0.7  0.8  0.9  1  20 10 0 −10 200 100 0 −100 −200  FIGURE DP11.7 The step response showing the actual and estimated states.  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  645  Computer Problems  Computer Problems CP11.1  The controllability and observablity matrices have nonzero determinants, as shown in Figure CP11.1. Therefore, the system is observable and controllable. >>  A=[-6 2 0;4 0 7;-10 1 11]; b=[5;0;1]; c=[ 1 2 1]; d=[0]; sys = ss(A,b,c,d); Co=ctrb(sys); dt_Co=det(Co) Ob=obsv(sys); dt_Ob=det(Ob)  dt_Co = -84933 dt_Ob = -3.6030e+03  FIGURE CP11.1 Determining controllability and observability.  CP11.2  The system is controllable since the determinant of the controllability matrix is nonzero , as shown in Figure CP11.2. a=[0 1;-6 -5]; b=[0;6]; c=[1 0]; d=[0]; sys_ss = ss(a,b,c,d); Pc=ctrb(sys_ss); dt_Pc=det(Pc) Ob=obsv(sys_ss); dt_Ob=det(Ob) sys_tf=tf(sys_ss)  dt_Pc = -36 dt_Ob = 1  Transfer function: 6 ------------s^2 + 5 s + 6  FIGURE CP11.2 M-file script to determine controllability and to compute equivalent transfer function model.  CP11.3  The gain matrix (computed as shown in Figure CP11.3) is K =  a=[0 1;-1 -2]; b=[1;1]; c=[1 -1]; d=[0]; p=[-1;-2]; K=acker(a,b,p) K= 0.5000  0.5000  FIGURE CP11.3 M-file script to place the closed-loop system poles using state feedback.  h  i  0.5 0.5 .  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  646  CHAPTER 11  CP11.4  The Design of State Variable Feedback Systems  The constant velocity guided missile is not controllable since the controllablity matrix, Co , has a zero determinant, as shown in Figure CP11.4. Using the tf function (see Figure CP11.4), we determine that the transfer function is G(s) =  s5  5s . + 0.5s4 + 0.1s3  Cancelling common terms in the transfer function yields G(s) =  s4  5 . + 0.5s3 + 0.1s2  Then, using the ss function, we determine a state-space representation of G(s). As shown in Figure CP11.4, the state-space representation is ẋ = Ax + Bu y = Cx  A=[0 1 0 0 0;-0.1 -0.5 0 0 0;0.5 0 0 0 0;0 0 10 0 0;0.5 1 0 0 0]; b=[0;1;0;0;0]; c=[0 0 0 1 0]; d=[0]; sys_ss = ss(A,b,c,d); Transfer function: % Part (a) dt_Co = 5s Co=ctrb(sys_ss); dt_Co=det(Co) 0 ----------------------% Part (b) s^5 + 0.5 s^4 + 0.1 s^3 sys_tf = tf(sys_ss) sys_new = minreal(sys_tf ); sys_new_ss=ss(sys_new) a= % Part (c) x1 x2 x3 x4 Co_new=ctrb(sys_new_ss); dt_Co_new=det(Co_new) x1 -0.50000 -0.10000 0 0 % Part (d) x2 1.00000 0 0 0 evalues=eig(sys_new_ss) x3 0 1.00000 0 0 dt_Co_new = x4 0 0 2.00000 0 32 b= u1 evalues = x1 2.00000 0 x2 0 0 x3 0 -0.2500 + 0.1936i x4 0 -0.2500 - 0.1936i c= x1 x2 x3 x4 y1 0 0 0 1.25000 d= u1 y1 0 Continuous-time system.  FIGURE CP11.4 Analysis of the constant velocity guided missile state-space model.  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  647  Computer Problems  where      −0.5 −0.1 0 0     1 0 0 0     A=      0     0  1  0  0  0  2 0       2     0     B=    0       and C =  h  0 0 0 1.25  i  0  The reduced system is controllable but not stable, since there are two poles at the origin. Systems that are not controllable have too many states. After eliminating unnecessary states, a controllable system of minimal complexity (i.e. states) is obtained. In this case, the number of states is reduced from five to four. CP11.5  The eigenvalues of A are e1 = −2.0727 e2 = −0.2354 e3,4 = 0.2761 ± 0.2593j The system is unstable since there are two eigenvalues in the right halfplane, see Figure CP11.5. The characteristic polynomial is  A = [-0.0389 0.0271 0.0188 -0.4555; 0.0482 -1.0100 0.0019 -4.0208; >> 0.1024 0.3681 -0.7070 1.4200; 0 0 1 0]; evalues = b1 = [0.4422;3.5446;-6.0214;0]; 0.2761 + 0.2593i b2 = [0.1291;-7.5922;4.4900;0]; 0.2761 - 0.2593i % Part (a) -0.2354 evalues = eig(A) -2.0727 %part (b) p= p = poly(A) 1.0000 1.7559 -0.6431 0.0618 0.0700 r = roots(p) % Part (c) dt1 = Co1 = ctrb(A,b1); dt1 = det(Co1) -1.8451e+03 Co2 = ctrb(A,b2); dt2 = det(Co2)  r= -2.0727 0.2761 + 0.2593i 0.2761 - 0.2593i -0.2354  dt2 = -90.6354  FIGURE CP11.5 Analysis of the VTOL aircraft model.  p(s) = s4 + 1.7559s3 − 0.6431s2 + 0.0618s + 0.0700 . The roots of the characteristic equation are the same as the eigenvalues. Also, the system is controllable from either u1 or u2 . If the aircraft should lose the control of the vertical motion through u1 , then the control u2 can  .  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  648  CHAPTER 11  The Design of State Variable Feedback Systems  be used to control both vertical and horizontal motion, and vice versa. CP11.6  The m-file script to analyze the translunar halo orbit problem is shown in Figure CP11.6. The translunar equilibrium point is not a stable point  A=[0 0 0 1 0 0;0 0 0 0 1 0; 0 0 0 0 0 1;7.3809 0 0 0 0 -2.1904 0 -2 0 0; 0 0 -3.1904 0 0 0]; c=[0 1 0 0 0 0];d=[0]; b1=[0;0;0;1;0;0]; b2=[0;0;0;0;1;0]; b3=[0;0;0;0;0;1]; sys_ss_1 = ss(A,b1,c,d); sys_ss_2 = ss(A,b2,c,d); dt1 = sys_ss_3 = ss(A,b3,c,d); 0 % Part (a) evalues=eig(A) dt2 = % Part (b) 0 Cb1=ctrb(sys_ss_1); dt1=det(Cb1) dt3 = % Part (c) 0 Cb2=ctrb(sys_ss_2); dt2=det(Cb2) % Part (d) Cb3=ctrb(sys_ss_3); dt3=det(Cb3) % Part (e) sys_tf = tf(sys_ss_2); sys_tf=minreal(sys_tf ) % Part (f ) sys_ss=ss(sys_tf ); Co=ctrb(sys_ss); dt_Co=det(Co) if dt_Co ~= 0 disp('System is completelly Controllable') else disp('System in uncontrollable') end % Part (g) P = [-1+i; -1-i;-10;-10]; [A,B]=ssdata(sys_ss); K = acker(A,B,P) dt_Co =  2 0;  evalues = 2.1587 -2.1587 0 + 1.8626i 0 - 1.8626i 0 + 1.7862i 0 - 1.7862i Transfer function: s^2 - 7.381 ---------------------------s^4 - 1.19 s^2 - 16.17  a= x1 x2 x3 x4  x1 x2 x3 x4 0 0.59525 0 2.02089 2.00000 0 0 0 0 2.00000 0 0 0 0 2.00000 0  x1 x2 x3 x4  u1 1.00000 0 0 0  b=  c= y1  x1 0  y1  u1 0  x2 0.50000  x3 0  x4 -0.92261  d=  64 System is completelly Controllable  FIGURE CP11.6 Analysis of the translunar satellite halo orbit.  as evidenced by the eigenvalues of A in the right half-plane; the system is not completely controllable from any ui individually. The transfer function  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  649  Computer Problems  from u2 to η is T (s) =  s6  s4 − 4.191s2 − 23.55 . + 2s4 − 19.97s2 − 51.58  A careful analysis reveals that T (s) can be reduced by eliminating common factors. The common factors are s2 + 3.1834. The reduced transfer function is T (s) =  s2 − 7.3815 . s4 − 1.1837s2 − 16.2030  Using state feedback u2 = −Kx the gain matrix K which places the desired poles (using Ackermann's formula) is K= CP11.7  h  22 71.56 60 27.02  i  .  The m-file script to determine the initial state is shown in Figure CP11.7a. Given three data points at t = 0, 2, 4, we construct the three equations A=[0 1 0;0 0 1;-2 -4 -6]; b=[0;0;0]; c=[1 0 0]; d=[0]; sys=ss(A,b,c,d); % % Part (b) v1=c*expm(0*A); v2=c*expm(2*A); v3=c*expm(4*A); V=[v1;v2;v3]; Vi=inv( V ); n=[1;-0.0256;-0.2522]; x0=Vi*n % % Part (c) t=[0:0.1:4]; u=0.0*t; [y,x]=lsim(sys,u,t,x0'); plot(t,y,[0 2 4],[1;-0.0256;-0.2522],'*'), grid xlabel('Time (sec)'), ylabel('y(t)') title('Data points denoted by *')  FIGURE CP11.7 (a) Script to determine the initial state from three observations.  y(0) = 1 = Ce0A x0 y(2) = −0.0256 = Ce2A x0  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  CHAPTER 11  The Design of State Variable Feedback Systems  y(4) = −0.2522 = Ce4A x0 or, in matrix form   Ce0A     Ce2A    Ce4A        1        x0 =  −0.0256  .       −0.2522  The problem is solvable if the matrix     0A  Ce     Ce2A      Ce4A     is invertible. In this case, the inverse does exist and the solution is     x0 =     1 −1 1.9998       .    The simulation is shown in Figure CP11.7b.  Data points denoted by * 1*  0.8  0.6  0.4  y(t)  650  0.2  0  *  -0.2 *  -0.4  0  0.5  1  1.5  2  2.5  3  Time (sec)  FIGURE CP11.7 CONTINUED: (b) System simulation using computed initial state.  3.5  4  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  651  Computer Problems  CP11.8  Suppose we are given   A=  0    1    −1 0    B=    0    1  and the feedback u = −Kx = −[K1 K2 ]x . Solving HT P + PH = −I for P yields K22 + K12 + 3K1 + 2 2(K1 + 1)K2 1 = 2(K1 + 1) K1 + 2 = 2(K1 + 1)K2  p11 = p12 p22  Then, with xo T = [1, 0] we find that J = xo T Pxo = p11 . Computing the partial of J with respect to K2 yields 1 ∂J 1 K1 + 2 = − ∂K2 2 K1 + 1 K22     .  Setting ∂J =0 ∂K2 and solving for K2 , we find that K2 =  q  (K1 + 2)(K1 + 1) .  For a given value of K1 , the value of K2 that minimizes J can be computed via the above equation. With K2 given as above, we can compute J to be J=  s  K1 + 2 . K1 + 1  A plot of J versus K1 (with K2 equal to the minimizing value) is shown in Figure CP11.8. As K1 increases, the performance index J decreases. However, we see that the rate of decrease slows considerably after K1 > 20. Also, K2 increases as K1 increases. We want to keep both gains as  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  652  CHAPTER 11  The Design of State Variable Feedback Systems  small as possible, while still having a small J. A reasonable selection is K1 = 20  and K2 = 21.5 .  Performance index J versus K1  1.5 1.4  J  1.3 1.2 1.1 1 0  5  10  15  20  25 K1  30  35  40  45  50  30  35  40  45  50  K2 versus K1  60  K2  40 20 0 0  5  10  15  20  25 K1  FIGURE CP11.8 Performance index as a function of K1 and K2 .  CP11.9  In this problem, A = −1 and B = 1. Computing Q yields Q = (1 + λ(−k)2 ) = 1 + λk 2 . Define H = A − Bk = −1 − k . Solving H T P + P H = −Q yields p=  1 + λk 2 . 2(k + 1)  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  653  Computer Problems  0.5 0.49 0.48  J/x0^2  0.47 0.46 0.45 0.44 0.43 0.42 0.41  0  0.1  0.2  0.3  0.4  0.5  0.6  0.7  0.8  0.9  1  6  7  8  9  10  k 2.5  2  k min  1.5  1  0.5  0 0  1  2  3  4  5 lambda  FIGURE CP11.9 Plot of J/x20 versus k and the minimizing k versus λ.  The performance index is J = x20 p which implies  J/x20 =  1 + λk 2 . 2(k + 1)  The plot of J/x20 versus k is shown in Figure CP11.9. The minimum value is achieved when k = 0.41. To arrive at this result analytically, take the  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  654  CHAPTER 11  The Design of State Variable Feedback Systems  partial of J/x20 with respect to k, set the result to zero and solve for k: ∂J/x20 = 0 when k 2 + 2K − 1/λ = 0 . ∂k p  Solving for k yields k = −1 ± 1 + 1/λ. So, when λ = 1, k = 0.41. The plot of kmin versus λ is shown in Figure CP11.9. CP11.10  The m-file is shown in Figure CP11.10. A=[0 1;-18.7 -10.4]; B=[10.1; 24.6]; C=[1 0]; D=[0]; % Controller Gains p=[-2;-2 ]; K=acker(A,B,p)  >> K= -0.3081 -0.1337  % Observer Gains q=[-20+4*j;-20-4*j]; L = acker(A',C',q); L=L  L= 29.6000 89.4600  FIGURE CP11.10 Using the acker function to compute the controller gains and the observer gains.  