Rourkela

PCEC4303CONTROL SYSTEM ENGINEERING

Module1.............................................................................................................................................. 4
BasicConceptsofControlSystems,Openloopandclosedloopsystems ........................................ 4
Openloopsystems:....................................................................................................................... 4
Closedloopsystems ...................................................................................................................... 4
TransferFunction........................................................................................................................... 6
DifferentialEquationsofPhysicalSystems:Mechanical,TranslationalSystems ............................. 7
Rotationalsystems,GearTrains,ElectricalSystems....................................................................... 10
DerivationofTransferfunctions,BlockDiagramAlgebra............................................................... 16
SignalflowGraphs,MasonsGainFormula. ................................................................................... 25
FeedbackcharacteristicsofControlSystems:Effectofnegativefeedbackonsensitivity. ............ 27
Bandwidths,Disturbance,linearizingeffectoffeedback,Regenerativefeedback ........................ 28
A.C.Tachometer,Synchro,StepperMotors. .................................................................................. 31
Module2............................................................................................................................................ 42
TimeresponseAnalysis:StandardTestSignals............................................................................... 42
Timeresponseoffirstordersystemstounitstepandunitrampinputs ....................................... 45
TimeResponseofSecondordersystemstounitstepinput .......................................................... 46
TimeResponsespecifications,SteadyStateErrors. ....................................................................... 49
StaticErrorConstantsofdifferenttypesofsystems.Generalizederrorseriesandgeneralized
errorcoefficients............................................................................................................................. 50
StabilityandAlgebraicCriteria,conceptofstability,Necessaryconditionsofstability................. 51
Hurwitzstabilitycriterion,Routhstabilitycriterion,ApplicationoftheRouthstability. ............... 53
RootlocusTechnique:Rootlocusconcepts,RulesofConstructionofRootlocus......................... 56
Systemswithtransportationlag,PhaseLeadCompensation,phaseLagCompensation.............. 58
PhaseLagLeadCompensation,FeedbackCompensation............................................................ 61
Module3............................................................................................................................................ 63
FrequencyResponseAnalysis:Frequencydomainspecifications .................................................. 63
Procedurefordrawingthebodeplots.DeterminationofGainMarginandPhaseMarginfrom
Bodeplot ......................................................................................................................................... 72

1. IntroductiontoControlSystems:Basic elements of control system

Open loop and closed loop systems
TrackingSystem,Regulators
Differential equation
Transfer function
2. Modeling of electric systems: Translational and rotational mechanical systems
Block diagram reduction techniques.
Signal flow graph,MasonsGainFormula.
3. Feedback characteristics of Control Systems: Effect of negative feedback on
sensitivity.
Bandwidth,Disturbance.
Linearizingeffectoffeedback,Regenerativefeedback.
4. ControlComponents:D.C.Servomotors,A.C.Servomotors.
A.C.Tachometer.
Synchros,StepperMotors.

ModuleII
5. TimeresponseAnalysis:StandardTestSignals
Time response of first order systems to unit step and unit
rampinputs.
TimeResponseofSecondordersystemstounitstepinput.
TimeResponsespecifications,SteadyStateErrors.
GeneralisederrorseriesandGensalisederrorcoefficients.
6. Stability Theory:StabilityandAlgebraicCriteria
Conceptofstability,Necessaryconditionsofstability.
Hurwitzstabilitycriterion,Routhstabilitycriterion.
Application of the Routh stability criterion to linear
feedbacksystem.
Relativestabilitybyshiftingtheorigininsplane.
7. RootlocusTechnique:Rootlocusconcepts,RulesofConstructionofRootlocus
DeterminationofRootsfromRootlocusforaspecifiedopen
loopgain.
EffectofaddingopenlooppolesandzerosonRootlocus.
Rootcontours.
8. Compensation Technique:Systemswithtransportationlag.
Lead compensation.
Lag compensation
Lead-Lag compensation

ModuleIII
9. FrequencyResponseAnalysis:Frequencydomainspecifications
Correlation between Time and Frequency Response with

10. Stabilityinfrequencydomain:Principleofargument,Nyquiststabilitycriterion.
Applicationof Nyquist stabilitycriterion for linear feedback
system.
ConstantMcircles.
ConstantNCircles.
Nicholschart.
11.Controllers:ConceptofProportional,DerivativeandIntegralControlactions.
P,PD,PI,PIDcontrollers.
ZeiglerNicholsmethodoftuningPIDcontrollers.

ControlSystems

Module1
Introduction to Control Systems: Basic elements of control system Open loop and closed loop systems,
Tracking System, Regulators, and Differential equation, Transfer function. Modeling of electric systems
Translational and rotational mechanical systems. Block diagram reduction techniques. Signal flow graph,
Masons Gain Formula. Feedback characteristics of Control Systems: Effect of negative feedback on
sensitivity. Bandwidth, Disturbance. Linearizing effect of feedback, Regenerative feedback. Control
Components:D.C.Servomotors,A.C.Servomotors.A.C.Tachometer,Synchros,StepperMotors.

Lecture1
BasicConceptsofControlSystems,Openloopandclosedloopsystems

1. Basic elements of control system:In recent years, control systems have gained an increasingly importance in the development and
advancementofthemoderncivilizationandtechnology.Figureshowsthebasiccomponentsofacontrol
system.Disregardthecomplexityofthesystem;itconsistsofaninput(objective),thecontrolsystemand
itsoutput(result).Practicallydaytodayactivitiesareaffectedbysometypeofcontrolsystems.Thereare
twomainbranchesofcontrolsystems:
1)Openloopsystemsand
2)Closedloopsystems.

Open-loop systems:
Theopenloopsystemisalsocalledthenonfeedbacksystem.Thisisthesimplerofthetwosystems.
AsimpleexampleisillustratedbythespeedcontrolofanautomobileasshowninFigure12.Inthisopen
loopsystem,thereisnowaytoensuretheactualspeedisclosetothedesiredspeedautomatically.The
actualspeedmightbewayoffthedesiredspeedbecauseofthewindspeedand/orroadconditions,suchas
uphillordownhilletc.ExampleAutomaticwashingMachine,immersionrod,Afieldcontrold.cmotorand
automaticcontroloftrafficlamp.
.

(Fig1.2 Basic open-loop system)

Closed-loop systems:
The closedloop system is also called the feedback system. A simple closedsystem is shown in
Figure13.Ithasamechanismtoensuretheactualspeedisclosetothedesiredspeedautomatically.In
closedloopcontrolsystemsthecontrolactionisdependentondesiredoutput.Ifanysystemhavingoneor
morefeedbackpathsformingaclosedloopsystem.Exampleairconditionersareprovidedwiththermostat.

BPUT

Page4

ControlSystems

BPUT

Page5

ControlSystems
Lecture2
ServoMechanism/TrackingSystem,Regulators.
Transfer Function
A simpler system or element may be governed by first order or second order differential equation.
Whenseveralelementsareconnectedinsequence,saynelements,eachonewithfirstorder,thetotal
orderofthesystemwillbenthorder.Ingeneral,acollectionofcomponentsorsystemshallberepresented
bynthorderdifferentialequation

In control systems, transfer function characterizes the input output relationship of

componentsorsystemsthatcanbedescribedbyLinerTimeInvariantDifferentialEquation.
In the earlier period, the input output relationship of a device was represented graphically. In a system
havingtwoormorecomponentsinsequence,itisverydifficulttofindgraphicalrelationbetweentheinput
ofthefirstelementandtheoutputofthelastelement.Thisproblemissolvedbytransferfunction.
Definition of Transfer Function
Transfer function of a LTIV system is defined as the ratio of the Laplace Transform of the output
variabletotheLaplaceTransformoftheinputvariableassumingalltheinitialconditionaszero.
Properties of Transfer Function
The transfer function of a system is the mathematical model expressing the differential equation
thatrelatestheoutputtoinputofthesystem.
Thetransferfunctionisthepropertyofasystemindependentofmagnitudeandthenatureofthe
input.
The transfer function includes the transfer functions of the individual elements. But at the same
time,itdoesnotprovideanyinformationregardingphysicalstructureofthesystem.
Thetransferfunctionsofmanyphysicallydifferentsystemsshallbeidentical.
Ifthetransferfunctionofthesystemisknown,theoutputresponsecanbestudiedforvarioustypes
ofinputstounderstandthenatureofthesystem.
Ifthetransferfunctionisunknown,itmaybefoundoutexperimentallybyapplyingknowninputsto
thedeviceandstudyingtheoutputofthesystem.
How you can obtain the transfer function (T. F.)
Writethedifferentialequationofthesystem.
TaketheL.T.ofthedifferentialequation,assumingallinitialconditiontobezero.
Taketheratiooftheoutputtotheinput.ThisratioistheT.F.
Mathematical Model of control systems
A control system is a collection of physical object connected together to serve an objective. The
mathematicalmodelofacontrolsystemconstitutesasetofdifferentialequation.

BPUT

Page6

ControlSystems
Lecture-3
DifferentialEquationsofPhysicalSystems:Mechanical,TranslationalSystems.

2. Modelingofelectricsystems:
MechanicalTranslationalsystems
Themodelofmechanicaltranslationalsystemscanobtainbyusingthreebasicelementsmass,spring
anddashpot.Whenaforceisappliedtoatranslationalmechanicalsystem,itisopposedbyopposingforces
due to mass, friction and elasticity of the system. The force acting on a mechanical body is governed by
Newtonssecondlawofmotion.Fortranslationalsystemsitstatesthatthesumofforcesactingonabody
iszero.
Forcebalanceequationsofidealizedelements
Consideranidealmasselementshowninfig.whichhasnegligiblefrictionandelasticity.Letaforcebe
appliedonit.Themasswillofferanopposingforcewhichisproportionaltoaccelerationofabody.

Let

= applied force
= opposing force due to mass
Here

By Newtons second law,

Consideranidealfrictionalelementdashpotshowninfig.whichhasnegligiblemassandelasticity.Leta
forcebeappliedonit.Thedashpotwillbeofferanopposingforcewhichisproportionaltovelocityofthe
body.

Let

= applied force
= opposing force due to friction

Here,
By Newtons second law,
Consideranidealelasticelementspringshowninfig.whichhasnegligiblemassandfriction.