CP11.11  The m-file is shown in Figure CP11.11(a). The compensator can be repA=[0 1 0;0 0 1;-4.3 -1.7 -6.7]; B=[0;0;0.35]; C=[0 1 0]; D=[0]; % Controller Gains p=[-1.4+1.4*j;-1.4-1.4*j;-2]; K=acker(A,B,p) % Observer Gains q=[-18+5*j;-18-5*j;-20]; L = acker(A',C',q); L=L'  >> K= 10.1143 22.3429 -5.4286  L= 1.0e+003 *  % Simulation of closed-loop system with the observer Ac=[A -B*K;L*C A-B*K-L*C]; Bc=[zeros(6,1)]; Cc=eye(6); Dc=zeros(6,1); sys=ss(Ac,Bc,Cc,Dc); x0=[1;0;0;0.5;0.1;0.1]; t=[0:0.001:3.5]; [y,t]=initial(sys,x0,t); subplot(311) plot(t,y(:,1),t,y(:,4),'--'), grid subplot(312) plot(t,y(:,2),t,y(:,5),'--'), grid subplot(313) plot(t,y(:,3),t,y(:,6),'--'), grid  FIGURE CP11.11 (a) M-file.  -1.6223 0.0493 0.7370  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  655  Computer Problems  resented as x̂˙ = (A − BK − LC)x̂ + Ly  and  u = −Kx̂ .  Since y = Cx, we can write x̂˙ = (A − BK − LC)x̂ + LCx . Similarly, with ẋ = Ax + Bu  and u = −Kx̂  we obtain ẋ = Ax − BKx̂ . In matrix form, we have    ẋ x̂˙      =  −BK  A  LC A − BK − LC     x x̂     ,  with initial conditions h  x(0)T  x̂(0)T  iT  =  h  1 0 0 0.5 0.1 0.1  iT  .  The response of the system is shown in Figure CP11.11(b). 5  Estimated state (dashed line)  x1 0 True state 5 0 (solid line) 1  1  2  3  4  0  1  2  3  4  0  1  3  4  x2 0 1 2  x3  0 2  2 Time (sec)  FIGURE CP11.11 CONTINUED: (b) Response of system to an initial condition.  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  656  CHAPTER 11  CP11.12  The Design of State Variable Feedback Systems  The Simulink block diagram is shown in Figure CP11.12.  FIGURE CP11.12 Simulink block diagram.  CP11.13  The m-file to design the compensator is shown in Figure CP11.13(a). The Simulink simulation is shown in Figure CP11.13(b). The output shown on the x-y graph depicts the state x of the system. The initial conditions selected for the simulation are      1     0     x(0) =     0       0       0.5     0.1     and x̂(0) =    .  0.1       0.1  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  657  Computer Problems  A=[0 1 0 0;0 0 1 0;0 0 0 1;-2 -5 -1 -13]; B=[0;0;0;1]; C=[1 0 0 0]; D=[0]; >> K=  % Controller Gains p=[-1.4+1.4*j;-1.4-1.4*j;-2+j;-2-j]; K=acker(A,B,p)  17.6000 24.6800 19.1200 -6.2000  % Observer Gains q=[-18+5*j;-18-5*j;-20;-20]; L = acker(A',C',q); L=L'  L=  % Simulation of closed-loop system with the observer Ac=[A -B*K;L*C A-B*K-L*C]; Bc=[zeros(8,1)]; Cc=eye(8); Dc=zeros(8,1); sys=ss(Ac,Bc,Cc,Dc); x0=[1;0;0;0;0.5;0.1;0.1;0.1]; t=[0:0.001:10]; [y,t]=initial(sys,x0,t); subplot(311) 100 plot(t,y(:,1),t,y(:,4),'--'), grid subplot(312) 0 plot(t,y(:,2),t,y(:,5),'--'), grid subplot(313) ?100 0 2 plot(t,y(:,3),t,y(:,6),'--'), grid 2  63 1369 10495 1479  4  6  8  10  0 ?2 10  0  2  4  6  8  10  0  2  4  6  8  10  0 ?10  FIGURE CP11.13 (a) M-file to design the compensator, including the observer.  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  658  CHAPTER 11  The Design of State Variable Feedback Systems  . ^x=[A-BK-LC]x+Ly ^ ^ u=-Kx  FIGURE CP11.13 CONTINUED (b) The Simulink simulation.  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  C H A P T E R  1 2  Robust Control Systems  Exercises E12.1  The plant transfer function is G(s) =  3 . s+3  Try a PI controller, given by Gc = K1 +  K2 . s  The ITAE characteristic equation is s2 + 1.4ωn s + ωn2 , where ωn = 30. Then K1 = 13  and  K2 = 300 .  Without a prefilter, the closed-loop system is Y (s) 39s + 900 = 2 , R(s) s + 42s + 900 and with a prefilter, the closed-loop system is Y (s) 900 = 2 , R(s) s + 42s + 900 where Gp (s) =  23.07 . s + 23.07  The step response, with and without the prefilter, is shown in Figure E12.1.  659  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  660  CHAPTER 12  Robust Control Systems  1.4 Without prefilter With prefilter 1.2  1  y(t)  0.8  0.6  0.4  0.2  0  0  0.05  0.1  0.15  0.2  0.25 Time (sec)  0.3  0.35  0.4  0.45  0.5  FIGURE E12.1 Step response: (a) w/o prefilter (solid line), and (b) w/prefilter (dashed line).  The disturbance response is shown in Figure E12.2.  0.05  0.04  0.03  y(t)  E12.2  0.02  0.01  0  −0.01  0  0.05  0.1  0.15  0.2  FIGURE E12.2 Disturbance response for system in E12.1.  0.25 Time (sec)  0.3  0.35  0.4  0.45  0.5  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  661  Exercises  E12.3  The closed-loop transfer function is T (s) =  s2  25 , + bs + 25  and the sensitivity function is S(s) =  s2 + bs , s2 + bs + 25  where b = 8, nominally. The sensitivity of T to changes in b is determined to be SbT =  ∂T b −bs = 2 . ∂b T s + bs + 25  The plot of T (s) and S(s) is shown in Figure E12.3, where b = 8.  10  0  Gain dB  −10  20log|T|  −20 20log|S| −30  −40  −50  −60 −1 10  0  1  10  10 Frequency (rad/sec)  FIGURE E12.3 Plot of T (s) and the sensitivity function S(s).  E12.4  The plant transfer function is G(s) =  1 , (s + 20)(s + 36)  2  10  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  662  CHAPTER 12  Robust Control Systems  and the PID controller is given by K3 (s + a)(s + b) . s  Gc (s) =  Let a=20, b=500, and K3 = 200. Then, the closed-loop system is T (s) =  200s2 + 4000s + 100000 . s3 + 256s2 + 4720s + 100000  The closed-loop poles are s1 =-237.93 and s2,3 = −9.04 ± j18.5 and the zeros are s1,2 = −10±j20. Therefore, there is an approximate cancellation of the complex poles and zeros and the approximate system is T̂ (s) =  238 . s + 238  The actual response and approximation are shown in Figure E12.4.  1.4  1.2 actual 1  approximation  y(t)  0.8  0.6  0.4  0.2  0  0  0.05  0.1  0.15  0.2  0.25 0.3 Time (sec)  0.35  0.4  0.45  FIGURE E12.4 Step response for closed-loop actual and approximate transfer functions.  E12.5  The loop transfer function is L(s) = Gc (s)G(s) =  10KD (s + KP /KD ) . s(s + 3)(s + 10)  0.5  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  663  Exercises  Select KP /KD = 10. Then L(s) = Gc (s)G(s) =  10KD , s(s + 3)  and the closed-loop transfer function is T (s) =  10KD . s2 + 3s + 10KD  Let ζ = 0.69, which implies P.O. < 5%. Also, 2ζωn = 3, so ωn = 2.17. Thus, 10KD = ωn2 = 4.72 . Thus, the controller is Gc (s) = 0.47(s + 10). The settling time is Ts = 2.8 s and the percent overshoot is P.O. = 4.6%. As K increases, the percent overshoot increases from 0% to 16% and the settling time generally decreases from 3.8 sec to 2.6 sec. E12.6  The loop transfer function with the PID controller is Gc (s)Gs(s) =  KD s2 + KP s + KI 1 . s (s + 5)2  The ITAE step response requires s3 + 1.75ωn s2 + 2.15ωn2 s + ωn3 = s3 + (10 + KD )s2 + (25 + KP )s + KI . For n = 3 we estimate the normalized settling time to be ωn Ts ≈ 8 seconds. Thus, ωn ≈ 6, and KD = 0.5,  KP = 52.4,  and  KI = 216.  The step response is shown in Figure E12.6. The transfer function from the disturbance to the output is Y (s) G(s) s = = 3 . 2 Td (s) 1 + Gc (s)G(s) s + 10.5s + 77.4s + 216 The disturbance response is shown in Figure E12.6. The system is effective in reducing the effects of the disturbance, and the maximum output is reduced by 1/100 for a step disturbance.  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  664  CHAPTER 12  Robust Control Systems  −3  (a) Step response  1.2  10  1  8  0.8  6  x 10  (b) Disturbance response  y(t)  12  y(t)  1.4  0.6  4  0.4  2  0.2  0  0  0  0.5  1 TIme (s)  1.5  2  −2  0  0.5  1 TIme (s)  1.5  2  FIGURE E12.6 (a) Step response: w/o prefilter (solid line) and w/prefilter (dashed line); and (b) disturbance response.  E12.7  The plant transfer function is G(s) =  1 , (s + 4)2  and the PID controller is Gc (s) =  K1 s + K2 + K3 s2 . s  Using the ITAE criteria and selecting ωn = 10 yields K3 = 9.5  K2 = 1000  and K1 = 199 .  The step response is shown in Figure E12.7. The disturbance response is also shown in Figure E12.7. The maximum y(t) = 0.0041, so the system is effective in rejecting the step disturbance.  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  665  Exercises x10 -3  (a) step response 1.4  (b) disturbance  4.5 4  1.2 3.5 1  3 2.5  y(t)  y(t)  0.8  0.6  2 1.5 1  0.4  0.5 0.2 0 0 0  0.5  1  1.5  -0.5  2  0  0.5  Time (sec)  1  1.5  2  Time (sec)  FIGURE E12.7 (a) Step response: w/o prefilter (solid line) and w/prefilter (dashed line); and (b) disturbance response.  The maximum ωn = 60. Then K1 = 3600 and K2 = 80. The maximum control input is max |u(t)| ≈ 80. The plot of the step response and the control input u(t) is shown in Figure E12.8. (a) step response  (b) control input u(t)  1.2  90 80  1  70 60  0.8  u(t)  50  y(t)  E12.8  0.6  40 30  0.4  20 10  0.2  0 0 0  0.05  0.1  0.15  0.2  -10 0  Time (sec)  FIGURE E12.8 Step response w/o prefilter; and (b) control input u(t).  0.05  0.1 Time (sec)  0.15  0.2  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  666  CHAPTER 12  E12.9  Robust Control Systems  One possible PD controller is Gc (s) = 27.6s + 8.25s . When K=1, the system roots are s1,2 = −3.2 ± j4.3 s3 = −9.5 . The step response is shown in Figure E12.9 for K = 0.5, 1, and 1.5. K=1 (solid); K=0.5 (dashed); and K=1.5 (dotted) 1.4  1.2  1  y(t)  0.8  0.6  0.4  0.2  0  0  0.2  0.4  0.6  0.8  1 1.2 Time (sec)  1.4  1.6  1.8  2  FIGURE E12.9 Step response for K = 0.5, 1, and 1.5.  E12.10  One possible PI controller is Gc (s) =  2.2s + 22 . s  When K = 1, the system roots are s1,2 = −1.31 ± j1.31, and s3 = −6.37. The step response is shown in Figure E12.10.  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  667  Exercises  K=1 (solid); K=0.5 (dashed); and K=1.5 (dotted) 1.4  1.2  1  y(t)  0.8  0.6  0.4  0.2  0  0  0.5  1  1.5  2  2.5 3 Time (sec)  3.5  4  4.5  5  FIGURE E12.10 Step response for K = 0.5, 1, and 1.5.  The plot is shown in Figure E12.11.  100 90 80 70 60 P.O. (%)  E12.11  50 40 30 20 10 0  0  0.5  1  1.5  2 Time (sec)  2.5  3  FIGURE E12.11 Percent overshoot as a function of k in the interval 0.1 ≤ k ≤ 4.  3.5  4  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  668  CHAPTER 12  The controllability matrix is   Pc =   and  c1  c2  c2 −ac1 − bc2     det Pc = c22 + [bc1 ]c2 + ac21 . For controllability we require det Pc 6= 0, hence c22 + [bc1 ]c2 + ac21 6= 0 implies c2 b q 6= − ± (b/2)2 − a c1 2 where (b/2)2 − a ≥ 0. For real-valued c1 and c2 , if (b/2)2 − a < 0, all real values of c1 and c2 are valid. Valid values of the constants are c1 = 0, c2 = 10, a = 10, and b = 3. The step response is shown in Figure E12.12.  Step Response 1.4  1.2  1  Amplitude  E12.12  Robust Control Systems  0.8  0.6  0.4  0.2  0  0  0.5  1  1.5  2 Time (sec)  2.5  FIGURE E12.12 Step response with c1 = 0, c2 = 10, a = 10, and b = 3.  3  3.5  4  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  669  Problems  Problems The closed-loop transfer function is T (s) =  4(s + 2) s2 + 4s + 8  and the sensitivity function is S(s) =  s2 . s2 + 4s + 8  The plot of 20 log |T | and 20 log |S| is shown in Figure P12.1. The bandwidth is ωB = 6.31 rad/sec . Then T |SK |ωB = 0.98 T ω |SK | B = 0.78 2  T ω |SK | B 4  = 0.30 .  10  0  -10 20log|T| 20log|S| -20 Gain dB  P12.1  -30  -40  -50  -60 -1 10  0  10 Frequency (rad/sec)  FIGURE P12.1 Plot of T (s) and the sensitivity function S(s).  10  1  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  670  CHAPTER 12  (a) The loop transfer function is given by Gc (s)G(s) =  K . s(0.02s + 1)(0.002s + 1)  When K = 100 , the peak magnitude is Mpω = 1.84 . (b) The plot of 20 log |T | and 20 log |S| is shown in Figure P12.2a. 20 20log|S| 0  -20  Gain dB  P12.2  Robust Control Systems  20log|T|  -40  -60  -80  -100  -120 101  102  103 Frequency (rad/sec)  FIGURE P12.2 (a) Plot of T (s) and the sensitivity function S(s).  (c) The bandwidth is ωB = 117 rad/sec , and T |SK |ωB = 1.47 T ω |SK | B = 0.39 4  T ω |SK | B = 1.62 . 2  104  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  671  Problems  (c) The disturbance response is shown in Figure P12.2b. x10 -7 8 7 6  Amplitude  5 4 3 2 1 0 0  0.05  0.1  0.15  0.2  0.25  0.3  Time (secs)  FIGURE P12.2 CONTINUED: (b) Disturbance response for K = 100.  P12.3  (a) The loop transfer function is L(s) = Gc (s)G(s) =  K(s − 4)(s − 1) . (s + 0.02)(s + 2)2  The characteristic equation is 1 + Gc (s)G(s) = 1 + K  (s − 4)(s − 1) =0 (s + 0.02)(s + 2)2  or s3 + (4.02 + K)s2 + (4.08 − 5K)s + 0.08 + 4K = 0 . Using Routh-Hurwitz we find that the system is stable for −4.6987 < K < 0.6947 . (b) The steady-state error is ess =  1 . 1 + 50K  Select K = 0.18 to obtain a steady-state error to a unit step of 0.1.  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  672  CHAPTER 12  Robust Control Systems  (c,d) The plots of y(t) for K = 0.18 K = 0.21 K = 0.15  (nominal) (+15%) (−15%)  are shown in Figure P12.3.  K=0.18 (solid) & K=0.21 (dashed) & K=0.15 (dotted) 1.2  1  0.8  y(t)  0.6  0.4  0.2  0  −0.2  0  5  10  15  20  25  Time (sec)  FIGURE P12.3 Step input response for K = 0.18, K = 0.21 and K = 0.15.  P12.4  (a) The plant is given by G=  s  1  . +1  s 25  We desire P.O. < 10% and Ts < 100 ms. Using a PD controller Gc (s) = 100 + 2.2s , we determine that P.O. = 7%, Ts < 100 ms and ess = input. The plot of y(t) is shown in Figure P12.4.  A 100  for a ramp  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  673  Problems  (b) The sensitivity is r |SK | = 27.95 1  when K1 = 1. (c) The plot of y(t) when K1 = 2 (the compensator Gc (s) is unchanged) is shown in Figure P12.4. (d) The disturbance response is shown in Figure P12.4. (b) disturbance 0.012  1  0.01  0.8  0.008  y(t)  y(t)  (a) step response 1.2  0.6  0.006  0.4  0.004  0.2  0.002  0 0  0.05  0.1  0.15  0 0  0.2  0.05  Time (sec)  0.1  0.15  0.2  Time (sec)  FIGURE P12.4 (a) Step response: K1 = 1 (solid line) and K1 = 2 (dashed line); and (b) disturbance response.  P12.5  (a) The plant is given by G(s) =  1 s(s + p)  where p = 2, nominally. One solution is Gc (s) =  18.7(s + 2.9) . (s + 5.4)  Then, T (s) =  18.7(s + 2.9) √ √ . (s + 3.41)(s + 2 + 2 3j)(s + 2 − 2 3j)  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  674  CHAPTER 12  Robust Control Systems  (b,d) The step responses are shown in Figure P12.5 for p = 2 and p = 1. (c,d) The disturbance responses are shown in Figure P12.5 for p = 2 and p = 1. (a) step response  (b) disturbance  1.6  0  1.4  -0.02  1.2  -0.04  1  y(t)  y(t)  -0.06 0.8  -0.08 0.6 -0.1  0.4  -0.12  0.2 0 0  -0.14  5  0  Time (sec)  5 Time (sec)  FIGURE P12.5 (a) Step response: p = 2 (solid line) and p = 1 (dashed line); and (b) disturbance response: p = 2 (solid line) and p = 1 (dashed line).  P12.6  (a) The plant is given by G(s) =  s(s2  1 , + 4s + 5)  and the PID controller is Gc (s) =  K(s + z)2 . s  When z = 1.25 and K=4, all roots are s = −1 ± j1.22 .  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  675  Problems  Then, the closed-loop transfer function is T (s) =  4(s + 1.25)2 . s4 + 4s3 + 9s2 + 10s + 6.25  (b,c) The step responses with and without a prefilter are shown in Figure P12.6. (d) The disturbance response is shown in Figure P12.6. (a) step response  (b) disturbance  1.6  0.02  1.4 0 1.2 -0.02  y(t)  y(t)  1 0.8  -0.04  0.6 -0.06 0.4 -0.08 0.2 0 0  5  10  -0.1  0  Time (sec)  5  10  Time (sec)  FIGURE P12.6 (a) Step response: w/o prefilter (solid line) and w/prefilter (dashed line); and (b) disturbance response.  P12.7  (a) The loop transfer function is Gc (s)G(s) =  10Ka (5s + 500 + 0.0475s2 ) . s3  When Ka = 374.5 , the phase margin is P.M. = 40o . (b) The root locus is shown in Figure P12.7a.  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  676  CHAPTER 12  Robust Control Systems  150 *  100  o  Imag Axis  50 0  x  *  -50 o  -100 *  -150 -150  -100  -50  0  50  100  150  Real Axis  FIGURE P12.7 10(0.0475s2 +5s+500) = 0. (a) Root locus for 1 + Ka s3  When Ka = 374.5 , the roots are s1 = −139.8 s2,3 = −19.1 ± j114.2 . (c) The transfer function from Td (s) to Y (s) is Y (s) −s = 3 . 2 Td (s) s + 182s + 19150s + 1915000 The maximum is max |y(t)| = 0.0000389 . (d) The step responses, with and without a prefilter, are shown in Figure P12.7b. P12.8  The polynomial under investigation is s3 + 3s2 + 3s + 4 = 0 .  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  677  Problems  1.6 1.4 1.2  y(t)  1 0.8 0.6 0.4 0.2 0 0  0.05  0.1  0.15  0.2  0.25  0.3  Time (sec)  FIGURE P12.7 CONTINUED: (b) Step response: w/o prefilter (solid line) and w/prefilter (dashed line).  From the uncertainty bounds on the coefficients, we define α0 = 4 α1 = 1 α2 = 2  β0 = 5 β1 = 4 β2 = 4  Then, we must examine the four polynomials: s3 + 2s2 + 4s + 5 = 0 s3 + 4s2 + s + 4 = 0 s3 + 4s2 + 4s + 4 = 0 s3 + 2s2 + s + 5 = 0 The fourth polynomial is not stable—therefore, the system is not stable for the uncertain parameters. P12.9  One possible PID controller is Gc (s) =  0.058s2 + 2.17s + 16.95 . s  A first-order Pade approximation was used in the design to account for the delay system. The step input response is shown in Figure P12.9. A prefilter should also be used with the PID controller. A suitable prefilter  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  678  CHAPTER 12  Robust Control Systems  is Gp (s) =  K2 . K3 s2 + K1 s + K2  1.2  1  y(t)  0.8  0.6  0.4  0.2  0 0  0.05  0.1  0.15  0.2  0.25  0.3  0.35  0.4  0.45  0.5  Time (sec)  FIGURE P12.9 Step response with the PID controller and prefilter.  P12.10  The PID controller is given by Gc (s) =  KD s2 + KP s + KI . s  Using the ITAE method, we desire the characteristic polynomial to be q(s) = s3 + 1.75ωn s2 + 2.15ωn2 s + ωn3 = 0 , where we select ωn = 4 to obtain a peak time of Tp = 1 second. Here we use the approximation for ITAE third-order systems that ωn Tp ≈ 4 from Figure 5.30(c) in Dorf and Bishop. The actual characteristic equation is s3 + 25KD s2 + 25KP s + 25KI = 0 . Equating coefficients and solving for the gains yields KP = 1.376 ,  KD = 0.28 ,  and  KI = 2.56 .  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  679  Problems  The step response is shown in Figure P12.10, with the prefilter Gp (s) =  KD  s2  KI . + KP s + KI  Step Response 1.4  1.2  Amplitude  1  0.8  0.6  0.4  0.2  0  0  0.5  1  1.5 Time (sec)  2  2.5  3  FIGURE P12.10 Step response with the PID controller and prefilter.  P12.11  We will design for the case where K = 1 and p = 1. The design plant is G(s) =  1 . s(s + 1)(s + 4)  The nominal plant is given by G(s) =  2.5 , s(s + 2)(s + 4)  and the PID controller is Gc (s) =  KD s2 + KP s + KI . s  Using the ITAE method, we desire the characteristic polynomial to be q(s) = s4 + 2.1ωn s3 + 3.4ωn2 s2 + 2.7ωn3 s + ωn4 = 0 ,  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  680  CHAPTER 12  Robust Control Systems  where we select ωn = 2.38 to obtain a peak time around Tp = 3 seconds. The actual characteristic equation (with the worst-case plant) is s4 + 5s3 + (4 + KD )s2 + KP s + KI = 0 . Equating coefficients and solving for the gains yields KP = 36.40, KI = 32.08, and KD = 15.26. The step response is shown in Figure P12.11, with the prefilter Gp (s) =  KD  s2  KI . + KP s + KI  1.4 Worst−case plant Nominal plant 1.2  1  y(t)  0.8  0.6  0.4  0.2  0  0  0.5  1  1.5  2  2.5 Time (sec)  3  3.5  4  4.5  5  FIGURE P12.11 Step response with the prefilter: nominal plant (dashed line) & worst-case plant (solid line).  P12.12  The transfer function is G(s) = C(sI − A)−1 B =  h  The sensitivity is G SK =  2 0  i     s  −3  5 s+K     ∂G K −Ks . = 2 ∂K G s + Ks + 5  0 1    =  s2  −6 . + Ks + 5  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  681  Advanced Problems  Advanced Problems AP12.1  Let Gp (s) = 1. A viable PID controller is Gc (s) = KP +  1000s2 + 3000s + 100 KI + KD s = . s s  The loop transfer function is Gc (s)G(s) =  1000s2 + 3000s + 100) . s(50s2 + 1)  We can check that Kv = 100, as desired. The step response is shown in Figure AP12.1. Step Response 1.4 System: syscl Peak amplitude: 1.1 Overshoot (%): 9.5 At time (sec): 0.234  1.2  Amplitude  1  0.8  0.6  0.4  0.2  0  0  0.2  0.4  0.6 Time (sec)  0.8  1  1.2  FIGURE AP12.1 Step response with PID controller.  AP12.2  For all three controllers, choose K = 1 as the design value. Also, use as the nominal points a = 2 and b = 5 for each design. ITAE methods were employed in all designs, although this did not work well for the PI controller. (a) PI controller: Let Gp (s) = 1 . Not all specifications could be met simultaneously with a PI con-  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  682  CHAPTER 12  Robust Control Systems  troller. The best over-all results are achieved when using a = 3 and b = 4.5 as the design values. An acceptable PI controller is Gc (s) = 1.2 +  3.96 . s  Controller  P.O.  Ts  Tp  |u(t)|max  PI  0%  2.29s  n.a.  4.43  PD  4.6%  1.72s  1.26s  12.25  PID  1.97%  0.65s  0.47s  37.25  TABLE AP12.2  PI, PD, and PID controller performance summary.  The final design is based on root locus methods since the ITAE methods did not produce an effective controller. The closed-loop transfer function is T (s) =  1.2s + 3.96 . s3 + 3s2 + 5.7s + 3.96  (b) PD controller: Let Gp (s) =  12.25 . 7.25 + 2.9s  The closed-loop transfer function is T (s) =  7.25 + 2.9s , s2 + 4.9s + 12.25  where the PD controller (based on ITAE methods) is Gc (s) = 7.25 + 2.9s .  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  683  Advanced Problems  (c) PID controller: Let Gp (s) =  15.5s2  1000 . + 210s + 1000  The closed-loop transfer function is T (s) =  15.5s2 + 210s + 1000 . s3 + 17.5s2 + 215s + 1000  And the PID controller (based on ITAE methods) is Gc (s) =  15.5s2 + 210s + 1000 . s  The performance of each controller is summarized in Table AP12.2. AP12.3  (a) The PID controller is Gc (s) =    KD s2 +  KP KD s  +  KI KD  s    .  