BPUT

Page7

ControlSystems

Let = applied force

= opposing force due to elasticity
Here,
By Newtons second law,

MechanicalRotationalSystems
The model of rotational mechanical systems can be obtained by using three elements, moment of
inertia [J] of mass, dash pot with rotational frictional coefficient [B] and torsion spring with
stiffness[k].When a torque is applied to a rotational mechanical system, and it is opposed by opposing
torques due to moment of inertia, friction and elasticity of the system. The torque acting on rotational
mechanicalbodiesisgovernedbyNewtonssecondlawofmotionforrotationalsystems.
Torquebalanceequationsofidealizedelements
Consider an ideal mass element shown in fig. which has negligible friction and elasticity. The opposing
torqueduetomomentofinertiaisproportionaltotheangularacceleration.

Let T = applied torque

T j =opposing torque due to moment of inertia of the body
Here, T j d2 /dt2
By Newtons law
T= T j = J d2/dt2
Consider an ideal frictional element dash pot shown in fig. which has negligible moment of inertia and
elasticity. Let a torque be applied on it. The dash pot will offer an opposing torque is proportional to
angularvelocityofthebody.

Let T = applied torque

Tb =opposing torque due to friction
Here

BPUT

Page8

ControlSystems
By Newtons law

Consideranidealelasticelement,torsionspringasshowninfig.whichhasnegligiblemomentofinertiaand
friction.Letatorquebeappliedonit.Thetorsionspringwillofferanopposingtorquewhichisproportional
toangulardisplacementofthebody.

Let T = applied torque

=opposing torque due to friction
Here,
By Newtons law

BPUT

Page9

ControlSystems
Lecture4
Rotationalsystems,GearTrains,ElectricalSystems.

Modeling of electrical system

Electrical circuits involving resistors, capacitors and inductors are considered. The behavior of
such systems is governed by Ohms law and Kirchhoffs laws.
Resistor: Consider a resistance of R carrying current I Amps as shown in Fig (a), then the
voltage drop across it is v = R I

Inductor: Consider an inductor L H carrying current I Amps as shown in Fig (a), then the
voltage drop across it can be written as

Capacitor: Consider a capacitor C F carrying current I Amps as shown in Fig (a), then the
voltage drop across it can be written as

Steps for modeling of electrical system

Apply Kirchhoffs voltage law or Kirchhoffs current law to form the differential equations
describing electrical circuits comprising of resistors, capacitors, and inductors.
Form Transfer Functions from the describing differential equations.
Then simulate the model.
Example

BPUT

Page10

ControlSystems
Electrical systems
LRC circuit. Applying Kirchhoffs voltage law to the system shown. We obtain the following equation
Resistance circuit

(1)
(2)
Equation (1) & (2) give a mathematical model of the circuit. Taking the L.T. of equations (1) & (2),
assuming zero initial conditions, we obtain

The transfer function

Armature-Controlled dc motors
Thedcmotorshaveseparatelyexcitedfields.Theyareeitherarmaturecontrolledwithfixedfieldor
fieldcontrolled with fixed armature current. For example, dc motors used in instruments employ a fixed
permanentmagnetfield,andthecontrolledsignalisappliedtothearmatureterminals.
Considerthearmaturecontrolleddcmotorshowninthefollowingfigure

Ra = armature-winding resistance, ohms

La = armature-winding inductance, henrys
I a = armature-winding current, amperes
I f = field current, a-pares
E a = applied armature voltage, volt
E b = back emf, volts
= angular displacement of the motor shaft, radians
T = torque delivered by the motor, Newton*meter
J = equivalent moment of inertia of the motor and load referred to the motor shaft kg .m2
f = equivalent viscous-friction coefficient of the motor and load referred to the motor shaft. Newton*m/rad/s
T = k1 I a where is the air gap flux, = k f I f, k1 is constant
BPUT

Page11

ControlSystems
For the constant flux
(1)
Where Kb is a back emf constant
The differential equation for the armature circuit
(2)
The armature current produces the torque which is applied to the inertia and friction; hence
(3)
Assuming that all initial conditions are condition are zero/and taking the L.T. of equations (1), (2) & (3), we
obtain

The T.F can be obtained is

BPUT

Page12

ControlSystems
Lecture5
AnalogybetweenMechanicalandelectricalquanties,Thermalsystems,fluidsystems.
Analogous Systems
Let us consider a mechanical (both translational and rotational) and electrical system as shown in the
fig.

From the fig (a)

We get
From the fig (b)

(1)

We get
From the fig (c)

(2)

(3)
We get
Where
They are two methods to get analogous system. These are (i) force- voltage (f-v) analogy and (ii) forcecurrent (f-c) analogy
(i)Force Voltage (f-v) Analogy
Translational
Force (F)
Mass (M)
Damper (D)
Spring (K)
Displacement (x)
Velocity (u)

Electrical
Voltage (V)
Inductance (L)
Resistance (R)
Elastance (1/C)
Charge (q)
Current (I)

(ii)Force Current (f-c) Analogy

Translational
Electrical
Force (F)
Current (I)
Mass (M)
Capacitance (C)
Damper (D)
Reciprocal of Inductance (1/L)
Spring (K)
Conductance (1/K)
Displacement (x)
Flux Linkage ( )
Velocity
BPUT

Voltage

Rotational
Torque (T)
Inertia (J)
Damper (D)
Spring (K)
Displacement ( )
Velocity ( )

Rotational
Torque (T)
Inertia (J)
Damper (D)
Spring (K)
Displacement (

Velocity
Page13

ControlSystems

Problem
Find the system equation for system shown in the fig. And also determine f-v and f-i analogies

For free body diagram M1

(1)
For free body diagram M2
(2)
Force voltage analogy
From Eq. (1) we get
(3)
From Eq. (2) we get
(4)
From Eq. (3) and (4) we can draw f-v analogy

Forcecurrentanalogy

BPUT

Page14

ControlSystems
From Eq. (1) we get

(5)
From eq (2) we get

(6)
From eq (5) and (6) we can draw force-current analogy

BPUT

Page15

ControlSystems
Lecture-6
DerivationofTransferfunctions,BlockDiagramAlgebra.

Thesystemcanberepresentedintwoforms:
Blockdiagramrepresentation
Signalflowgraph

Blockdiagram
Apictorialrepresentationofthefunctionsperformedbyeachcomponentandoftheflowofsignals
Basicelementsofablockdiagram
Blocks
Transferfunctionsofelementsinsidetheblocks
Summingpoints
Takeoffpoints
Arrow

Blockdiagram
A control system may consist of a number of components. A block diagram of a system is a pictorial
representationofthefunctionsperformedbyeachcomponentandoftheflowofsignals.Theelementsofa
blockdiagramareblock,branchpointandsummingpoint.

Block

Inablockdiagramallsystemvariablesarelinkedtoeachotherthroughfunctionalblocks.Thefunctional
block or simply block is a symbol for the mathematical operation on the input signal to the block that
producestheoutput.

Summingpoint
Althoughblocksareusedtoidentifymanytypesofmathematicaloperations,operationsofadditionand
subtractionarerepresentedbyacircle,calledasummingpoint.AsshowninFigureasummingpointmay
haveoneorseveralinputs.Eachinputhasitsownappropriateplusorminussign.
Asummingpointhasonlyoneoutputandisequaltothealgebraicsumoftheinputs.

BPUT

Page16

ControlSystems

A takeoff point is used to allow a signal to be used by more than one block or summing point. The
transferfunctionisgiveninsidetheblock
TheinputinthiscaseisE(s)
TheoutputinthiscaseisC(s)
C(s)=G(s)E(s)

Functionalblockeachelementofthepracticalsystemrepresentedbyblockwithits
Brancheslinesshowingtheconnectionbetweentheblocks
Arrowassociatedwitheachbranchtoindicatethedirectionofflowofsignal
Closedloopsystem
Summingpointcomparingthedifferentsignals
Takeoffpointpointfromwhichsignalistakenforfeedback
AdvantagesofBlockDiagramRepresentation
Verysimpletoconstructblockdiagramforacomplicatedsystem
Functionofindividualelementcanbevisualized
Individual&Overallperformancecanbestudied
Overalltransferfunctioncanbecalculatedeasily
DisadvantagesofBlockDiagramRepresentation
Noinformationaboutthephysicalconstruction
Sourceofenergyisnotshown

BPUT

Page17

ControlSystems
SimpleorCanonicalformofclosedloopsystem

R(s)Laplaceofreferenceinputr(t)
C(s)Laplaceofcontrolledoutputc(t)
E(s)Laplaceoferrorsignale(t)
B(s)Laplaceoffeedbacksignalb(t)
G(s)Forwardpathtransferfunction
H(s)Feedbackpathtransferfunction
Blockdiagramreductiontechnique
Because of their simplicity and versatility, block diagrams are often used by control engineers to
describe all types of systems. A block diagram can be used simply to represent the composition and
interconnectionofasystem.Also,itcanbeused,togetherwithtransferfunctions,torepresentthecause
andeffectrelationshipsthroughoutthesystem.TransferFunctionisdefinedastherelationshipbetweenan
inputsignalandanoutputsignaltoadevice
Blockdiagramrules

Cascadedblocks

Movingasummerbeyondtheblockmoving

BPUT

Page18

ControlSystems

Movingasummeraheadofblock

Movingapickoffaheadofblock

Movingapickoffbeyondablock

Eliminatingafeedbackloop

CascadedSubsystems

BPUT

Page19

ControlSystems

ParallelSubsystems

Feedback Control System

ProceduretosolveBlockDiagramReductionProblems
Step1:Reducetheblocksconnectedinseries
Step2:Reducetheblocksconnectedinparallel
Step3:Reducetheminorfeedbackloops
Step4:TrytoshifttakeoffpointstowardsrightandSummingpointtowardsleft
Step5:Repeatsteps1to4tillsimpleformisobtained
Step6:ObtaintheTransferFunctionofOverallSystem

BPUT

Page20

ControlSystems
Problem1

ObtaintheTransferfunctionofthegivenblockdiagram

CombineG1,G2whichareinseries

Combine G3, G4 which are in Parallel

BPUT

Page21

ControlSystems
Reduce minor feedback loop of G1, G2 and H1

Transferfunction

2.Obtainthetransferfunctionforthesystemshowninthefig

BPUT

Page22

ControlSystems

3.ObtainthetransferfunctionC/Rfortheblockdiagramshowninthefig

ThetakeoffpointisshiftedaftertheblockG2

Reducingthecascadeblockandparallelblock

Replacingtheinternalfeedbackloop

BPUT

Page23

ControlSystems

Equivalentblockdiagram

Transferfunction

BPUT

Page24

ControlSystems
Lecture-7
SignalflowGraphs,MasonsGainFormula.
Signal Flow Graph Representation
SignalFlowGraphRepresentationofasystemobtainedfromtheequations,whichshowstheflowofthe
signal

Signalflowgraph
A signal flow graph is a diagram that represents a set of simultaneous linear algebraic equations. By
taking Laplace transfer, the time domain differential equations governing a control system can be
transferredtoasetofalgebraicequationinsdomain.Asignalflowgraphconsistsofanetworkinwhich
nodes are connected by directed branches. It depicts the flow of signals from one point of a system to
anotherandgivestherelationshipsamongthesignals.