Since we want P.O. < 4% and Ts < 1s, we choose the dominant closed-loop poles to have ωn = 6 and ζ = 0.8. Therefore, we place the zeros at s2 +  KP KI s+ = s2 + 10s + 36 . KD KD  Solving for the constants yields, KP = 10 , KI  KI = 36 . KD  Then, using root locus methods, we choose KD = 91 to place the roots near the zeros. The PID controller gains are computed to be KP = 910, KI = 3276 and KD = 91. (b) The loop transfer function is Gc (s)G(s) =  KD s2 + KP s + KI . s2 (s2 + 5s + 4)  The closed-loop system characteristic equation is s3 + 5s2 + 4s + KD s2 + KP s + KI = 0 . Solving for the PID gains yields KP = 73.4, KI = 216 and KD = 5.5.  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  684  CHAPTER 12  Robust Control Systems  Therefore, the controller is Gc (s) =  5.5(s2 + 13.35s + 39.3) . s  Using the prefilter Gp (s) =  s2  39.3 , + 13.35s + 39.3  we obtain the closed-loop transfer function T (s) =  s3  +  216 . + 77.4s + 216  10.5s2  The percent overshoot is P.O. ≈ 3.5% and the settling time is Ts ≈ 1.67 sec. The PID controller is Gc (s) =    KD s2 +  KP KI KD s + KD  s    .  The bounds 1 ≤ a ≤ 2 and 4 ≤ b ≤ 12 imply that 2 ≤ ωn ≤ 3.46 and 0.5 ≤ ζωn ≤ 1. One solution is to place the PID controller zeros at 1. 4 1. 2 1 Amplitude  AP12.4  0. 8 0. 6 0. 4 0. 2 0  0  0. 5  1  1. 5 Time (sec)  2  2. 5  3  FIGURE AP12.4 Family of step response with PID controller with nominal case (a, b) = (1.5, 9) denoted by the solid line.  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  685  Advanced Problems  √ s = −1 ± j 8 (i.e. ζωn = 1 and ωn = 3). So, s2 +  KI KP s+ = s2 + 2ζωn s + ωn2 = s2 + 4s + 9 . KD KD  The nominal case for design is chosen to be a = 1.5 and b = 9. Using root locus, we select KD = 2.1 to place the closed-loop characteristic roots near the zeros. Then, the PID controller gains are computed to be KP = 8.4, KI = 18.9, and KD = 2.1. The plot of the response to a step input is shown in Figure AP12.4. The off-nominal cases shown in the simulations are (a, b) = (1.2, 4), (1.4, 6), (1.6, 10), and (1.8, 12). AP12.5  To obtain a phase margin of P.M. = 49.77o , select K = 1.5, b = 36 and choose Gp (s) = 1. The PID controller is Gc (s) =  1.5(s2 + 20s + 36) . s  When K1 = 0.75, the phase margin is reduced to P.M. = 45.45o ; and when K1 = 1.25, the phase margin is increased to P.M. = 52.75o . AP12.6  With the settling time Ts = 1 and percent overshoot P.O. < 10% specifications, we target for dominant closed-loop poles with ωn = 10. Here we estimate ωn Ts ≈ 10 associated with the ITAE performance. The closedloop transfer function is T (s) = Gp (s)  1.5(KD s2 + KP s + KI ) , (1 + 1.5KD )s2 + 1.5KP s + 1.5KI  where we have neglected τ . Using the ITAE method, the desired characteristic polynomial is s2 +  √  2ωn s + ωn2 = s2 +  1.5Kp 1.5KI s+ . 1 + 1.5KD 1 + 1.5KD  Let KD = 0.25. Then solving for the remaining PID gains yields KP = 12.96 and KI = 91.67. The pre-filter is Gp (s) =  0.375s2  137.5 . + 19.45s + 137.5  Then the closed-loop transfer function (with τ = 0.001) is T (s) =  0.001s3  137.5 . + 1.375s2 + 19.45s + 137.5  The transfer function from the disturbance to the output is Y (s)/Td (s) =  0.001s3  1.5s . + 1.375s2 + 19.45s + 137.5  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  686  CHAPTER 12  Robust Control Systems  The step input response and disturbance response are shown in Figure AP12.6. (a)  (b) 0.05  1.4  1.2  0.04  1  Amplitude  Amplitude  0.03 0.8  0.02  0.6 0.01 0.4  0  0.2  0  0  0.2  0.4 0.6 Time (sec)  0.8  1  −0.01  0  0.2  0.4 0.6 Time (sec)  0.8  1  FIGURE AP12.6 (a) Input response; (b) Disturbance response.  AP12.7  The PI controller is given by Gc (s) =  KP s + KI . s  We will also use the prefilter Gp (s) =  KI . KP s + KI  Using the ITAE method, we determine that √ KP = 2ωn and KI = ωn2 . Let ωn = 2.2. Then KP = 3.11 and KI = 4.8. The step response and control u(t) are shown in Figure AP12.7.  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  687  Advanced Problems (a)  (b)  1.2  1.2  1  1  0.8  Amplitude  Amplitude  0.8  0.6  0.6  0.4  0.4 0.2 0.2  0  0 0  5  -0.2  0  5  Time (secs)  Time (secs)  FIGURE AP12.7 (a) Input response; (b) Control history u(t).  AP12.8  (a) A suitable PD controller is given by Gc (s) = 0.6 + 0.4s . The percent overshoot is P.O. = 18.8% and the peak time is Tp = 2.4 sec. (b) A suitable PI controller is given by Gc (s) = 0.15 +  0.01 . s  The percent overshoot is P.O. = 23.7% and the peak time is Tp = 7.8 sec. (c) A suitable PID controller is given by Gc (s) = 0.6 +  0.01 + 0.4s . s  The percent overshoot is P.O. = 19.9% and the peak time is Tp = 2.5 sec. (d) The PD or PID controllers are the best choices. AP12.9  A robust PID controller designed with ITAE methods will be a suitable controller. From the settling time specification we select ωn = 10, where  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  688  CHAPTER 12  Robust Control Systems  we have used ζ = 0.8. The worst case is a = 1 and  K=2.  The desired closed-loop transfer function is T (s) =  ωn3 s3 + 1.75ωn s2 + 2.15ωn2 s + ωn3  and the actual characteristic equation is q(s) = s3 + (2a + KKD )s2 + (a2 + KKP )s + KKI . Equating like terms, we find that KP = 107  KD = 7.75 .  We use as the design plant G(s) =  s+2 . s(s + 3)  1.4  1.2  1  Amplitude  AP12.10  KI = 500  0.8  0.6  0.4  0.2  0 0  0.01  0.02  0.03  0.04  0.05  0.06  0.07  0.08  0.09  0.1  Time(sec)  FIGURE AP12.10 Family of step responses with the design plant (p, q, r) = (3, 0, 2) denoted by the solid line.  Select p1 = 2  and z1 = 3  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  689  Advanced Problems  to cancel a design plant pole and zero. Then, choose p2 = 0 to have zero steady-state error to a unit step. The remaining variables K and z2 are selected based on ITAE methods, where ωn = 100. A suitable compensator is Gc (s) =  141.42(s + 3)(s + 70.71) . s(s + 2)  A plot of the step responses for various values of p, q and r is shown in Figure AP12.10. A suitable compensator is Gc (s) =  1000(s + 1.8)(s + 3.5)(s + 5.5) . s(s + 600)  1.4  1.2  1  Amplitude  AP12.11  0.8  0.6  0.4  0.2  0 0  1  2  3  4  5  6  7  8  9  10  Time (sec)  FIGURE AP12.11 Step responses with nominal plant (solid line) and off-nominal plant with all poles reduced by 50% (dashed line).  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  690  CHAPTER 12  Robust Control Systems  Design Problems The plant model with parameters given in Table CDP2.1 in Dorf and Bishop is given by: θ(s) 26.035 = , Va (s) s(s + 33.142) where we neglect the motor inductance Lm and where we switch off the tachometer feedback (see Figure CDP4.1 in Dorf and Bishop). With a PID controller ,the closed-loop system characteristic equation is s3 + (33.142 + 26.035KD )s2 + 26.035KP s + 26.035KI = 0 . A suitable PID controller is Gc (s) = 50 + s +  0.1 . s  This PID controller places the closed-loop system poles to the left of the −ζωn line necessary to meet the settling time requirement. The step response is shown below. The settling time is Ts = 0.12 second. In the steady-state the error due to a step disturbance is zero.  1.2  1  0.8 Amplitude  CDP12.1  0.6  0.4  0.2  0  0  0.05  0.1  0.15  0.2  0.25 Time (secs)  0.3  0.35  0.4  0.45  0.5  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  691  Design Problems  The closed-loop transfer function is Y (s) Km Gc (s) . = 2 R(s) s + (2 + Km K1 )s + Gc (s)Km (a) When Gc = K, we have T (s) =  s2  15K , + (2 + 15K1 )s + 15K  where Km = 15. Using ITAE criteria and ωn = 10, we determine that K1 = 0.81 and K = 6.67. For the disturbance, we have −1 Y (s) = 2 . TL (s) s + 14.14s + 100 The input and disturbance responses are shown in Figure DP12.1, without prefilters. (a) Step response  (b) Disturbance response  1.2  0  1  -0.002  0.8  -0.004  y(t)  y(t)  DP12.1  0.6  -0.006  0.4  -0.008  0.2  -0.01  0 0  0.5  1  Time (sec)  -0.012  0  0.5  1  Time (sec)  FIGURE DP12.1 (a) Step response: Gc (s) = K (solid line) and Gc (s) = KP + KD s (dashed line); and (b) disturbance response (same for both compensators).  (b) When Gc = KP + KD s, we have Y (s) 15(KP + KD s) = 2 . R(s) s + (2 + 15K1 + 15KD )s + 15KP For ωn = 10 and with the ITAE criteria, we determine that (with  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  692  CHAPTER 12  Robust Control Systems  KD = 0.1) Y (s) 15(6.67 + 0.1s) = 2 . R(s) s + 14.14s + 100 The nominal plant is given by 1 . s(s + 5)  G(s) = The closed-loop transfer function is T (s) =  K(KD s2 + KP s + KI ) . s3 + (5 + KKD )s2 + KKP s + KKI  Let KP = 450 ,  KI = 750 ,  and KD = 150 .  A family of responses is shown in Figure DP12.2 a for various values of K. The percent overshoot for 0.1 ≤ K ≤ 2 is shown in Figure DP12.2b.  1.4  1.2  1 Step response  DP12.2  0.8  0.6  0.4  0.2  0  0  0.5  1  1.5  2 Time (s)  FIGURE DP12.2 (a) Family of step responses for various values of K.  2.5  3  3.5  4  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  693  Design Problems  9 8  Percent overshoot  7 6 5 4 3 2 1 0  0  0.2  0.4  0.6  0.8  1 K  1.2  1.4  1.6  1.8  FIGURE DP12.2 CONTINUED: (b) Percent overshoot for various values of K.  DP12.3  (a) The dexterous hand model is given by G(s) =  Km , s(s + 5)(s + 10)  where Km = 1, nominally. The PID controller is Gc (s) =  KD (s2 + 6s + 18) . s  The root locus is shown in Figure DP12.3a. If we select KD = 90 , the roots are s1,2 = −5.47 ± j6.6 s3,4 = −2.03 ± j4.23 . Thus, all roots have ζωn > 4/3  2  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  CHAPTER 12  Robust Control Systems  to meet the design specification Ts < 3 sec . (b) The step responses for Km = 1 and Km = 1/2 are shown in Figure DP12.3b. When K = 1/2 , an off-nominal value, the settling time specification is no longer satisfied. 20 15 10 *  5  Imag Axis  694  o  0  x  *  x  x o *  -5 *  -10 -15 -20 -20  -15  -10  -5  0 Real Axis  FIGURE DP12.3 s2 +6s+18 (a) Root locus for 1 + KD s2 (s+5)(s+10) = 0.  5  10  15  20  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  695  Design Problems  1.6 1.4 1.2  y(t)  1 0.8 0.6 0.4 0.2 0 0  0.5  1  1.5  2  2.5  3  3.5  4  4.5  5  Time (sec)  FIGURE DP12.3 CONTINUED: (b) Step response (without prefilters): PID with K3 = 90 and Km = 1 (solid line) and PID with K3 = 90 and Km = 0.5 (dashed line).  DP12.4  The nominal plant is G(s) =  s(s2  17640 , + 59.4s + 1764)  and the PID controller is Gc (s) =  KI (τ1 s + 1)(τ2 s + 1) . s  (a) Using ITAE methods, we determine that ωn = 28.29, KI = 36.28, τ1 + τ2 = 0.0954 and τ1 τ2 = 0.00149. So, Gc (s) =  36.28(0.00149s2 + 0.0954s + 1) . s  (b) The step response for the nominal plant and the PID controller is shown in Figure DP12.4a, with and without a prefilter. (c) The disturbance response is shown in Figure DP12.4b. (d) The off-nominal plant is G(s) =  s(s2  16000 . + 40s + 1600)  The step response for the off-nominal plant is shown in Figure DP12.4a.  