BasicElementsofaSignalflowgraph

Nodeapointrepresentingasignalorvariable.
Branchunidirectionallinesegmentjoiningtwonodes.
Path a branch or a continuous sequence of branches that can be traversed from one node to
anothernode.
Loopaclosedpaththatoriginatesandterminatesonthesamenodeandalongthepathnonodeis
mettwice.
Nonteachingloopstwoloopsaresaidtobenontouchingiftheydonothaveacommonnode.

Masonsgainformula
Therelationshipbetweenaninputvariableandanoutputvariableofsignalflowgraphisgivenbythenet
gainbetweentheinputandtheoutputnodesisknownasoverallgainofthesystem.Masonsgainrulefor
thedeterminationoftheoverallsystemgainisgivenbelow.

Where

=gainbetweenXinandXout

=outputnodevariable

=inputnodevariable

=totalnumberofforwardpaths

=pathgainofthekthforwardpath

=1(sumofloopgainsofallindividualloop)+(sumofgainproductofallpossiblecombinationsoftwo
nontouchingloops)(sumofgainproductsofallpossiblecombinationofthreenontouchingloops)
BPUT

Page25

ControlSystems
Problem

Forwardpathgain:
Closedloopgain
1.

2.

3.

4.

Nontouchingloopstakentwoatatime
5. Loop1andloop2:
6. Loop1andloop3:
7. Loop2andloop3:

Nontouchingloopstakenthreeatatime
8. Loop1,2,3:

Now,

Portionof nottouchingtheforwardpath

Hence,

BPUT

Page26

ControlSystems
Lecture-8
FeedbackcharacteristicsofControlSystems:Effectofnegativefeedbackonsensitivity.

3.FeedbackcharacteristicsofControlSystems:
Incontrolsystem,thefeedbackreducestheerror,alsoreducesthesensitivityofthesystemto
parameter variations .The parameter may vary due to some change in conditions .The variation in
parameteraffects theperformance of the system. So it is necessaryto make the system in sensitive to
suchparametervariations.
Effectoffeedbackonsensitivity
Theparametersofanycontrolsystemchangeswiththechangeenvironmentconditions.Alsothese
parameterscannotbeconstantthroughoutthelife.Theseparametervariationsaffecttheperformanceof
the system. For example, the resistance of winding of a motor changes due to change in temperature
duringitsoperation.
Sensitivitytomodeluncertainties

Openloop:

Closedloop:

Reduced

belowthatoftheopenloopsysbyincreasingG*H(>>1.0).

*ifGH>>1.0

Feedbackcomponentsshouldnot bevariedwithenvironmentalchangeschangeinH(s)directlyaffects
outputresponse

BPUT

Page27

ControlSystems

Lecture-9
Bandwidths,Disturbance,linearizingeffectoffeedback,Regenerativefeedback
Findtransferfunctionofafeedbackcontrolsystem

Closedloopcontrol

Hasabilitytoreducesystemsensitivity

If

forallcomplexfrequencyofinterest,then:

By increasing the gain of

it reduces the effect of

parametersoftheprocess,

But,making

WhenProcess,

on the input

variation of the

,isreduced

canleadtohighlyoscillatoryandevenunstableresponse
,ischanged

Openloop

Closedloop

Thenthechangeintheoutputis

BPUT

Page28

ControlSystems

,asisoftenthecase,wehave

when

Changeoftheoutputisreducedby

Disturbanceinasystem

StateVariableModel

Thetransferfunction

T.FfromthecontrolInput
BPUT

T.FfromD(s)tothe
Page29

ControlSystems
M(s)totheoutput

output

Disturbanceinaclosedloopsystem:

Where

Theloopgain

mustbemadelargetoreducethesystemsensitivity

ReducingDisturbance

Reducethegain

Increasetheloopgain

Reducethedisturbanced(t)

Feedforwardmethodifthedisturbancecanbemeasured

BPUT

(Choiceof

Page30

ControlSystems

Lecture-10
A.C.Tachometer,Synchro,StepperMotors.

4.ControlComponents:
SYNCHROS:
A synchro is a type of rotary electrical transformer that is used for measuring the angle of a rotating
machinesuchasanantennaplatform.Initsgeneralphysicalconstruction,itismuchlikeanelectricmotor.
Theprimarywindingofthetransformer,fixedtotherotor,isexcitedbyanalternatingcurrent,whichby
electromagneticinduction,causescurrentstoflowinthreestarconnectedsecondarywindingsfixedat120
degreestoeachotheronthestator.Therelativemagnitudesofsecondarycurrentsaremeasuredandused
todeterminetheangleoftherotorrelativetothestator,orthecurrentscanbeusedtodirectlydrivea
receiversynchrothatwillrotateinunisonwiththesynchrotransmitter

SynchroOperation:
Onapracticallevel,Synchroresemblesmotors,inthatthereisarotor,stator,andashaft.Ordinarily,slip
rings and brushes connect the rotor to external power. A synchro transmitter's shaft is rotated by the
mechanism that sends information, while the synchro receiver's shaft rotates a dial, or operates a light
mechanical load. Single and threephase units are common in use, and will follow the other's rotation
whenconnectedproperly.Onetransmittercanturnseveralreceivers;iftorqueisafactor,thetransmitter
mustbephysicallylargertosourcetheadditionalcurrent.
UsesofSynchro:
SynchrosystemswerefirstusedinthecontrolsystemofthePanamaCanalintheearly1900sto
transmitlockgateandvalvestempositionsandwaterlevelstothecontroldesks
FirecontrolsystemdesignsdevelopedduringWorldWarIIusedsynchrosextensively,totransmit
angularinformationfromgunsandsightstoananalogfirecontrolcomputer,andtotransmitthe
desiredgunpositionbacktothegunlocation.

BPUT

Page31

ControlSystems
Tachometers
Tachometerisanelectromechanicalunitwhichgeneratesanelectricaloutputproportionaltothespeedof
theshaft.Inautomaticcontrolsystemtachometerperformstwomainfunctions:

Stabilizationofsystem
Computationofclosedloopsinacontrolsystem

ACTachometer:

TheACtachometerisadevice,whichissimilartoatwophaseinductionmotor,inwhichtwostator
windings are placed in quadrature with each other and rotor is short circuited. In AC Tachometer, a
sinusoidalvoltageofratedvalueisappliedtotheprimarywinding,whichisknownasreferencewinding,
the secondary winding is placed 90 degrees apart from primary winding. The magnitude of sinusoidal
outputvoltageisdirectlyproportionaltothespeedofrotor.

D.C. Tachometer
Incontrolsystemsmostcommontypeoftachometersared.c.tachometers.D.C.Tachometercontainsan
ironcorerotorandpermanentmagnet.Themagneticfieldisprovidedbythepermanentmagnetandno
externalsupplyvoltageisnecessary.Theinputtothetachometeristhespeedoftheshaftandtheoutput
isvoltagewhichisproportionaltotheangularspeedoftheshaft.

Where

e=tachometergeneratorvoltage
K=tachometersensitivity

=angularspeedofshaft

Laplacetransformofequation,

Hencetransferfunctionoftachometeris

BPUT

Page32

ControlSystems

Ind.ctachometerthewindingonrotorareconnectedtothecommutatorandtheoutputvoltageistaken
acrossthebrushes.Thepermanentmagnettachometersarecompactandreliablebuthavinghighinertia.
Forreducingthevoltagedropacrossthebrushes,metalbrusheswithsilvertipsareused.

BPUT

Page33

ControlSystems
Lecturer-11
Stepper motors, Amplidyne
StepperMotors:
In stepper motors, the movement of rotor is in discrete steps. A stepper motor is electromechanical
device.Therearethreetypesofsteppermotors.
1. Variablereluctancemotors
2. Permanentmagnetmotors
3. Hybridtype

Conventionalservomotorsareclassifiedascontinuousrotationmotors
Steppermotorsrotatethroughaspecificnumberofdegrees,orsteps,thenstop
Eachincomingpulseresultsintheshaftturningaspecificangulardistance
Steppermotorscancontrolvelocity,distance,anddirectionofmechanicalload

PermanentMagnetStepperMotor:

BPUT

PMsteppermotorshaverotorteethmadeofpermanentmagnets
Reactionoftherotorteethtostatorfieldsprovidestorqueforthemotor.
Signalsareappliedtothestatortodeterminedirectionandsteprateoftherotor.

Page34

ControlSystems

StepperMotorSpeed

Steppermotorspeeddependsuponthestepangleandsteppingrate
N=(YS)/6
Wheren=speedinRPM
Y=stepangleindegrees
S=Stepspersecond
6=Formulaconstant

Amplidyne:
AnAmplidyneisarotatingamplifier.Itisaprimemoverdrivend.c.generatorwhoseoutputpower
canbecontrolledbyasmallfieldpowerinput.Anamplidyneiscapableofgivingacontrolledpoweroutput
intherangeofafewhundredtofewthousandwattswithapoweramplificationoftheorderof10,000
ormoreandhencefindswideapplicationinfeedbackcontrolsystem.
Studyofamplidynewhoseoutputpowercanbecontrolledbyasmallfieldpowerinput.Forstudying
thecharacteristics.
(i) We plot a graph of Output voltage against effective field current with no load, full load (500W), and
withoutcompensationwindingeffect
(ii)Wedrawtheschematicdiagramofanamplidynesystem

BPUT

Page35

ControlSystems

BPUT

Page36

ControlSystems
Lecture12
D.C.Servomotors,A.C.Servomotors.
Servomotor:
Theservosystemistheone,inwhichtheoutputissomemechanicalvariablesuchasposition,velocity
or acceleration. The motors used in the servo systems arecalled servomotors. These motors are usually
coupledtotheoutputshaftforpowermatching.Therearetwotypesofservomotors.
1. DCServomotors
2. ACServomotors
1.DCServomotors: D.C.servomotorsareseparatelyexcitedorpermanentmagnetd.c.servomotor
.The armature of d.c servomotor has a large resistance, therefore torque speed characteristics is linear.
Thetorquespeedcharacteristics

Showsinfig(c)andfig(a)showstheschematicdiagramofseparatelyexcitedd.c.servomotor.

DCservomotorsarecontrolledbyDCcommandsignalsapplieddirectlytocoils.Timeconstantfor
fieldcircuitislarge,duetolargetimeconstant,the
Responseisslowandthereforetheyarenotcommonlyused.
Themagneticfieldsthatareformedinteractwithpermanentmagnetsandcausetherotating
membertoturn.
OnetypeofPMusesawoundarmatureandbrusheslikeaconventionalDCmotor,butuses
magnetsaspolepieces
Anothertypeuseswoundfieldcoilsandapermanentmagnetrotor.

2.ACServomotors: Thesemotorhavingtwopartsnamelystatorandrotor.A.C.Servomotorsaretwophase
inductionmotor.Thestatorhastwodistributedwindings.Thesewindingsaredisplayedfromeachotherby
900.Onewindingiscalledmainwindingorreferencewinding.Thereferencewindingisexcitedbyconstant
a.c.voltage.Otherwindingiscalledcontrolwinding,thesewindingisexcitedbyvariablecontrolvoltageof
the same frequency as the reference winding but having a phase displacement of 900 electrical. The
variablecontrolvoltageforcontrolwindingisobtainedfromaservoamplifier.
Therotorofa.c.servomotorsareoftwotypes(a)squirrelcagerotor(b)dragcuptyperotor.The
squirrelcagerotorhavinglargelengthandsmalldiameter,soitsresistanceisveryhigh.Theairgapof
squirrelcageiskeptsmall.Indragcuptypetherearetwoairgaps.

BPUT

Page37

ControlSystems
Fortherotoracupofnonmagneticconductingmaterialisused.Astationaryironcoreisplacedbetween
theconductingcuptocompletethemagneticcircuit.Theresistanceofdragcuptypeishighandhaving
highstartingtorque.Fig(a)showstheschematicdiagramoftwophasea.c.servomotorandfig(b)shows
thetwotypesofrotor.

Fig(a)

Fig(b)
ControlledbyACcommandsignalsappliedtothecoils.
ACBrushlessServoMotorOperatesonthesameprincipleassinglephaseinductionmotor.
Torquespeedcharacteristic:

BPUT

Page38

ControlSystems
Applicationofservomotors:
Servomotoriswidelyusedinradars,electromechanicalactuators,computers,machinetools,tracking
andguidancesystem,processcontrollersandrobots.

UNIT I
CONTROL SYSTEM MODELLING

1. Whatiscontrolsystem?
Asystemconsistsofanumberofcomponentsconnectedtogethertoperformaspecificfunction.In
asystemwhentheoutputquantityiscontrolledbyvaryingtheinputquantitythenthesystemis
calledcontrolsystem.
2. Whatarethetwomajortypesofcontrolsystem?
Thetwomajortypesofcontrolsystemareopenloopandclosedloop.
3. Defineopenloopcontrolsystem.
The control system in which the output quantity has no effect upon the input quantity is called
openloopcontrolsystem.Thismeansthattheoutputisnotfeedbacktotheinputforcorrection.
4. Defineclosedloopcontrolsystem.
Thecontrolsysteminwhichtheoutputhasaneffectupontheinputquantitysoastomaintainthe
desiredoutputvaluesarecalledclosedloopcontrolsystem.
5. Whatarethecomponentsoffeedbackcontrolsystem?
Thecomponentsoffeedbackcontrolsystemareplant,feedbackpathelements,errordetectorand
controller.
6. Definetransferfunction.
TheT.FofasystemisdefinedastheratiooftheLaplacetransformofoutputtoLaplacetransform
ofinputwithzeroinitialconditions.
7. Whatarethebasicelementsusedformodelingmechanicaltranslationalsystem.
Mass,springanddashpot
8. Whatarethebasicelementsusedformodelingmechanicalrotationalsystem?
MomentofinertiaJ,dashpotwithrotationalfrictionalcoefficientBandtorsionspringwithstiffness
K
9. Nametwotypesofelectricalanalogousformechanicalsystem.
The two types of analogies for the mechanical system are Force voltage and force current
analogy.
10. Whatisblockdiagram?
A block diagram of a system is a pictorial representation of the functions performed by each
componentofthesystemandshowstheflowofsignals.Thebasicelementsofblockdiagramarrow
block,branchpointandsummingpoint.
11. Whatisthebasisforframingtherulesofblockdiagramreductiontechnique?
The rules for block diagram reduction technique are framed such that any modification made on
thediagramdoesnotaltertheinputoutputrelation.
12. Whatisasignalflowgraph?

BPUT

Page39

ControlSystems
Asignalflowgraphisadiagramthatrepresentsasetofsimultaneousalgebraicequations.Bytaking
L.Tthetimedomaindifferentialequationsgoverningacontrolsystemcanbetransferredtoasetof
algebraicequationsinsdomain.
13. Whatistransmittance?
Thetransmittanceisthegainacquiredbythesignalwhenittravelsfromonenodetoanothernode
insignalflowgraph.
14. Whatissinkandsource?
Source is the input node in the signal flow graph and it has only outgoing branches. Sink is an
outputnodeinthesignalflowgraphandithasonlyincomingbranches.
15. Definenontouchingloop.
Theloopsaresaidtobenontouchingiftheydonothavecommonnodes.
16. WriteMasonsGainformula.
MasonsGainformulastatesthattheoverallgainofthesystemis

and

WhereM= gainbetween
=
outputnodevariable
=
inputnodevariable
N =
totalnumberofforwardpaths
=
pathgainofthektheforwardpath
=1(sum of loop gains of all individual loop) + (sum of gain product of all
possible combinations of two non touching loops) (sum of gain products of all
possiblecombinationofthreenontouchingloops)
17. Writetheanalogouselectricalelementsinforcevoltageanalogyfortheelementsofmechanical
translationalsystem.
Force

voltagee
Velocity
v

currenti
Displacementx

chargeq
FrictionalcoefficientB

ResistanceR
Mass
M

InductanceL
Stiffness
K

Inverseofcapacitance1/C
18. Writetheanalogouselectricalelementsinforcecurrentanalogyfortheelementsofmechanical
translationalsystem.
Force

currenti
Velocityv

voltagev
Displacementx

flux
FrictionalcoefficientB

conductance1/R
MassM

capacitanceC
StiffnessK

Inverseofinductance1/L
19. Writetheforcebalanceequationofidealmasselement.

BPUT

Page40

ControlSystems
20. Writetheforcebalanceequationofidealdashpotelement.

21. Writetheforcebalanceequationofidealspringelement.

22. Distinguishbetweenopenloopandclosedloopsystem
OpenloopandClosedloop
Inaccurate
Simpleandeconomical
Thechangesinoutputduetoexternaldisturbancearenotcorrected.
Theyaregenerallystable
Accurate
Complexandcostlier
Thechangesinoutputduetoexternaldisturbancesarecorrectedautomatically.
Greateffortsareneededtodesignastablesystem.
23. Whatisservomechanism?
The servomechanism is a feedback control system in which the output is mechanical position (or
timederivativesofpositionvelocityandacceleration).

BPUT

Page41

ControlSystems

Module2
Lecture-13
TimeresponseAnalysis:StandardTestSignals.

5.TimeresponseAnalysis:
Introduction

Afterderivingamathematicalmodelofasystem,thesystemperformanceanalysiscanbedonein
variousmethods.
Inanalyzinganddesigningcontrolsystems,abasisofcomparisonofperformanceofvariouscontrol
systemsshouldbemade.Thisbasismaybesetupbyspecifyingparticulartestinputsignalsandby
comparingtheresponsesofvarioussystemstothesesignals.
The system stability, system accuracy and complete evaluation are always based on the time
responseanalysisandthecorrespondingresults
Nextimportantstepafteramathematicalmodelofasystemisobtained.
Toanalyzethesystemsperformance.
Normallyusethestandardinputsignalstoidentifythecharacteristicsofsystemsresponse
1. Stepfunction
2. Rampfunction
3. Impulsefunction
4. Parabolicfunction
5. Sinusoidalfunction

Itisanequationoraplotthatdescribesthebehaviorofasystemandcontainsmuchinformationaboutit
withrespecttotimeresponsespecificationas
Overshooting,settlingtime,peaktime,risetimeandsteadystateerror.Timeresponseisformedbythe
transientresponseandthesteadystateresponse.
Timeresponse=Transientresponse+Steadystateresponse

Transienttimeresponse(Naturalresponse)describesthebehaviorofthesysteminitsfirstshorttime
untilarrivesthesteadystatevalueandthisresponsewillbeourstudyfocus.Iftheinputisstepfunction
thentheoutputortheresponseiscalledsteptimeresponseandiftheinputisramp,theresponseiscalled
ramptimeresponse...etc.
ClassificationofTimeResponse
Transientresponse
Steadystateresponse
y(t)=yt(t)+yss(t)

TransientResponse
Thetransientresponseisdefinedasthepartofthetimeresponsethatgoestozeroastimebecomes
verylarge.Thusyt(t)hastheproperty

Limyt(t)=0
t>

BPUT

Page42

ControlSystems
Thetimerequiredtoachievethefinalvalueiscalledtransientperiod.Thetransientresponsemaybe
exponential or oscillatory in nature. Output response consists of the sum of forced response (form the
input)andnaturalresponse(fromthenatureofthesystem).Thetransientresponseisthechangeinoutput
response from the beginning of the response to the final state of the response and the steady state
responseistheoutputresponseastimeisapproachinginfinity(ornomorechangesattheoutput).