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  CHAPTER 12  Robust Control Systems  (b) off-nominal plant 1.6  1.4  1.4  1.2  1.2  1  1  y(t)  y(t)  (a) nominal plant 1.6  0.8  0.8  0.6  0.6  0.4  0.4  0.2  0.2  0 0  0.5  0 0  1  0.5  Time (sec)  1  Time (sec)  FIGURE DP12.4 (a) Step response for (i) nominal plant: w/o prefilter (solid line) and w/prefilter (dashed line); and (ii) for off-nominal plant: w/o prefilter (solid line) and w/prefilter (dashed line).  disturbance response 0.3  0.25  0.2  0.15  y(t)  696  0.1  0.05  0  -0.05  0  0.1  0.2  0.3  0.4  0.5  0.6  0.7  0.8  Time (sec)  FIGURE DP12.4 CONTINUED: (b) Disturbance response for the nominal plant.  0.9  1  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  697  Design Problems  DP12.5  One possible solution is Gc (s) = 0.08  (0.01s + 1)(0.99s + 1) . s  The phase margin with this controller is P.M. = 45.5o . The step response is shown in Figure DP12.5 for the nominal plant (with and without a prefilter); the step response for the off-nominal plant is also shown in Figure DP12.5. The prefilter is Gp (s) =  13.97s2  1411 . + 1411s + 1411  (a) nominal plant  (b) off nominal plant  1. 2  1. 2  1  1  0. 8  0. 8 y(t)  1. 4  y(t)  1. 4  0. 6  0. 6  0. 4  0. 4  0. 2  0. 2  0  0  10 Time (sec)  20  0  0  10 Time (sec)  20  FIGURE DP12.5 (a) Step response for nominal plant: w/o prefilter (solid line) and w/prefilter (dashed line); and (b) for off-nominal plant: w/o prefilter (solid line) and w/prefilter (dashed line).  DP12.6  Using ITAE methods, three controllers are designed for the nominal plant: (i) PID controller: Gc (s) =  0.225s2 + 0.535s + 34.3 s  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  698  CHAPTER 12  Robust Control Systems  (ii) PI controller: Gc (s) =  0.9s + 22.5 s  (iii) PD controller: Gc (s) = 0.9s + 22.5 The step responses for each controller is shown in Figure DP12.6. The responses for the PID and PI controller are the same since the gains were selected to obtain the same ITAE characteristic equation. An appropriate prefilter is used in all cases. (b) off-nominal plant 1.2  1  1  0.8  0.8  y(t)  y(t)  (a) nominal plant 1.2  0.6  0.6  0.4  0.4  0.2  0.2  0 0  0.5  1  1.5  Time (sec)  2  0 0  0.5  1  1.5  2  Time (sec)  FIGURE DP12.6 (a) Step response for nominal plant: PID (solid line); PI (dashed line); and PD (dotted line); (b) for off-nominal plant: PID (solid line); PI (dashed line); and PD (dotted line).  DP12.7  The loop transfer function is G(s) =  Ka Km K = (0.5s + 1)(τf s + 1)s(s + 1) s(s + 2)(s + 1)  since τf is negligible. A suitable PID controller is Gc (s) =  300(s2 + 2.236s + 2.5) KKD (s2 + as + b) = . s s  The step response is shown in Figure DP12.7. The percent overshoot is  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  699  Design Problems  P.O. = 4.6% and the settling time is Ts = 3.74 seconds.  1.2  1  Amplitude  0.8  0.6  0.4  0.2  0 0  0.5  1  1.5  2  2.5  3  3.5  4  4.5  Time (secs)  FIGURE DP12.7 Step response for the elevator position control.  DP12.8  The system transfer function is Y (s) =    G(s)Gc (s)Gp (s) R(s) . 1 + G(s)Gc (s)   We are given G(s) = e−sT  where T = 1 second .  Using a second-order Pade approximation yields G(s) ≈  s2 − 6s + 12 . s2 + 6s + 12  Three controllers that meet the specifications are 0.5 (Integral controller) s 0.04s + 0.4 Gc2 (s) = (PI controller) s 0.01s2 + 0.04s + 0.4 Gc3 (s) = (PID controller) . s Gc1 (s) =  5  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  700  CHAPTER 12  Robust Control Systems  In all cases, the steady-state error is zero. Integral  PI  PID  P.O.(%)  4.05  0  0  Ts (sec)  6.03  6.12  6.02  Tp (sec)  4.75  N/A N/A  |V (t)|max (volts)  1.04  1  1  The prefilter Gp (s) = 1 is used in all designs. To compute the voltage, the transfer function is V (s) =  DP12.9  Gp (s)Gc (s) R(s) . 1 + Gc (s)G(s)  The space robot transfer function is G(s) =  1 . s(s + 10)  (a) Consider Gc (s) = K. Then T (s) =  Gc (s)G(s) K = 2 . 1 + Gc (s)G(s) s + 10s + K  We determine that K = 50.73 for ζ = 0.702. Thus, we expect P.O. < 4.5%. So, Gc (s) = 50.73 . (b) Consider the PD controller Gc (s) = KP + KD s . Then T (s) =  KP + KD s . s2 + (10 + KD )s + KP  Using the ITAE method, we compute KP = 100  and KD = 4 .  Thus, Gc (s) = 4s + 100 ,  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  701  Design Problems  and the prefilter is Gp (s) =  100 . 4s + 100  (c) Consider the PI controller Gc (s) = KP +  KP s + KI KI = . s s  Then, T (s) =  s3  KP s + KI . + 10s2 + KP s + KI  Using the ITAE method, we have ωn = 5.7  KP = 70.2  and KI = 186.59 .  Thus, Gc (s) = 70.2 + 186.59/s , and the prefilter is Gp (s) =  186.59 . 70.2s + 186.59  (d) Consider the PID controller Gc (s) =  KD s2 + KP s + KI . s  Then, T (s) =  KD s2 + KP s + KI . s3 + 10s2 + KD s2 + KP s + KI  Using the ITAE method with ωn = 10, we have KD = 7.5  KP = 215  and KI = 1000 .  Thus, Gc (s) =  7.5s2 + 215s + 1000 , s  and the prefilter is Gp (s) =  7.5s2  1000 . + 215s + 1000  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  702  CHAPTER 12  Robust Control Systems  A summary of the performance is given in Table DP12.9.  Gc (s)  P.O.  tp  ts  yss  max|y(t)|  K  4.5%  0.62 s  0.84 s  0  0.026  PD  5.2%  0.39 s  0.56s  0  0.010  PI  1.98%  0.81 s  1.32s  0  0.013  PID  1.98%  0.46 s  0.75 s  0  0.004  TABLE DP12.9  A summary of performance to a disturbance input.  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  703  Computer Problems  Computer Problems The closed-loop transfer function is T (s) =  s2  8K , + 2s + 8K  T , is and the sensitivity function, SK  S(s) =  s2 + s . s2 + 2s + 8K  The plot of T (s) and S(s) is shown in Figure CP12.1, where K = 10. nt=[80]; dt=[1 2 80]; syst = tf(nt,dt); ns=[1 2 0];ds=[1 2 80]; syss = tf(ns,ds); w=logspace(-1,2,400); [magt,phaset]=bode(syst,w);magtdB(1,:) = 20*log10(magt(1,1,:)); [mags,phases]=bode(syss,w); magsdB(1,:) = 20*log10(mags(1,1,:)); semilogx(w,magtdB,w,magsdB,'--') legend('20log|T|','20log|S|') xlabel('Frequency (rad/sec)') ylabel('Gain dB') grid 20 20log|T| 20log|S| 10  0  −10 Gain dB  CP12.1  −20  −30  −40  −50  −60 −1 10  0  1  10  10 Frequency (rad/sec)  FIGURE CP12.1 Plot of T (s) and the sensitivity function S(s).  2  10  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  704  CHAPTER 12  CP12.2  Robust Control Systems  A reasonable value of the gain K = 4. The family of step responses is shown in Figure CP12.2.  p=[0.5:0.5:20]; K=4; t=[0:0.01:1]; for i=1:length(p) n=[K*p(i)]; d=[1 p(i)]; sys = tf(n,d); sys_cl = feedback(sys,[1]); y=step(sys_cl,t); Y(:,i)=y; [y2,t2]=step(sys_cl); S=stepinfo(y2,t2); Ts(i)=S.SettlingTime; end plot(t,Y) , xlabel('Time (sec)'), ylabel('Step response')  0.9 0.8 0.7  Step response  0.6 0.5 0.4 0.3 0.2 0.1 0  0  0.1  0.2  0.3  0.4  0.5 0.6 Time (sec)  0.7  FIGURE CP12.2 Family of step responses for 0.5 < p < 20.  CP12.3  The closed-loop characteristic equation is 1 + KD  s2 + as + b =0 Js3  where a = KP /KD b = KI /KD .  0.8  0.9  1  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  705  Computer Problems  We select a = 1 and  b=2  to move the root locus into the left hand-plane (see Figure CP12.3a). Then, we choose KD = 71 from the root locus using the rlocfind function. The closed-loop Bode plot in Figure CP12.3b verifies that the bandwidth ωB < 5 rad/sec. Also, the phase margin is P.M. = 45.7o , which meets the design specification. The plot of phase margin versus J is shown in Figure CP12.3c. We see that as J increases, the phase margin decreases. J=25; a=1; b=2; ng=[1];dg=[J 0 0]; sysg=tf(ng,dg); nc=[1 a b]; dc=[1 0]; sysc=tf(nc,dc); sys=series(sysc,sysg); rlocus(sys) Root Locus 2 1.5  Imaginary Axis  1 0.5 0 −0.5 −1 −1.5 −2 −1.2  −1  −0.8  −0.6  FIGURE CP12.3 2 +s+2 (a) Root locus for 1 + KD s 10s = 0. 3  −0.4 Real Axis  −0.2  0  0.2  0.4  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  CHAPTER 12  Robust Control Systems  J=25; a=1; b=2; KD=71; KP=a*KD; KI=b*KD; ng=[1]; dg=[J 0 0]; sysg=tf(ng,dg); nc=[KD KP KI]; dc=[1 0]; sysc = tf(nc,dc); sys=series(sysc,sysg); sys_cl = feedback(sys,[1]); bode(sys_cl); [GM,PM]=margin(sys); PM  PM = 45.7093  20  Magnitude (dB)  10 0 −10 −20 −30  XY 10  0  1  10  2  10  10  Frequency (rad/sec)  FIGURE CP12.3 CONTINUED: (b) Closed-loop Bode plot with ωB < 5 rad/sec.  Ji=[10:1:40]; for i=1:length(Ji) numc=[KD KP KI]; denc=[Ji(i) 0 0 0]; sysc = tf(numc,denc); [gm,pm]=margin(sysc); Pm(i)=pm; end plot(Ji,Pm), grid xlabel('J'), ylabel('Phase Margin (deg)')  90 80 70 60 Phase Margin (deg)  706  50 40 30 20 10 0 −10  0  5  10  15  FIGURE CP12.3 CONTINUED: (c) Phase margin versus J.  20 J  25  30  35  40  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  707  Computer Problems  The closed-loop characteristic equation is  CP12.4  1+K  s2  1 =0 + bs + a  where a = 8 and the nominal value of b = 4. The root locus is shown in Figure CP12.4a.  clf, hold off a=8; b=4; num=[1]; den=[1 b a]; sys = tf(num,den); rlocus(sys), hold on zeta=0.59; wn=1.35; x=[-10:0.1:-zeta*wn]; y=-(sqrt(1-zeta^2)/zeta)*x; xc=[-10:0.1:-zeta*wn];c=sqrt(wn^2-xc.^2); plot(x,y,':',x,-y,':',xc,c,':',xc,-c,':') rlocfind(sys)  ÈSelect a point in the graphics window selected_point = -2.0165 + 2.5426i  K  ans = 2.4659  4 3 +  2  x  Imag Axis  1 0 -1 -2  x +  -3 -4 -4  -3  -2  -1  0  1  2  3  4  Real Axis  FIGURE CP12.4 1 (a) Root locus for 1 + K s2 +4s+8 .  The performance region is specified by ζ = 0.59  and  ωn = 1.35 ,  which derives from the design specifications Ts < 5 sec and P.O. < 10% .  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  708  CHAPTER 12  Robust Control Systems  Using an m-file, the value of K = 2.5 is selected with the rlocfind function. The step responses for b = 0, 1, 4 and b = 40 are shown in Figure CP12.4b. When b = 0, the system is marginally stable; b = 1 results in a stable system with unsatisfactory performance. The nominal case b = 4 is stable and all performance specs are satisfied. When b = 40, the system is heavily damped: the percent overshoot specification is satisfied, but the settling time is too long.  0.5  b=0  0.45 0.4  b=1  Amplitude  0.35 0.3 0.25 0.2 0.15  b=4  0.1 0.05 0 0  b=40 1  2  3  4  5  6  7  8  9  10  Time (secs)  FIGURE CP12.4 CONTINUED: (b) Step responses for b = 0, 1, 4 and 40.  CP12.5  (a) An acceptable lead compensator (designed with root locus methods) is Gc (s) = K  s+a s + 0.3 =5 . s+b s+2  The compensated root locus is shown in Figure CP12.5a, where K=5 is selected to place the closed-loop poles in the performance region. (b) The step responses for ζ = 0, 0.005, 0.1 and 1 are shown in Figure CP12.5b.  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  709  Computer Problems  4 3 +  2 +  x o  Imag Axis  1 0  +o  x  x  -1 +  -2  o x  +  -3 -4 -4  -3  -2  -1  0  1  2  3  4  Real Axis  FIGURE CP12.5 (a) Compensated root locus.  (c) You would like the actual structural damping to be greater than the design value, if it must be different at all. zeta=0,0.005 (solid); zeta=0.1 (dashed); zeta=1 (dotted) 1.4  1.2  Amplitude  1  0.8  0.6  0.4  0.2  0 0  2  4  6  8  10  12  14  Time (sec)  FIGURE CP12.5 CONTINUED: (b) Step responses for ζ = 0, 0.005, 0.1 and 1.  16  18  20  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  710  CHAPTER 12  CP12.6  Robust Control Systems  The m-file script which computes the phase margin as a function of the time delay (using the pade function) is shown in Figure CP12.6. The maximum time delay (for stability) is td = 4.3 seconds. K=5; numg=K*[1]; deng=[1 10 2]; sysg = tf(numg,deng); time delay vector td=[0:0.1:5]; for i=1:length(td) [ndelay,ddelay]=pade(td(i),2); sysd = tf(ndelay,ddelay); sys = series(sysg,sysd); [mag,phase,w]=bode(sys); [gm,pm,w1,w2]=margin(mag,phase,w); pmv(i)=pm; end plot(td,pmv), grid xlabel('time delay [sec]') ylabel('phase margin [deg]')  120  100  phase margin [deg]  80  60  40  20  0  -20  0  0.5  1  1.5  2  2.5  3  3.5  4  4.5  5  time delay [sec]  FIGURE CP12.6 Phase margin versus time delay.  CP12.7  The m-file script is shown in Figure CP12.7a. The steady-state error (shown in Figure CP12.7b) is zero when a = 0.5 and increases rapidly as a increases past a = 0.5. The maximum initial undershoot is shown in Figure CP12.7c. As a increases, the initial undershoot increases linearly. The gain margin is shown in Figure CP12.7d. It  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  711  Computer Problems  can be seen that as a increases, the gain margin decreases very rapidly. a=[0.01:0.01:0.99]; t=[0:0.1:30]; for i=1:length(a) num=a(i)*[1 -0.5]; den=[1 2 1]; sys_o = tf(num,den); [mag,phase,w]=bode(sys_o); [gm,pm,w1,w2]=margin(mag,phase,w); gain margin gmv(i)=gm; sys_cl = feedback(sys_o,[1]); [y,x]=step(-sys_cl,t); negative unit step input yf(i)=1-y(length(t)); steady-state tracking error ym(i)=-min(y)*100; max initial undershoot end figure(1), plot(a,gmv), grid, xlabel('a'), ylabel('gm') figure(2), plot(a,yf ), grid, xlabel('a'), ylabel('steady-state error') figure(3), plot(a,ym), grid, xlabel('a'), ylabel('maximum initial undershoot [%]')  FIGURE CP12.7 Script to generate all the plots.  1 0.9 0.8  steady−state error  0.7 0.6 0.5 0.4 0.3 0.2 0.1 0  0  0.1  0.2  0.3  0.4  0.5 a  FIGURE CP12.7 CONTINUED: (b) Steady-state tracking error.  0.6  0.7  0.8  0.9  1  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  CHAPTER 12  Robust Control Systems  25  maximum initial undershoot [%]  20  15  10  5  0  0  0.1  0.2  0.3  0.4  0.5 a  0.6  0.7  0.8  0.9  1  0.6  0.7  0.8  0.9  1  FIGURE CP12.7 CONTINUED: (c) Maximum initial undershoot.  250  200  150 gm  712  100  50  0  0  0.1  0.2  FIGURE CP12.7 CONTINUED: (d) Gain margin.  0.3  0.4  0.5 a  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  713  Computer Problems  CP12.8  The plant (balloon and canister dynamics plus motor) is given by G(s) =  1 , (s + 2)(s + 4)(s + 10)  and the PID controller is Gc (s) =  KD (s2 + as + b) . s  Let a = 6. Then using the root locus methods, we determine that with KD = 12.5  and  b = 10  we have the roots s1 = −8.4 s2 = −4.7 s3,4 = −1.43 ± j1.05 . Thus, ζ = 0.8. The plot of y(t) is shown in Figure CP12.8. The percent overshoot is less that 3%, as desired.  1.4  1.2  With prefilter 1  y(t)  0.8  Without prefilter 0.6  0.4  0.2  0  0  FIGURE CP12.8 Simulation of the GRID device.  0.5  1  1.5  2  2.5 Time (sec)  3  3.5  4  4.5  5  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  C H A P T E R  1 3  Digital Control Systems  Exercises E13.1  (a) Elevation contours on a map are discrete signals. (b) Temperature in a room is a continuous signal. (c) A digital clock display is a discrete signal. (d) The score of a basketball game is a discrete signal. (e) The output of a loudspeaker is a continuous signal.  E13.2  (a) Using long-division we determine that Y (z) = z −1 + 3z −2 + 7z −3 + 15z −4 + · · · Therefore, with Y (z) =  ∞ X  y(kT )z −k  k=0  we have y(0) = 0  y(T ) = 1  y(2T ) = 3 y(3T ) = 7 y(4T ) = 15 .  (b) The exact solution is y(kT ) = ek ln 2 − 1 . E13.3  For the system response y(kT ) = kT where k ≥ 0, we have Y (z) =  E13.4  The partial fraction expansion of Y (s) is Y (s) =  714  Tz . (z − 1)2  5 0.25 0.0625 0.3125 = + − . s(s + 2)(s + 10) s s + 10 s+2  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  715  Exercises  Then, using Table 13.1 in Dorf and Bishop, we determine that z z z + 0.0625 − 0.3125 −10T z −1 z −e z − e−2T z z z = 0.25 + 0.0625 − 0.3125 , z −1 z − 0.135 z − 0.670  Y (z) = 0.25  where T = 0.1. E13.5  The Space Shuttle and robot arm control block diagram is shown in Figure E13.5. The human operator uses information from the computer generated data display and visual sensory data from the TV monitor and by looking out the window. He/she commands the robot arm via a joystick command to the computer.  data display  measurement digital  analog  A/D  joint angle & rate sensors  digital human operator joystick command  ref.  +  Computer  digital  analog  D/A  -  Robot arm & motors/gears  tip position  measurement  TV monitor & window view  FIGURE E13.5 The Space Shuttle/robot arm control block diagram.  E13.6  From Section 10.8 in Dorf and Bishop, we find that the design resulted in the compensator Gc (s) =  6.66s + 1 s + 0.15 = 0.1 . 66.6s + 1 s + 0.015  Using the relationships A = e−aT , we compute  B = e−bT ,  and C  1−A a =K , 1−B b  A = e−0.15(0.001) = 0.99985 , B = e−0.015(0.001) = 0.999985 , and C = 0.1 .  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  716  CHAPTER 13  Digital Control Systems  Therefore, D(z) = C E13.7  z − 0.99985 z−A = 0.1 . z−B z − 0.999985  Using long-division, we determine that Y (z) = 1 + 3.5z −1 + 5.75z −2 + 6.875z −3 + · · · Therefore, with Y (z) =  ∞ X  y(kT )z −k  k=0  we have y(0) = 1 y(T ) = 3.5 E13.8  y(2T ) = 5.75  y(3T ) = 6.875 .  The closed-loop system with T (z) =  z2  z + 0.2z − 1.0  is unstable since one of the poles of the transfer function (z = −1.1 and z = 0.90) lies outside the unit circle in the z-plane. E13.9  (a) Using long-division we determine that Y (z) = z −1 + z −2 + z −3 + z −4 + · · · Therefore, with Y (z) =  ∞ X  y(kT )z −k  k=0  we have y(0) = 0 y(T ) = 1 y(2T ) = 1  y(3T ) = 1 y(4T ) = 1 .  (b) The exact solution is y(kT ) = 1 − δ(k) where δ(k) = 1 when k = 0 and δ(k) = 0 when k 6= 0. E13.10  We compute T /τ = 1.25. (a) Using Figure 13.19 in Dorf and Bishop, we determine that Kτ = 0.8 which implies K = 100. (b) Using Figure 13.21 in Dorf and Bishop, we determine that ess = 0.75.  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  717  Exercises  (c) Using Figure 13.20 in Dorf and Bishop, we determine that Kτ = 0.7 which implies K = 88. (a) The transfer function (including the zero-order hold) is Go (s)Gp (s) =  100(1 − e−sT ) . s(s2 + 100)  Expanding into partial fractions yields s 1 − 2 G(z) = (1 − z )Z s s + 100   z z(z − cos 10T ) −1 = (1 − z ) − . z − 1 z 2 − 2 cos 10T z + 1     −1  When T = 0.05 we ha,ve G(z) =  0.1224(z + 1) . − 1.7552z + 1  z2  (b) The system is marginally stable since the system poles, z = −0.8776± 0.4794j, are on the unit circle. (c) The impulse response and sinusoidal input response are shown in Figure E13.11.  Amplitude  0.5  0  -0.5  2  0  4  6  8  10  12  14  16  No. of Samples 40  Amplitude  E13.11  20 0 -20 -40 0  10  20  30  40  50  60  70  No. of Samples  FIGURE E13.11 Impulse and sinusoidal (natural frequency) input response.  80  90  100  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  718  CHAPTER 13  E13.12  Digital Control Systems  The partial fraction expansion of X(s) is X(s) =  s2  s+1 1 2 = − . + 5s + 6 s+3 s+2  Then, with T = 1, we have X(z) = E13.13  z 2z z 2z − = − . −3 −2 z−e z−e z − 0.0498 z − 1353  The root locus is shown in Figure E13.13. For stability: 2.2 < K < 5.8. Root Locus 2  1.5  Imaginary Axis  1  0.5 K=5.8 0 K=2.2 −0.5  −1  −1.5  −2 −2  −1.5  −1  −0.5  0 Real Axis  0.5  1  1.5  2  FIGURE E13.13 Root locus with unit circle (dashed curve).  E13.14  Given Gp (s), we determine that (with K = 5) G(z) =  5(1 − e−1 )z . z(z − e−1 )  The closed-loop characteristic equation is z 2 + 1.792z + 0.368 = 0 and the system is unstable, since there is a pole at z = −1.55. The system is stable for 0 < K < 4.32 .  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  719  Exercises  E13.15  The transfer function G(z) is G(z) =  z2  The sampling time is T = 1 s. E13.16  0.1289z + 0.02624 . − 0.3862z + 0.006738  The transfer function G(z) is G(z) =  0.2759z + 0.1982 . − 1.368z + 0.3679  z2  The sampling time is T = 0.5 s.  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  720  CHAPTER 13  Digital Control Systems  Problems P13.1  The plot of the input to the sampler and the output r ∗ (t) is shown in Figure P13.1.  1 0.8 0.6 0.4  r(t), r*(t)  0.2 0 -0.2 -0.4 -0.6 -0.8 -1 0  0.2  0.4  0.6  0.8  1  1.2  1.4  1.6  1.8  2  Time (sec)  FIGURE P13.1 Plot of r(t) = sin(ωt) and r ∗ (t).  The plot of the input and the output is shown in Figure P13.2.  1 0.8 0.6 0.4 0.2  r(t)  P13.2  0 -0.2 -0.4 -0.6 -0.8 -1 0  0.2  0.4  0.6  0.8  1  1.2  Time (sec)  FIGURE P13.2 Plot of r(t) = sin(ωt) and output of sample and hold.  1.4  1.6  1.8  2  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  721  Problems  P13.3  The transfer function Y (z)/R∗ (z) = G(z) =  z . z − e−T  The ramp input is represented by R(z) =  Tz . (z − 1)2  The output Y (z) = G(z)R(z) is obtained by long division: h  i  Y (z) = T z −1 + T (2 + e−T )z −2 − T (1 + 2e−T ) − (2 + e−T )2 z −3 h  + T e−T + (1 + 2e−T )(2 + e−T )   − (2 + e−T ) (1 + 2e−T ) − (2 + e−T )2 P13.4  i  z −4 + · · ·  The transfer function Y (s)/R∗ (s) =  1 − e−sT . s(s + 2)  The partial fraction expansion (with T = 1) yields G(z) = (1 − z −1 )Z = P13.5    0.5 0.5 − s s+2  0.4323 . z − 0.1353    = (1 − z −1 )    0.5z 0.5z − z − 1 z − 0.1353    The step input is R(z) =  z . z−1  Also, T (z) =  G(z) 0.6321 = . 1 + G(z) z + 0.2643  So, Y (z) = T (z)R(z) =  0.6321 z 0.6321z = 2 . z + 0.2643 z − 1 z − 0.7357z − 0.2643  Using long-division we determine that Y (z) = 0.6321z −1 + 0.4650z −2 + 0.5092z −3 + 0.4975z −4 + 0.5006z −5 + · · ·  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  722  CHAPTER 13  Digital Control Systems  Therefore, with Y (z) =  ∞ X  y(kT )z −k  k=0  we have y(0) = 0, y(T ) = 0.6321, y(2T ) = 0.4650, y(3T ) = 0.5092, y(4T ) = 0.4975, and y(5T ) = 0.5006. P13.6  Using the final value theorem (see Table 13.1 in Dorf and Bishop), we determine that (for a step input) Yss = lim (z − 1)Y (z) = lim (z − 1) z→1  z→1  0.6321 z 0.6321 = = 0.5 . z + 0.2643 z − 1 1.2643  And using the initial value theorem, we compute Yo = lim Y (z) = lim z→∞  z→∞  0.6321 z =0. z + 0.2643 z − 1  P13.7  Using Figures 13.19 and 13.21 in Dorf and Bishop, we determine that the performance specifications are satisfied when Kτ = 0.5 and Tτ = 2. Computing K and T (with τ = 0.5) yields K = 1 and T = 1.  