SteadyStateResponse:Thesteadystateresponseisthepartofthetotalresponsethatremainsafterthe
transienthasdiedout.Forapositioncontrolsystem,thesteadystateresponsewhencomparedtowiththe
desired reference position gives an indication of the final accuracy of the system. If the steady state
responseoftheoutputdoesnotagreewiththedesiredreferenceexactly,thesystemissaidtohavesteady
stateerror.
Timeresponse=Transientresponse+Steadystateresponse
TypicalInputSignals
1.ImpulseSignal
2.StepSignal
3.RampSignal
4.ParabolicSignal

BPUT

Page43

ControlSystems

TimeResponseAnalysis&Design

Twotypesofinputscanbeappliedtoacontrolsystem
CommandInputorReferenceInputyr(t)
DisturbanceInputw(t)(Externaldisturbancesw(t)aretypicallyuncontrolledvariationsintheload
onacontrolsystem)
Insystemscontrollingmechanicalmotions,loaddisturbancesmayrepresentforces.
Involtageregulatingsystems,variationsinelectricalloadareamajorsourceofdisturbances.

TestSignals
1. Inputr(t)R(S)
2. StepinputAA/S
3. RampinputAtA/S2
4. ParabolicinputAt2/2A/S3
5. Impulseinput(t)1
BPUT

Page44

ControlSystems
Lecture14
Timeresponseoffirstordersystemstounitstepandunitrampinputs

Firstordersystemtimeresponse

Transient
Steadystate
FirstOrderSystem

StepResponseofFirstOrderSystem
Evolution of the transient response is determined by the pole of the transfer function at s=1/t
wheretisthetimeconstantAlso,thestepresponsecanbefound:

Impulseresponse

Exponential

Stepresponse

Step,Exponential

Rampresponse

BPUT

Ramp,Step,Exponential

Page45

ControlSystems

Lecture15
TimeResponseofSecondordersystemstounitstepinput
Second-order systems: LTI second-order system

Second-Order Systems

Second order system responses

Over damped response:
Poles: Two real at
BPUT

-1 -2
Page46

ControlSystems
Natural response: Two exponentials with time constants equal to the reciprocal of the pole location

Under damped response:

Poles:twocomplexesat

Natural response: Damped sinusoid with an exponential envelope whose time constant is equal to the
reciprocalofthepolesradianfrequencyofthesinusoid,thedampedfrequencyofoscillation,isequalto
theimaginarypartofthepoles

Un-damped response:
Poles:Twoimaginaryat
Naturalresponse:

Undammedsinusoidwithradianfrequencyequaltotheimaginarypartofthepoles

Critically damped responses:

Poles: Two real at
Natural response: One term is anexponentialwhose time constant is equal to the reciprocal of the pole
location.Anothertermproductoftimeandanexponentialwithtimeconstantequaltothereciprocalofthe
polelocation

Second order system responses damping cases

BPUT

Page47

ControlSystems

Secondorderstepresponse
Complexpoles

BPUT

Page48

ControlSystems
Lecture-16
TimeResponsespecifications,SteadyStateErrors.

Steady State Error:

- Consider a unity feedback system
- Transfer function between e(t) and r(t)

Steady state error is

BPUT

Page49

ControlSystems
Lecture-17 ,18
StaticErrorConstantsofdifferenttypesofsystems.Generalizederrorseriesandgeneralizederrorcoefficients

Errorconstants

Steadystateerror

Typeofsystem
Unitstepinput Unitrampinput Unitparabolicinput

1/(1+K)

1/K

1/K

Output Feedback Control Systems

Feedback only the output signal

Easy access
Obtainable in practice
6.Stability Theory: Stability and Algebraic Criteria: A system is stable if any bounded input produces a
bounded output for all bounded initial conditions.

BPUT

Page50

ControlSystems
Lecture-19
StabilityandAlgebraicCriteria,conceptofstability,Necessaryconditionsofstability.

Basic concept of stability

Stability of the system and roots of characteristic equations

Characteristic Equation
BPUT

Page51

ControlSystems
Consider an nth-order system whose the characteristic equation (which is also the denominator of the transfer
function) is

BPUT

Page52

ControlSystems
Lecture20,21
Hurwitzstabilitycriterion,Routhstabilitycriterion,ApplicationoftheRouthstability.

Routh Hurwitz Criterion

Goal: Determining whether the system is stable or unstable from a characteristic equation in
polynomial form without actually solving for the roots
Rouths stability criterion is useful for determining the ranges of coefficients of polynomials
for stability, especially when the coefficients are in symbolic (non numerical) form
To find Kmar &

A necessary condition for Rouths Stability

A necessary condition for stability of the system is that all of the roots of its characteristic
equation have negative real parts, which in turn requires that all the coefficients be positive.
A necessary (but not sufficient) condition for stability is that all the coefficients of the
polynomial characteristic equation are positive & none of the co-efficient vanishes.
Rouths formulation requires the computation of a triangular array that is a function of the
coefficients of the polynomial characteristic equation.
A system is stable if and only if all the elements of the first column of the Routh array are
positive

Method for determining the Routh array

Consider the characteristic equation

Routh array method

Then add subsequent rows to complete the Routh array
Compute elements for the 3rd row:

BPUT

Page53

ControlSystems
Given the characteristic equation,

Is the system described by this characteristic equation stable?

Answer:

All the coefficients are positive and nonzero

Therefore, the system satisfies the necessary condition for stability
We should determine whether any of the coefficients of the first column of the Routh array are
negative
S6:1314
S5:4240
S4:2.504
S3:22.40
S2:34
S1:1.60
S0:4
The elements of the 1st column are not all positive. Then the system is unstable.
Special cases of Rouths criteria
Case 1: All the elements of a row in a RA are zero

Form Auxiliary equation by using the co-efficient of the row which is just above the row of zeros

Find derivative of the A.E.

Replace the row of zeros by the co-efficient of dA(s)/ds

complete the array in terms of these coefficients

analyze for any sign change, if so, unstable

no sign change, find the nature of roots of AE

non-repeated imaginary roots - marginally stable

repeated imaginary roots unstable

Case 2: First element of any of the rows of RA is

BPUT

Page54

ControlSystems

Zero and the same remaining row contains at least one non-zero element
Substitute a small positive no. in place of zero and complete the array.

Examine the sign change by taking Lt = 0

BPUT

Page55

ControlSystems
Lecture-22, 23
RootlocusTechnique:Rootlocusconcepts,RulesofConstructionofRootlocus

7.RootlocusTechnique:

IntroducedbyW.R.Evansin1948
Graphicalmethod,inwhichmovementofpolesinthesplaneissketchedwhensomeparameter
isvaried
ThepathtakenbytherootsofthecharacteristicequationwhenopenloopgainKisvariedfrom
0toarecalledrootloci
DirectRootLocus=0<k<
InverseRootLocus=<k<0

RootLocusAnalysis:

Therootsoftheclosedloopcharacteristicequationdefinethesystemcharacteristicresponses
Their location in the complex splane lead to prediction of the characteristics of the time domain
responsesintermsof:
odampingratio,
onaturalfrequency,wn
odampingconstant,firstordermodes
oConsiderhowtheserootschangeastheloopgainisvariedfrom0to

BasicsofRootLocus:

Symmetricalaboutrealaxis
RLbranchstartsfromOLpolesandterminatesatOLzeroes
No.ofRLbranches=No.ofpolesofOLTF
Centroidiscommonintersectionpointofalltheasymptotesontherealaxis
AsymptotesarestraightlineswhichareparalleltoRLgoingtoandmeettheRLat
No.ofasymptotes=No.ofbranchesgoingto
AtBreakAwaypoint,theRLbreaksfromrealaxistoenterintothecomplexplane
AtBIpoint,theRLenterstherealaxisfromthecomplexplane

ConstructingRootLocus:

LocatetheOLpoles&zerosintheplot
Findthebranchesontherealaxis
Findangleofasymptotes&centroid
a=180(2q+1)/(nm)
a=(poleszeroes)/(nm)

FindBAandBIpoints

FindAngleOfdeparture(AOD)andAngleOfArrival(AOA)
AOD=180(sumofanglesofvectorstothecomplexpolefromallotherpoles)+
(Sumofanglesofvectorstothecomplexpolefromallzero)
BPUT

Page56

ControlSystems

AOA=180(sumofanglesofvectorstothecomplexzerofromallotherzeros)+(sumofanglesof
vectorstothecomplexzerofrompoles)
FindthepointofintersectionofRLwiththeimaginaryaxis.

ApplicationoftheRootLocusProcedure:

Step1:Writethecharacteristicequationas

Step2:Rewriteprecedingequationintotheformofpolesandzerosasfollows

Step3:
Locate the poles and zeros with specific symbols, the root locus begins at the openloop poles
andendsattheopenloopzerosasKincreasesfrom0toinfinity
If openloop system has nm zeros at infinity, there will be nm branches of the root locus
approachingthenmzerosatinfinity
Step4:
Therootlocusontherealaxisliesinasectionoftherealaxistotheleftofanoddnumberofreal
polesandzeros
Step5:
Thenumberofseparatelociisequaltothenumberofopenlooppoles
Step6:
Therootlocimustbecontinuousandsymmetricalwithrespecttothehorizontalrealaxis
Step7:
Thelociproceedtozerosatinfinityalongasymptotescenteredatcentroidandwithangles

Step8:

Theactualpointatwhichtherootlocuscrossestheimaginaryaxisisreadilyevaluatedbyusing
Routhscriterion

Step9:
Determinethebreakawaypointd(usuallyontherealaxis)
Step10:
Plottherootlocusthatsatisfythephasecriterion

Step11:
DeterminetheparametervalueK1ataspecificrootusingthemagnitudecriterion

BPUT

Page57

ControlSystems
Lecture-24, 25
Systemswithtransportationlag,PhaseLeadCompensation,phaseLagCompensation

8.Compensation Technique:
Series Compensation or Cascade Compensation
This is the most commonly used system where the controller is placed in series with the controlled
process.
Figure shows the series compensation

Feedback Compensation or Parallel Compensation

Series-Feedback Compensation

Lead Compensator
It has a zero and a pole with zero closer to the origin. The general form of the transfer function of the load
compensator is

BPUT

Page58

ControlSystems

Here,

Subsisting
;
Transfer function

Lag Compensator
It has a zero and a pole with the zero situated on the left of the pole on the negative real axis. The general
form of the transfer function of the lag compensator is

Therefore, the frequency response of the above transfer function will be

BPUT

Page59

ControlSystems

Now comparing with

Therefore

BPUT

Page60

ControlSystems
Lecture-26, 27
PhaseLagLeadCompensation,FeedbackCompensation
Lag-Lead Compensator
The lag-lead compensator is the combination of a lag compensator and a lead compensator. The lag section is
provided with one real pole and one real zero, the pole being to the right of zero, where as the lead section
has one real pole and one real came with the zero being to the right of the pole.
The transfer function of the lag-lead compensator will be

The figure shows lag lead compensator

Where

BPUT

Page61

ControlSystems
The above transfer functions are comparing with

Then

Therefore

Where

BPUT

Page62

ControlSystems

Module3
Lecture-28, 29
FrequencyResponseAnalysis:Frequencydomainspecifications
WhatisFrequencyResponse?
Considerasystemwithasinusoidalinput

Understeadystate,thesystemoutputaswellassignalsatallotherpointsinthesystemaresinusoidal.The
steadystateoutputmaybewrittenas

The magnitude and phase relationship between the sinusoidal input and the steady state output of a
system is termed as frequency response. In linear timeinvariant systems, the frequency response is
independentoftheamplitudeandphaseoftheinputsignal.