P13.8  We can select K = 1 and r = 0.2. The step responses for the compensated and uncompensated systems are shown in Figure P13.8.  1.2  1 Uncompensated  Amplitude  0.8  0.6  Compensated  0.4  0.2  0 0  10  20  30 40 No. of Samples  FIGURE P13.8 Plot of compensated and uncompensated systems.  50  60  70  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  723  Problems  P13.9  Consider the compensator Gc (s) = K  s+a . s+b  Then, using Bode methods we can select a = 1, b = 4, and K = 1. The compensated system phase margin is P.M. = 50o and the gain margin is G.M. = 15dB. The crossover frequency is ωc = 2.15 rad/sec. Utilizing the Gc (s)-to-D(z) method and selecting T = 0.01 second, we determine D(z) = C We use the relationships A = e−aT ,  z − 0.99 z−A = . z−B z − 0.96  B = e−bT ,  and C  1−A a =K , 1−B b  to compute A = e−0.01 = 0.99, B = e−0.04 = 0.96, and C = 1. P13.10  (a) The transfer function G(z)D(z) is G(z)D(z) = K  0.0037z + 0.0026 . − 1.368z + 0.3679  z2  (b) The closed-loop system characteristic equation is 1+K  0.0037z + 0.0026 =0. − 1.368z + 0.3679  z2  (c) Using root locus methods, the maximum value of K is found to be Kmax = 239. (d) Using Figure 13.19 in Dorf and Bishop for T /τ = 1 and a maximum overshoot of 0.3, we find that K = 75. (e) The closed-loop transfer function (with K = 75) is T (z) =  0.2759z + 0.1982 . − 1.092z + 0.5661  z2  The step response is shown in Figure P13.10. (f) The closed-loop poles with K = 119.5 are z = 0.4641 ± 0.6843j. The overshoot is 0.55. (g) The step response is shown in Figure P13.10 (for K = 119.5).  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  724  CHAPTER 13  Digital Control Systems  K=75  Amplitude  1.5  1  0.5  0  2  0  4  6  8  10  12  14  16  No. of Samples K=119.5  Amplitude  2 1.5 1 0.5 0  0  2  4  6  8  10  12  14  16  18  20  No. of Samples  FIGURE P13.10 Step response for K = 75 and K = 119.5.  P13.11  (a) Consider the compensator Gc (s) = K  s+a . s+b  Then, using Bode methods we can select a = 0.7, b = 0.1, and K = 150. The compensated system overshoot and steady-state tracking error (for a ramp input) are P.O. = 30% and ess < 0.01. (b) Utilizing the Gc (s)-to-D(z) method (with T = 0.1 second), we determine D(z) = C We use the relationships A = e−aT ,  z−A z − 0.9324 = 155.3 . z−B z − 0.99  B = e−bT ,  to compute A = e−0.007 = 0.9324 ,  and C  1−A a =K , 1−B b  B = e−0.01 = 0.99 ,  and C = 155.3 .  (c) The step response for the continuous system with Gc (s) in part(a) and for the discrete system with D(z) in part (b) is shown in Figure P13.11a. (d) Utilizing the Gc (s)-to-D(z) method (with T = 0.01 second), we de-  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  725  Problems  T=0.1 sec 1.8 1.6 1.4  Amplitude  1.2 1 0.8 0.6 0.4 0.2 0 0  2  4  6  8  10  12  14  16  18  20  No. of Samples  FIGURE P13.11 (a) Step response for continuous and discrete systems (T=0.1s) in Parts (a) and (b).  termine D(z) = C We use the relationships  to compute  z − 0.993 z−A = 150 . z−B z − 0.999  A = e−aT B = e−bT 1−A a C =K 1−B b A = e−0.07 = 0.993 B = e−0.001 = 0.999 C = 150 .  The step response for the continuous system with Gc (s) in and for the discrete system with D(z) in part (d) is shown ure P13.11b. (e) The ramp response for the continuous system with Gc (s) in and for the discrete system with D(z) in part (b) is shown ure P13.11c.  part(a) in Figpart(a) in Fig-  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  CHAPTER 13  Digital Control Systems T=0.01 sec 1.4  1.2  Amplitude  1  0.8  0.6  0.4  0.2  0  0  20  40  60  80  100  120  140  160  180  200  No. of Samples  FIGURE P13.11 CONTINUED: (b) Step response for continuous and discrete systems (T=0.01s) in Parts (a) and (d).  T=0.1 sec 2 1.8 1.6 1.4  Amplitude  726  1.2 1 0.8  Ramp input (dashed line)  0.6 0.4 0.2 0  0  2  4  6  8  10  12  14  16  18  20  No. of Samples  FIGURE P13.11 CONTINUED: (c) Ramp response for continuous and discrete systems (T=0.1s) in Parts (a) and (b).  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  727  Problems  P13.12  The root locus is shown in Figure P13.12. For stability: 0 < K < 2. 2  1.5 Unit circle (dashed line)  1  Imaginary Axis  0.5  0  −0.5  −1  −1.5  −2 −2  −1.5  −1  −0.5  0 Real Axis  0.5  1  1.5  2  FIGURE P13.12 z+0.5 = 0. Root locus for 1 + K z(z−1)  The root locus is shown in Figure P13.13. When K = 0.027, the characteristic equation has two equal roots: z1,2 = 0.7247 and z3 = 0.2593.  2 Unit circle (dashed line)  1.5 1 0.5  Imag Axis  P13.13  0  o  o  xx  x  -0.5 -1 -1.5 -2 -2  -1.5  -1  -0.5  0  0.5  Real Axis  FIGURE P13.13 z 2 +1.1206z−0.0364 Root locus for 1 + K z 3 −1.7358z 2 +0.8711z−0.1353 = 0.  1  1.5  2  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  728  CHAPTER 13  P13.14  Digital Control Systems  The root locus is shown in Figure P13.14. When K = 9.5655 × 10−5 , the two real roots break away from the real axis at z = 0.99. For stability: K < 9.7 × 10−5 . 2 Unit circle (dashed line)  1.5 1  Imag Axis  0.5 x  0  o  o  x  x x  -0.5 -1 -1.5 -2 -2  -1.5  -1  -0.5  0  0.5  1  1.5  2  Real Axis  FIGURE P13.14 z 3 +10.3614z 2 +9.758z+0.8353 Root locus for 1 + K z 4 −3.7123z 3 +5.1644z 2 −3.195z+0.7408 = 0.  P13.15  Given Gp (s) =  20 s−5  and the sample and hold (T=0.1s) as shown in Figure 13.18 in Dorf and Bishop, we determine that G(z) =  2.595 . z − 1.649  Then, with R(z) = z/(z − 1), we have Y (z) =  2.595z . (z − 1)(z + 0.9462)  Therefore, Y (z) = 2.59z −1 + 0.14z −2 + 2.46z −3 + 0.26z −4 + · · ·.  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  729  Problems  P13.16  Given Gp (s) and the sample and hold (T=1s) as shown in Figure 13.18 in Dorf and Bishop, we determine that 0.22775z + 0.088984 . − 1.0498z + 0.049787  G(z) =  z2  Then, with R(z) = z/(z − 1), we have Y (z) =  0.22775z + 0.088984 z . − 0.82203z + 0.13877 z − 1  z2  The plot of y(kT ) is shown in Figure P13.16.  1 0.9 0.8 0.7  y(kT)  0.6 0.5 0.4 0.3 0.2 0.1 0  1  2  3  4  5  6  7  kT  FIGURE P13.16 Plot of y(kT ) for a step input.  P13.17  The root locus is shown in Figure P13.17 for 1+K  0.39532z + 0.30819 =0. − 1.4724z + 0.47237  z2  The limiting value of the gain for stability is K = 1.71.  8  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  730  CHAPTER 13  Digital Control Systems  Root Locus 2 1.5  Imaginary Axis  1 0.5 0 −0.5 −1 Unit circle (dashed line)  −1.5 −2 −5  −4  −3  −2 −1 Real Axis  0  1  2  FIGURE P13.17 = 0. Root locus for 1 + K z 20.39532z+0.30819 −1.4724z+0.47237  P13.18  The plot of the step responses for 0 ≤ T ≤ 1.2 is shown in Figure P13.18. The overshoot and settling time summary is given in Table P13.18.  T  P.O.  Ts  0  0.2  0.4  0.6  0.8  1.0  1.2  16.3%  20.6%  25.6%  31.3%  36.9%  40.0%  51.0%  8.1  8.4  8.8  11.4  14.4  16.0  19.2  TABLE P13.18  Performance summary.  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  731  Problems 1.6  1.4  1.2  Amplitude  1  0.8  0.6  0.4  0.2  0 0  10  20  FIGURE P13.18 Step responses for 0 ≤ T ≤ 1.2.  30  40 50 60 No. of Samples  70  80  90  100  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  732  CHAPTER 13  Digital Control Systems  Advanced Problems AP13.1  Given the sample and hold with Gp (s), we determine that G(z) =  10.5K(z − 0.9048) . (z − 1)2  The root locus is shown in Figure AP13.1. For stability: 0 < K < 0.2. 2 Unit circle (dashed line)  1.5 1  Imag Axis  0.5 0  ox  -0.5 -1 -1.5 -2 -2  -1.5  -1  -0.5  0  0.5  1  1.5  2  Real Axis  FIGURE AP13.1 10.5(z−0.9048) = 0 with unit circle (dashed line). Root locus for 1 + K (z−1)2  The root locus is shown in Figure AP13.2a. The loop transfer function is  2 Unit circle (dashed line)  1.5 1 0.5  Imag Axis  AP13.2  0  o  x  x  -0.5 -1 -1.5 -2 -2  -1.5  -1  -0.5  0 Real Axis  FIGURE AP13.2 0.0379z = 0. (a) Root locus for 1 + K (z−1)(z−0.368)  0.5  1  1.5  2  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  733  Advanced Problems  G(z)D(z) = K  0.0379z . (z − 1)(z − 0.368)  For stability: Kmax = 72. We select K = 8.2. The step response is shown in Figure AP13.2b. 1.2  1  Amplitude  0.8  0.6  0.4  0.2  0 0  1  2  3  4  5  6  7  8  9  10  No. of Samples  FIGURE AP13.2 CONTINUED: (b) Step response with K = 8.2.  The root locus is shown in Figure AP13.3a.  The maximum gain for  Root Locus 2 Unit circle (dashed line)  1.5  System: sysz Gain: 5.99 Pole: 0.736 + 0.257i Damping: 0.596 Overshoot (%): 9.74 Frequency (rad/sec): 8.36  1 Imaginary Axis  AP13.3  0.5 0 −0.5 −1 −1.5 −2 −3  −2.5  −2  −1.5  −1  −0.5 Real Axis  FIGURE AP13.3 (a) Root locus for 1 + K z0.07441z+0.06095 2 −1.474z+0.6098 = 0.  0  0.5  1  1.5  2  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  734  CHAPTER 13  Digital Control Systems  stability is Kmax = 44.3. We select K = 6. The step response is shown in Figure AP13.3b.  Step Response 1.4  1.2  Amplitude  1  0.8  0.6  0.4  0.2  0  0  0.2  0.4  0.6 Time (sec)  0.8  1  1.2  FIGURE AP13.3 CONTINUED: (b) Step response with K = 6.  AP13.4  The loop transfer function is G(z) =  10(1 − e−T ) , z − e−T  and the closed-loop transfer function is T (z) =  10(1 − e−T ) . z − (11e−T − 10)  For stability, we require |11e−T − 10| < 1 . Solving for T yields 0 < T < 0.2 . Selecting T = 0.1s provides a stable system with rapid response; the settling time is Ts = 0.2s. The step response is shown in Figure AP13.4.  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  735  Advanced Problems  1 0.9 0.8  Amplitude  0.7 0.6 0.5 0.4 0.3 0.2 0.1 0  0  0.5  1  1.5  2  2.5  3  3.5  4  No. of Samples  FIGURE AP13.4 Step response with T = 0.1s.  The maximum gain for stability is Kmax = 63.15. Root Locus 2  Unit circle (dashed line)  1.5  1  Imaginary Axis  AP13.5  0.5  System: sysz Gain: 63.2 Pole: 0.725 − 0.686i Damping: 0.00308 Overshoot (%): 99 Frequency (rad/sec): 7.58  0  −0.5  −1  −1.5  −2 −3  −2.5  −2  −1.5  −1  −0.5 Real Axis  FIGURE AP13.5 Root locus for 1 + K 0.004535z+0.004104 z 2 −1.741z+0.7408 = 0.  0  0.5  1  1.5  2  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  736  CHAPTER 13  Digital Control Systems  Design Problems The plant model with parameters given in Table CDP2.1 in Dorf and Bishop is given by: 26.035 , s(s + 33.142)  Gp (s) =  where we neglect the motor inductance Lm and where we switch off the tachometer feedback (see Figure CDP4.1 in Dorf and Bishop). Letting     G(z) = Z G≀ (∫ )G√ (∫ ) we obtain G(z) =  1.2875e − 05(z + 0.989) . (z − 1)(z − 0.9674)  A suitable controller is D(z) =  20(z − 0.5) . z + 0.25  The step response is shown below. The settling time is under 250 samples. With each sample being 1 ms this means that Ts < 250 ms, as desired. Also, the percent overshoot is P.O. < 5%.  