AdvantagesofFrequencyResponseAnalysis:
Thefrequencyresponsetestonasystemoracomponentisnormallyperformedbykeepingtheamplitude
A fixed and determining B and for a suitable range of frequencies. Signal generators and precise
measuringinstrumentsarereadilyavailableforvariousrangesoffrequenciesandamplitudes.

BPUT

Page63

ControlSystems

Whenever it is not possible to obtain the form of the transfer function of a system through
analyticaltechniques,thenecessaryinformationtocomputethetransferfunctioncanbeextracted
byperformingthefrequencyresponsetestonthesystem.
Thedesignandparameteradjustmentoftheopenlooptransferfunctionofasystemforspecified
closed loop performance is carried out somewhat more easily in frequency domain than in time
domain.
Theeffectofnoisedisturbanceandparametervariationsarerelativelyeasytovisualizeandassess
throughfrequencyresponse.
TheNyquistStabilitycriterionisapowerfulfrequencydomainmethodofextractingtheinformation
regarding stability as well as relative stability of a system without the need to find roots of the
characteristicequation.

HowtoobtainSteadyStateOutputstoSinusoidalInputs?
The Laplace Transform of the output of a linear singleinput, singleoutput system with transfer function
G(s)canbeexpressedintermsoftheinputas

We know, in general that

. However, it will be shown here that, for sinusoidal steadystate

analysis,weshallreplacesbyitsimaginarycomponent

only,sinceinsteadystate,thecontributionof

therealpart willdisappearforastablesystem.
Considerthestable,linearsystemshownbelow.
r(t)

G(s)

R(s

c
C(s

Letusassumethattheinputsignal

SupposethatthetransferfunctionG(s)ofthesystemcanbewrittenasaratiooftwopolynomialsinsas

TheLaplaceTransformoftheoutputofthesystemisthen
Transformoftheinputr(t).

,whereR(s)istheLaplace

Where,
BPUT

istheconjugateof

.TheinverseLaplaceoftheaboveequationyields
Page64

ControlSystems

Forastablesystem,a,b,chavepositiverealparts.Hence,astapproaches atsteadystate,alltheterms
intheexpressionforc(t)willvanishexceptthelasttwoterms.Thusatsteadystate,theresponsebecomes

Regardless of whether there are simple or multiple poles of G(s), the contribution due to them to the
steadystateresponsewillzero.
WheretheconstantKcanbeevaluatedasfollows:
|s=jw
|s=jw
Since

isacomplexquantity,itcanbewrittenintheform

Where

representsthemagnitudeand representstheangleof

,
.

Similarly,

Wecannowwrite,

Where

=AM

Hence,forastable,LTIsystem,subjectedtosinusoidalinput,theamplitudeoftheoutputisgivenbythe
productofthatoftheinputand
,whilethephaseanglediffersfromthatoftheinputbyanamount

.Apositivephaseangle iscalledphaseleadwhereasanegativephaseangleiscalledphase

lag.

BPUT

Page65

ControlSystems

Input sinusoid

FrequencyDomainparametersofprototype2ndordersystems:
Theclosedlooptransferfunctionofaprototype2ndordersystemisgivenby

isthedampingfactorand

Where

istheundampednaturalfrequency.

Thesinusoidaltransferfunctionofthesystemisobtainedbysubstitutings=

Hence,

Where,

isthenormalizeddrivingsignalfrequency

Itisseenthatwhen

BPUT

Page66

ControlSystems

The Magnitude and Phase angle characteristics for normalized frequency

shownbelow.

for certain values of are

ThefrequencywhereMhasapeakvalueisknownasResonantFrequency.Atthisfrequency,slopeofthe
Magnitudecurveiszero.Let
frequency.Then

betheresonantfrequencyand

bethenormalizedresonant

BPUT

Page67

ControlSystems
(1)

or,
Themaximumvalueofthemagnitude,knownastheResonantPeakisgivenby

(2)
Thephaseangle of

attheresonantfrequencyisgivenby

From Eqn. (1) and Eqn. (2), it is seen that as approaches zero,

approaches

infinity.For0<
,theresonantfrequencyalwayshasavaluelessthan
hasavaluegreaterthan1.

and

approaches

andtheresonantpeak

For>1/ ,itisseenthat
,slopeofthemagnitudecurvedoesnotbecomezeroforanyrealvalue
of .Forthisrangeof,themagnitudeofMdecreasesmonotonicallyfromM=1atu=0withincreasingu,
asshownintheabovefigure.Itthereforefollowsthatfor>1/
thegreatestvalueofMequals1.

,thereisnoresonantpeakandassuch

Asisevidentfromtheaboveequations,forasecondordersystem,theresonantpeak

ofitsfrequency

response is indicative of its damping factor for 0 < 1/ , and the resonant frequency of the
frequencyresponseisindicativeofitsnaturalfrequencyforagivenandhenceindicativeofitsspeedof
response (as
.
and
of the frequency response could thus be used as
performanceindicesforasecondordersystem.
For
, M decreases monotonically. The frequency at which M has a value 1/ is of special
significanceandiscalledthecutofffrequency .Thesignalfrequenciesabovethecutofffrequencyare
greatlyattenuatedinpassingthroughasystem.
Forfeedbackcontrolsystems,therangeoffrequenciesoverwhichMisequalto orgreaterthan1/ is
defined as the bandwidth
. Control systems being lowpass filters (At zero frequency, M=1), the
bandwidth isequaltothecutofffrequency .
Ingeneral,thebandwidthofacontrolsystemindicatesthenoisefilteringcharacteristicofthesystem.Also,
thebandwidthgivesameasureofthetransientresponsepropertiesasobservedbelow.

BPUT

Page68

ControlSystems
Mr
M(

Bandwidth

Magnitude

M(

wr

wc
Frequency,

Thenormalizedbandwidth
asfollows:

ofthesecondordersystemunderconsiderationcanbedetermined

Solvingfor

,weget,

Itcanbeapproximatedinlinearformas

We thus observe that the normalized bandwidth is a function of damping only. The denormalized
bandwidthcanbewrittenas

CorrelationbetweenTimeDomainandFrequencyDomain:
Letusconsiderthestepresponseofthesecondordersystem.Thepeakovershoot
for0<
BPUT

ofthestepresponse

1is
Page69

ControlSystems

Thecomparisonof and plotsisshownbelow.Itshowsthatfor0< 1/ ,thetwoperformance

indicesarecorrelatedasbotharefunctionsofthesystemdampingfactoronly.Itmeansthatasystem
withagivenvalueof
astepinput.For

ofitsfrequencyresponse,mustexhibitacorrespondingvalueof
1/

,theresonantpeak

ifsubjectedto

doesnotexistandthecorrelationbreaksdown.

Similarly,theexpressionfordampednaturalfrequencyforasecondordersystemisgivenas

Thus, there exists definite correlation between

oscillationofthestepresponse.

of the frequency response and damped frequency of

It is further observed that the bandwidth, a frequency domain concept, is indicative of the undamped
naturalfrequencyofasystemforagiven ,andthereforeindicativeofthespeedofresponse
atimedomainconcept.
CommonlyusedfrequencyresponseanalysisMethods:

CommonlyusedfrequencyresponseanalysisMethodsare:

Bodeplot
BPUT

Page70

ControlSystems

BPUT

Nyquistplot
Nicholschart

Page71

ControlSystems
Lecture-30, 31
Procedurefordrawingthebodeplots.DeterminationofGainMarginandPhaseMarginfromBodeplot
Bodeplotconsistsoftwosimultaneousgraphs:

MagnitudeindB[(20log|G(j)|)(Base10)]vs.frequency(inlog)

Phase(indegrees)vs.frequency(inlog)
Inthelogarithmicrepresentation,thecurvesaredrawnonsemilogpaper,usingthelogscaleforfrequency
andthelinearscaleforeithermagnitude(inDecibels)orphaseangle(indegrees).
AdvantagesofBodePlot:

MultiplicationofMagnitudescanbeconvertedintoaddition

AsimplemethodofsketchingBodePlotisbasedonasymptoticapproximations.Suchinformation
on straight line asymptotes is sufficient if only rough information on frequency response
characteristicsisneeded.

Shouldtheexactcurvebedesired,correctionscanbemadeeasilytothesebasicasymptoticplots.
Lowfrequencyresponsecontainssufficientinformationaboutthephysicalcharacteristicsofmostof
thepracticalsystems.
ExperimentaldeterminationofatransferfunctionispossiblethroughBodeplotanalysis.

BodeDiagrams

InBodediagrams,frequencyratiosareexpressedintermsof:
Octave:itisafrequencybandfrom1to21.

Decade:itisafrequencybandfrom1to101,where1isanyfrequencyvalue.

Thebasicfactorswhichoccurfrequentlyinanarbitrarytransferfunctionare:
GainK

Integralandderivatives:(j)1
Firstorderfactors:
QuadraticFactors:

BodeDiagrams

(1+jT)1,T=1a

ForConstantGainK,logmagnitudecurveisahorizontalstraightlineatthemagnitudeof(20logK)dB
andphaseangleis0deg.
Varying the gain K, raises or lowers the logmagnitude curve of the transfer function by the

BPUT

Page72

ControlSystems

correspondingconstantamount,buthasnoeffectonthephasecurve
Logarithmic representation of the frequencyresponse curve of factor ( j ( / a) +1) can be
approximatedbytwostraightlineasymptotes
Frequencyatwhichthetwoasymptotesmeetiscalledthecornerfrequencyorbreakfrequency.

TheGainK:
MagnitudeResponse:
LogMagnitude=20logK
Asanumberincreasesbyafactorof10,thecorrespondingvalueincreasesbyafactorof20.Thismaybe
seenfromthefollowing:

Again,whenexpressedindecibels,thereciprocalofanumberdiffersfromitsvalueonlyinsign,i.e.,forthe
numberK,

IntegralandDerivativeFactors
LogMagnitudePlot:
LogMagnitudeof

is

PhasePlot:
Thephaseangleof isconstantandequalto90.

isplottedonalogarithmic,scale,itisastraightline.Todrawthisstraight
Ifthelogmagnitude
line,weneedtolocateonepoint(0dB,

onit.Since

Theslopeofthelineis20dB/decadeor6dB/octave.

Similarly,
LogMagnitudeof

is

Thephaseangleof

isconstantandequalto90.

Itcanbeseenthatthedifferencesinthefrequencyresponsesof
magnitudecurvesandinthesignsofthephaseangles.

BPUT

and

lieintheslopesofthelog

Page73

ControlSystems
Ifthetransferfunctioncontainsthefactor

or

,thelogmagnitudebecomesrespectively,

20log
20log

or

Theslopesofthelogmagnitudecurveforthefactors
dB /decade respectively. The phase angle of
whereasthatof
point(0dB,

is90

and

arethus20

is equal to 90

dB/decadeand20

over the entire frequency range,

overtheentirefrequencyrange.Themagnitudecurvewillpassthroughthe

FirstOrderFactors

LogMagnitudeCurve:

Thelogmagnitudeofthefirstorderfactor
Forlowfrequencies,suchthat
20log

is20log

=20log

dB

,thelogmagnitudemaybeapproximatedby

=20log1=0dB.

Thus, the log magnitude curve at low frequencies is the constant 0dB line. For high frequencies, such
that

,20log

20log

=20log

,thelogmagnitudeequals0dB;at

At

decreasesby20dBforeverydecadeof

20log dB.

,thelogmagnitudeis20dB.Thus,thevalueof20log
.For

,thelogmagnitudecurveisthusastraightline

withaslopeof20dB/decade(or6dB/octave).
Our analysis shows that the logarithmic representation of the frequencyresponse curve for the factor
1/

can be approximated by two straightline asymptotes, one a straightline at 0 dB for the

frequencyrange

andtheotherastraightlinewithslope20dB/decadeforthefrequencyrange

ThefrequenciesatwhichthetwoasymptotesmeetiscalledtheCornerFrequencyortheBreakFrequency.
For the factor

is the corner frequency. The corner frequency thus divides the frequency

responsecurveintotworegions:Thelowfrequencyregionandthehighfrequencyregion.
BPUT

Page74

ControlSystems
PhasePlot:
Theexactphaseangle ofthefactor1/

is

Atzerofrequency,thephaseangleis0.Atthecornerfrequency,thephaseangleis

Atinfinitefrequency,thephaseanglebecomes90.Sincethephaseangleisgivenbyaninversetangent
function,itisskewsymmetricabouttheinflectionpointat

ErrorintheMagnitudecurve:
TheerrorintheMagnitudecurvecausedbytheuseofasymptotescanbecalculated.
Error at a particular frequency = Actual value Approximate value of the logmagnitude curve at that
frequency
Themaximumerroroccursatthecornerfrequency
Actualvalue=20log

Approximatevalue=20log1=0dB.
Thus,erroratcornerfrequency=3dB.
Theerroratoneoctavebelowthecornerfrequency,i.e.,at

is

Theerroratoneoctaveabovethecornerfrequency,i.e.,at

is

20log

Thus,theerroratoneoctaveavoveorbelowthecornerfrequencyisapproximately1dB.

Thetransferfunction

hasthecharacteristicsofalowpassfilter.Forfrequenciesabove

logmagnitudefallsofrapidlytowards

BPUT

Page75

,the

ControlSystems

QuadraticFactors

Whentherearecomplexconjugatezeroes,theprototype2ndordersystemswillhavethetransferfunction
,

Whentherearecomplexconjugatepoles,theprototype2ndordersystemswillhavethetransferfunction
,

Forthecomplexconjugatepoles,

LogMagnitudeCurve:
BPUT

Page76

ControlSystems

Logmagnitude=

Forlowfrequencies,i.e.,

Logmagnitudebecomes20log1=0dB
Thelowfrequenciesasymptoteisthusahorizontallineat0dB.
Forhighfrequenciesi.e.,

,
=40log

Logmagnitudebecomes20log

dB=4040log

dB.

Thehighfrequencyasymptoteisthusastraightlinehavingtheslope40dB/decade.
Thehighfrequencyasymptoteintersectsthelowfrequencyoneat

,thecornerfrequency.

Thetwoasymptotesderivedareindependentof .Theresonantpeakoccursnearthefrequency

Thedampingration determinesthemagnitudeofthisresonantpeak.Themagnitudeoferrorscausedby
thestraightlineasymptotesdependonthevalueof Itislargeforsmallvaluesof
PhasePlot:

Thephaseangleofthequadraticfactor

is

Thephaseangleisafunctionofboth and

Thephaseanglecurveisskewsymmetricabouttheinflectionpointwhere

Thefrequencyresponseforthefactor

Can be obtained by merely reversing the sign of the log magnitude and that of the phase angle for the
factor

BPUT

Page77

ControlSystems

RelationshipbetweenSystemTypeandLogMagnitudeCurve:
Foraunityfeedbacksystemthestaticposition,velocityandaccelerationerrorconstantsdescribethelow
frequencybehavioroftype0,type1,andtype2systemsrespectively.Foragivensystem,onlyoneofthe
staticerrorconstantsisfiniteandsignificant.(Thelargerthevalueof thefinitestaticerrorconstant,the
highertheloopgainisas

approacheszero.)

The type of the system determines the slope of the logmagnitude curve at low frequencies. Thus,
information concerning the existence and magnitude of the steadystate error of a control system to a
given input can be determined from the observation of the lowfrequency region of the logmagnitude
curve.
DeterminationofStaticErrorconstants:
Assumethattheopenlooptransferfunctionofaunityfeedbacksystemisgivenby

Or

BPUT

Page78

ControlSystems
StaticPositionErrorconstant:
Thefigureshownbelowshowsanexampleofthelogmagnitudeplotofatype0system.Insuchasystem,
equals

themagnitudeof

atlowfrequencies,or

dB

20 log Kp

-20 dB/decade

-40 dB/decade

0
In log scale

StaticVelocityErrorconstant:
The figure given below shows an example of the logmagnitude of a type 1 unity feedback system. The
intersectionoftheinitial20dB/decadesegment(oritsextension)withtheline
20log .Thismaybeseenasfollows.

hasthemagnitude

Inatype1system,

Thus,
BPUT

Page79

ControlSystems

Theintersectionoftheinitial20dB/decadesegment(oritsextension)withthe0dBlinehasafrequency
numericallyequalto ,i.e.,ifthefrequencyatthisintersectionis ,then
or,

StaticAccelerationErrorconstant:
The figure given below shows an example of the logmagnitude of a type 2 unity feedback system. The
intersectionoftheinitial40dB/decadesegment(oritsextension)withtheline
20log .Thismaybeseenasfollows.

hasthemagnitude

Sinceatlowfrequencies,

Itfollowsthat

Thefrequency attheintersectionoftheinitial40dB/decadesegment(oritsextension)withthe0dB
linegivesthesquarerootof numerically.Thiscanbeseenfromthefollowing.

Whichyields
BPUT

Page80

ControlSystems
PhaseMargin(PM):
Phasemarginisthatamountofadditionalphaselagatthegaincrossoverfrequencyrequiredtobringthe
systemtothevergeofinstability.
GainCrossoverFrequency:
Gain crossover frequency is that frequency at which,
functionisunity.
The Phase margin PM is 180 plus the phase angle
crossoverfrequency.

, the magnitude of the open loop transfer

of the open loop transfer function at the gain

GainMargin(PM):
GainMarginisthereciprocalofthemagnitude

atthePhasecrossoverfrequency.

PhaseCrossoverFrequency:
Phasecrossoverfrequencyisthatfrequencyatwhich,
functionequals180.

,thephaseangleoftheopenlooptransfer

Thus,GainMargin,

Intermsofdecibels,

AFewCommentsonPhaseandGainMargins:

BPUT

For a stable nonminimum phase system, the gain margin indicates how much the gain can be
increasedbeforethesystembecomesunstable.Foranunstablesystem,thegainmarginindicates
howmuchthegaincanbedecreasedbeforethesystembecomesstable.
TheGainMarginofafirstandSecondordersystemisinfinitesincethepolarplotofsuchsystems
doesnotcrosstherealaxis.Thus,theoretically,the1stand2ndordersystemscannotbeunstable.
Itisimportanttopointoutthatconditionallystablesystemswillhavetwoormorephasecrossover
frequenciesand some higher ordersystems with complicated numerator dynamics mayalso have
two or more gain crossover frequencies. For stable systems having two or more gain crossover
frequencies,thePhaseMarginismeasuredatthehighestGainCrossoverFrequency.
EithertheGainMarginaloneorthePhaseMarginalonedoesnotgiveasufficientindicationofthe
relativestability.Bothshouldbegivenfordeterminationofstability.

Page81

ControlSystems

For satisfactory performance, PM should be between 30 and 60 and the GM should be greater
than6dB.
TherequirementthatthePMbebetween30and60meansthatinBodediagram,theslopeofthe
logmagnitudecurveatthegaincrossoverfrequencyshouldbemoregradualthan40dB/decade.
In most practical cases, a slope of 20 dB/decade is desirable. If the slope at the gain crossover
frequencyis60dB/decadeorsteeper,thesystemismostlikelyunstable.

PHASEANDGAINMARGINTHROUGHBODEPLOTS:

BPUT

Page82

ControlSystems
Lecture-31
Polarplots.Polarplotsofsomestandardtypefunction
Asinusoidaltransferfunction

isacomplexfunctionandisgivenby

Or,

It is seen that

can be represented as a phasor of magnitude M and phase angle

positivelyincounterclockwisedirection).Astheinputfrequency

isvariedfrom0to

M and the phase angle change and hence the tip of the phasor
plane.ThelocusthusobtainedisknownasPolarPlotasshownbelow.