1.2  1  0.8 Amplitude  CDP13.1  0.6  0.4  0.2  0 0  50  100  150 No. of Samples  200  250  300  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  737  Design Problems  (a) Given the sample and hold with Gp (s), we determine that KG(z) = K  0.1228 . z − 0.8465  The root locus is shown in Figure DP13.1a. For stablity: 0 ≤ K < 15.  Unit circle (dashed line)  1.5  1  0.5  Imag Axis  DP13.1  0  x  -0.5  -1  -1.5 -1.5  -1  -0.5  0  0.5  1  1.5  Real Axis  FIGURE DP13.1 0.1228 (a) Root locus for 1 + K z−0.8465 = 0 with unit circle (dashed line).  (b) A suitable compensator is Gc (s) =  15(s + 0.5) . s+5  Utilizing the Gc (s)-to-D(z) method (with T = 0.5 second), we determine D(z) = C  z−A z − 0.7788 = 6.22 . z−B z − 0.0821  We use the relationships A = e−aT ,  B = e−bT ,  and C  1−A a =K , 1−B b  to compute A = e−0.5(0.5) = 0.7788 ,  B = e−0.5(5) = 0.0821 ,  and C = 6.22 .  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  738  CHAPTER 13  Digital Control Systems  (c) The step response is shown in Figure DP13.1b.  0.8 0.7 0.6  Amplitude  0.5 0.4 0.3 0.2 0.1 0  0  2  4  6  8  10  12  14  16  18  20  No. of Samples  FIGURE DP13.1 CONTINUED: (b) Closed-loop system step response.  DP13.2  With the sample and hold (T=10ms), we have G(z) =  0.00044579z + 0.00044453 . z 2 − 1.9136z + 0.99154  A suitable compensator is D(z) = K  z − 0.75 , z + 0.5  √ where K is determined so that ζ of the system is 1/ 2. The root locus is shown in Figure DP13.2. We choose K = 1400.  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  739  Design Problems  Root Locus 1.5 Curve of constant zeta=0.707 (dashed line) 1  Imaginary Axis  0.5  0  −0.5  −1  −1.5 −5  −4  −3  −2 −1 Real Axis  0  1  2  FIGURE DP13.2 0.00044579z+0.00044453 Root locus for 1 + K z−0.75 = 0. z+0.5 z 2 −1.9136z+0.99154  The root locus is shown in Figure DP13.3a.  2 Curve of constant zeta=0.707 (dashed line) 1.5 1 0.5  Imag Axis  DP13.3  0  o  x  x  0.5  1  -0.5 -1 -1.5 -2 -2  -1.5  -1  -0.5  0  1.5  2  Real Axis  FIGURE DP13.3 z+1 (a) Root locus for 1 + K (z−1)(z−0.5) = 0.  The gain for ζ = 0.707 is K = 0.0627. The step response is shown in Figure DP13.3b. The settling time is Ts = 14T = 1.4s and P.O. = 5%.  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  740  CHAPTER 13  Digital Control Systems  1.2  1  Amplitude  0.8  0.6  0.4  0.2  0 0  2  4  6  8  10  12  14  16  18  No. of Samples  FIGURE DP13.3 CONTINUED: (b) Step response with K = 0.0627.  With the sample and hold (T=1s), we have G(z) =  0.484(z + 0.9672) . (z − 1)(z − 0.9048)  2 Curve of constant zeta=0.5 (dashed line) 1.5 1 0.5  Imag Axis  DP13.4  0  o  x  ox x  -0.5 -1 -1.5 -2 -2  -1.5  -1  -0.5  0  0.5  Real Axis  FIGURE DP13.4 0.484(z+0.9672) (a) Root locus for 1 + K z−0.88 z+0.5 (z−1)(z−0.9048) = 0.  1  1.5  2  20  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  741  Design Problems  1.2  1  Amplitude  0.8  0.6  0.4  0.2  0 0  2  4  6  8  10  12  14  16  No. of Samples  FIGURE DP13.4 CONTINUED: (b) Step response for K = 12.5.  A suitable compensator is D(z) = K  z − 0.88 , z + 0.5  where K is determined so that ζ of the system is 0.5. The root locus is shown in Figure DP13.4a. We choose K = 12.5. The step response is shown in Figure DP13.4b. Also, Kv = 1, so the steady-state error specification is satisfied. DP13.5  Select T = 1 second. With the sample and hold, we have G(z) =  0.2838z + 0.1485 . − 1.135z + 0.1353  z2  The root locus is shown in Figure DP13.5. To meet the percent overshoot specification, we choose K so that ζ of the system is 0.7. This results in K = 1. The step response has an overshoot of P.O. = 4.6%. Also, from Figure 13.21 in Dorf and Bishop, we determine that the steady-state error to a ramp input is ess = 2 (since T /τ = 2, and Kτ = 0.3).  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  742  CHAPTER 13  Digital Control Systems  2 Curve of constant zeta=0.7 (dashed line) 1.5 1  Imag Axis  0.5 0  o  x  x  -0.5 -1 -1.5 -2 -2  -1.5  -1  -0.5  0  0.5  1  1.5  2  Real Axis  FIGURE DP13.5 Root locus for 1 + K z 20.2838z+0.1485 −1.135z+0.1353 = 0.  DP13.6  With the sample and hold at T = 1 , we have G(z) =  z2  Consider the digital controller  0.298z + 0.296 . − 1.98z + 0.9802  Dz) = K  z − 0.9 . z + 0.6  The root locus is shown in Figure DP13.6. To meet the percent overshoot specification, we choose K so that ζ of the system is greater than 0.52. We select K = 2. The step response has an overshoot of P.O. = 11.9% and the settling time is Ts = 17.8s.  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  743  Design Problems  Root Locus 1 0.8 0.6  Imaginary Axis  0.4 0.2 0 −0.2 −0.4 −0.6 −0.8 −1 −5  −4  −3  −2 −1 Real Axis  0  1  2  FIGURE DP13.6 0.298z+0.296 Root locus for 1 + K z−0.9 z+0.6 z 2 −1.98z+0.9802 = 0.  Step Response 1.4  System: syscl Peak amplitude: 1.12 Overshoot (%): 11.9 At time (sec): 2  1.2  System: syscl Settling Time (sec): 17.8  Amplitude  1  0.8  0.6  0.4  0.2  0  0  5  10  15 20 Time (sec)  FIGURE DP13.6 CONTINUED: (b) Step response for K = 2.  25  30  35  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  744  CHAPTER 13  Digital Control Systems  Computer Problems CP13.1  The m-file script and unit step response are shown in Figure CP13.1. num=[0.2145 0.1609]; den=[1 -0.75 0.125]; sysd = tf(num,den,1); step(sysd,0:1:50) 1.2  1  Amplitude  0.8  0.6  0.4  0.2  0 0  5  10  15  20  25  30  35  40  45  50  No. of Samples  FIGURE CP13.1 Step response.  CP13.2  The m-file script utilizing the c2d function is shown in Figure CP13.2.  % Part (a) num = [1]; den = [1 0]; T = 1; sys = tf(num,den); sys_d = c2d(sys,T,'zoh') % % Part (b) num = [1 0]; den = [1 0 2]; T = 1; sys = tf(num,den); sys_d=c2d(sys,T,'zoh')  FIGURE CP13.2 Script utilizing the c2d function for (a) and (b).  Transfer function: 1 ----z-1 Transfer function: 0.6985 z - 0.6985 -----------------z^2 - 0.3119 z + 1  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  745  Computer Problems  % Part (c) num = [1 4]; den = [1 3]; T = 1; sys = tf(num,den); sys_d = c2d(sys,T,'zoh') % % Part (d) num = [1]; den = [1 8 0]; T = 1; sys = tf(num,den); sys_d = c2d(sys,T,'zoh')  Transfer function: z + 0.267 ----------z - 0.04979 Transfer function: 0.1094 z + 0.01558 ------------------z^2 - z + 0.0003355  FIGURE CP13.2 CONTINUED: Script utilizing the c2d function for (c) and (d).  The continuous system transfer function (with T = 0.1 sec) is T (s) =  s2  13.37s + 563.1 . + 6.931s + 567.2  The step response using the dstep function is shown in Figure CP13.3a. The contrinuous system step response is shown in Figure CP13.3b.  1.8 1.6 1.4 1.2  Amplitude  CP13.3  1 0.8 0.6 0.4 0.2 0  0  2  4  6  8  No. of Samples  FIGURE CP13.3 (a) Unit step response using the dstep function.  10  12  14  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  746  CHAPTER 13  Digital Control Systems  1.8 *  1.6 1.4 *  1.2  *  1  *  *  *  *  *  *  *  * *  0.8  *  *  0.6 0.4 0.2 0* 0  0.2  0.4  0.6  0.8  1  1.2  1.4  FIGURE CP13.3 CONTINUED: (b) Continuous system step response (* denote sampled-data step response).  The root locus in shown in Figure CP13.4. For stability: 0 < K < 2.45. Root Locus 2 1.5 1 Imaginary Axis  CP13.4  0.5 0 −0.5 −1 −1.5 −2 −2  −1.5  −1  FIGURE CP13.4 z Root locus for 1 + K z 2 −z+0.45 = 0.  −0.5  0 Real Axis  0.5  1  1.5  2  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  747  Computer Problems  CP13.5  The root locus in shown in Figure CP13.5. For stability: 0 < K < ∞.  1 0.8 0.6  Imag Axis  0.4 0.2 0 -0.2 -0.4 -0.6 -0.8 -1 -1  -0.8  -0.6  -0.4  -0.2  0 0.2 Real Axis  0.4  0.6  0.8  1  FIGURE CP13.5 (z−0.2)(z+1) Root locus for 1 + K (z−0.08)(z−1) = 0  The root locus is shown in Figure CP13.6.  Root Locus 1.5 1  Imaginary Axis  CP13.6  0.5 0 0.5 1 1.5 1.5  1  0.5  0 Real Axis  FIGURE CP13.6 Root locus.  0.5  1  1.5  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  748  CHAPTER 13  Digital Control Systems  We determine the range of K for stability is 0.4 < K < 1.06. % Part (a) num=[1 4 4.25 ]; den=[1 -0.1 -1.5]; sys = tf(num,den); rlocus(sys), hold on xc=[-1:0.1:1];c=sqrt(1-xc.^2); plot(xc,c,':',xc,-c,':') hold off % % Part (b) rlocfind(sys) rlocfind(sys)  ÈSelect a point in the graphics window selected_point = -0.8278 + 0.5202i ans = 0.7444  Kmax  Select a point in the graphics window selected_point = -0.9745 - 0.0072i ans = 0.3481  Kmin  FIGURE CP13.6 CONTINUED: Using the rlocus and rlocfind functions.  Using root locus methods, we determine that an acceptable compensator is Gc (s) = 11.7  s+6 . s + 20  With a zero-order hold and T = 0.02 sec, we find that  1.2  1 *  *  * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * *  * * *  0.8  Amplitude  CP13.7  * * *  0.6 * *  0.4 * *  0.2 * *  0* * 0  0.1  0.2  0.3  0.4  0.5  0.6  0.7  Time (sec)  FIGURE CP13.7 System step response (* denotes sampled-data response).  0.8  0.9  1  © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.  749  Computer Problems  D(z) =  11.7z − 10.54 . z − 0.6703  The closed-loop step response is shown in Figure CP13.7.

Source Exif Data:

File Type                       : PDF File Type Extension             : pdf MIME Type                       : application/pdf PDF Version                     : 1.6 Linearized                      : Yes Page Mode                       : UseOutlines XMP Toolkit                     : 3.1-701 Create Date                     : 2010:08:05 08:48:05-05:00 Creator Tool                    : dvips(k) 5.98 Copyright 2009 Radical Eye Software Modify Date                     : 2010:10:25 08:50:47+02:00 Metadata Date                   : 2010:10:25 08:50:47+02:00 Producer                        : GPL Ghostscript 8.70 Format                          : application/pdf Title                           : Instructor's Solutions Manual for for Modern Control Systems, 12th Ed Creator                         : Richard C. Dorf, Robert H. Bishop Subject                         : 9780136024989 Document ID                     : uuid:107e56d4-5005-604d-a266-e210e0c618a9 Instance ID                     : uuid:3e3e4aa5-ad61-4d55-bcd8-16fef4e09972 Has XFA                         : No Page Count                      : 754 Page Layout                     : SinglePage Author                          : Richard C. Dorf Keywords                        : 9780136024989

EXIF Metadata provided by EXIF.tools

Modern Control Systems Thirteenth Edition Solutions

Source: https://usermanual.wiki/Pdf/ModernControlSystems12thEditionSolutionsManual.2023358919/help

Rowland Maidest

Modern Control Systems Thirteenth Edition Solutions

Instructor's Solutions Manual For Modern Control Systems, 12th Ed Systems Edition

Menu Halaman Statis