(Measured

,themagnitude

traces a locus in the complex

ProcedureforSketchingofthePolarPlot:
TosketchthePolarPlotofofagivenOpenLoopTransferFunctionovertheentirefrequencyrange,

ExpressthegivenexpressionfortheOLTFin(1+sT)form.

Substitute
intheexpressionfor
Findouttheexpressionsfor
and

Tabulatevariousvaluesofmagnitudeandphaseanglefordifferentvaluesof
.
Thereareusuallyfourkeypointstobeknown.

(a)Thestartingoftheplotwhere
(b)Theendoftheplotwhere

andget

(c)ThepointwherethePolarplotcrossestherealaxis,i.e.,
(d)ThepointwherethePolarplotcrossestheimaginaryaxis,i.e.,
BPUT

startingfrom0to

Page83

ControlSystems

Fix all points in a polargraph sheet and join the points. (Polar graph sheet hasconcentric circles and
radial lines. The concentric circles represent the magnitude and the radial lines represent the phase
angles. Inpolarsheet,+vephaseangleismeasuredinACWfrom00andvephaseangleismeasuredin
CWfrom00

Examples:
PolarPlotof

Considera1stordersystemwithtransferfunction

Thesinusoidaltransferfunctionis

and

When
realaxis. As

Therefore,thephasorat

hasunitlengthandliesalongthepositive

increases,M decreases and phase angle increases negatively. When

As

, M becomes zero and

is 90 This is represented by a phasor of zero length

directed along the 90 axis in the complex plane. In fact, the locus of
semicircle.

can be shown to be a

Im
270

1
-180

45

increasing

1/

-90

PolarPlotof

Polar Plot of 1/ (1+j T)

Considernowthetransferfunction

BPUT

and

Page84

ControlSystems
Thistransferfunctionmayberearrangedas

The general shape of this transfer function is shown below. The plot is asymptotic to the vertical line
passingthroughthepoint(T,0).

PolarPlotof

Thelowandhighfrequencyregionsofthepolarplotofthefollowingsinusoidaltransferfunction

aregivenrespectivelyby
and
The Polar plot of this sinusoidal transfer function starts at
fromzerotoinfinity.Thus,thehighfrequencyportionof

BPUT

and ends at

as

increases

isatangenttothenegativerealaxis.

Page85

ControlSystems

The exact shape of a polar plot depends on the value of the damping ratio
sameforbothoverdampedandunderdampedcase.
For the underdamped case, at

, we have

Therefore,itcanbeseenthatthefrequencyatwhichthe
undamped natural frequency
vectorattheresonantfrequency

. The peak value of

Thus the general shape is

, and the phase angle at

is 90.

locusintersectstheimaginaryaxisisthe
is obtained as the ratio of magnitude of the

tothemagnitudeofthevectorat

For the overdamped case, as increases well beyond unity, the

locus approaches a semicircle..
Thismaybeseenfromthefactthat,foraheavilydampedsystem,thecharacteristicrootsarereal,andone
is much smaller than the other. Since, for sufficiently large , the effect of the larger root (larger in
absolutevalue)ontheresponsebecomesverysmall,thesystembehaveslikea1storderone.

BPUT

Page86

ControlSystems

Example:
ObtainthePolarPlotofthefollowingtransferfunction:

Since

canbewrittenas

Themagnitudeandphaseanglearerespectively

And

Im

1
Re

Since the magnitude decreases from unity monotonically, and the phase angle also decreases
monotonically, and indefinitely, the polar plot of the given transfer function is a spiral, as shown in the
abovefigure.
GeneralNatureofNyquistPlots:
Thepolarplotsofatransferfunctionoftheform

Where,
or the degree of the denominator polynomial is greater than that of the numerator will
havethefollowingshapes.

BPUT

Page87

ControlSystems

Generalshapesofthepolarplotsofsomeimportantfunctions
The general shapes of the polar plots of some important functions are shown below. From the figures,
followingobservationsaremade.
Additionofanonzeropoletoatransferfunctionresultsinfurtherrotationofthepolarplotthroughan
angleof

as

Addition of a pole at the origin to the transferfunction rotates the polar plot at zero and infinite
frequenciesbyafurtherangleof
.
Theeffectofadditionofazerotothetransferfunctionistorotatethehighfrequencyportionofthe
polarplotby

BPUT

inthecounterclockwisedirection.

Ifdegreeofthedenominatorpolynomialisgreaterthanthatofthenumerator,the
lociwill
convergetotheoriginclockwise.
Anycomplicatedshapeofthepolarplotcurvesarecausedbythenumeratordynamics,whichisby
thetimeconstantsinthenumeratorofthetransferfunction.

Page88

ControlSystems
Generalshapesofthepolarplotsofsomeotherimportantfunctions

RELATIVESTABILITYANALYSIS:
Indesigningacontrolsystem,werequirethatthesystembestable.Wealsorequirethatthesystemhas
adequaterelativestability.Theclosenessofapproachofthe

locustothe1+j0pointisanindication

oftherelativestabilityofastablesystem.Ingeneral,wemayexpectthatthecloserthe
locusisto
the 1+j0 point, the larger is the maximum overshoot in the transient response and the larger it takes to
dampout.Whenthe
locuspassesthroughthe1+j0point,thesystemisonthevergeofinstabilityand
exhibits sustained oscillations. The measures of relative stability in the frequency domain are the Gain
MarginandthePhaseMargin.

BPUT

Page89

ControlSystems

Lecture - 33, 34
Stabilityinfrequencydomain:Principleofargument,Nyquiststabilitycriterion

10.Stabilityinfrequencydomain:

Astabilitytestfortimeinvariantlinearsystemscanalsobederivedinthefrequencydomain.Itisknownas
Nyquiststabilitycriterion.
ItisbasedonthecomplexanalysisresultknownasCauchysprincipleofargument.Notethatthesystem
transferfunctionisacomplexfunction.ByapplyingCauchysprincipleofargumenttotheopenloopsystem
transfer function, we will get information about stability of the closedloop system transfer function and
arriveattheNyquiststabilitycriterion(Nyquist,1932).
TheimportanceofNyquiststabilityliesinthefactthatitcanalsobeusedtodeterminetherelativedegree
ofsystemstabilitybyproducingthesocalledphaseandgainstabilitymargins.Thesestabilitymarginsare
neededforfrequencydomaincontrollerdesigntechniques.
We present Only the Essence of the Nyquist stability Criterion and Define the Phase and Gain stability
margins.TheNyquistMethodisusedforstudyingthestabilityoflinearSystemswithPuretimedelay.
ForaSISOfeedbackSystemtheclosedlooptransferfunctionisgivenby:

where
representsthesystemand
isthefeedbackelement.
Since the system poles are determined as those values at which its transfer function becomes infinity, it
followsthattheclosedloopsystempolesareobtainedbysolvingthefollowingequation
which,infact,representstheSystemcharacteristicequation.
Inthefollowingweconsiderthecomplexfunction
Whosezerosaretheclosedlooppolesofthetransferfunction.Inaddition,itiseasytoseethatthepolesof
are the zeros of

. At the same time the poles of

are the openloop control system poles

since they are contributed by the poles of

, which can be considered as the openloop control
systemtransferfunctionobtainedwhenthefeedbackloopisopenatsomepoint.TheNyquiststabilitytest
is obtained by applying the Cauchy principle of argument to the complex function
Cauchysprincipleofargument.

. First, we state

Cauchysprincipleofargument
Let

beananalyticfunctioninaclosedregionofthecomplexplane

finitenumberofpoints(namely,thepolesof
the contour. Then, as
function

Figure4.6),with

BPUT

).Itisalsoassumedthat

travels around the contour in the

encircles the origin in the

giveninFigurebelowexceptata
isanalyticateverypointon

plane in the clockwise direction, the

plane in the same direction

times (see

givenby

Page90

ControlSystems
Where

and

stand for the number of zeros and poles (including their multiplicities) of the function

insidethecontour.
Theaboveresultcanbealsowrittenas
Whichjustifiestheterminologyused,theprincipleofargument.

Figure1.Cauchy'sprincipleofargument

NyquistPlot
TheNyquistplotisapolarplotofthefunction
When

travelsaroundthecontourgiveninFigurebelow.

Figure2.ContourinSplane

Thecontourinthisfigurecoversthewholeunstablehalfplaneofthecomplexplane
function
thepolesof

.Sincethe

,accordingtoCauchysprincipleofargument,mustbeanalyticateverypointonthecontour,
ontheimaginaryaxismustbeencircledbyinfinitesimallysmallsemicircles.

NyquistStabilityCriterion
Itstatesthatthenumberofunstableclosedlooppolesisequaltothenumberofunstableopenlooppoles
plusthenumberofencirclementsoftheoriginoftheNyquistplotofthecomplexfunction.
This can be easily justified by applying Cauchys principle of argument to the function with the plane
contourgiveninFigure2.Notethatandrepresentthenumbersofzerosandpoles,respectively,ofinthe
unstablepartofthecomplexplane.Atthesametime,thezerosofaretheclosedloopsystempoles,and
thepolesofaretheopenloopsystempoles(closedloopzeros).
Theabovecriterioncanbeslightlysimplifiedifinsteadofplottingthefunction,weplotonlythefunction
andcountencirclementoftheNyquistplotofaroundthepoint,sothatthemodifiedNyquistcriterionhas
thefollowingform.

BPUT

Page91

ControlSystems

StabilityviatheNyquistDiagram
WenowusetheNyquistdiagramtodetermineasystem'sstability,usingthesimpleequation.Thevaluesof
P,thenumberofopenlooppolesofG(s)H(s)enclosedbythecontour,andN,thenumberofencirclements
theNyquistdiagrammakesabout1,areusedtodetermineZ,thenumberofrighthalfplanepolesofthe
closedloopsystem.
Iftheclosedloopsystemhasavariablegainintheloop,onequestionwewouldliketoaskis,"Forwhat
rangeofgainisthesystemstable?"Thegeneralapproachistosettheloopgainequaltounityanddraw
the Nyquist diagram. Since gain is simply a multiplying factor, the effect of the gain is to multiply the
resultantbyaconstantanywherealongtheNyquistdiagram.

Figure3.Nyquiststabilitycontouranddiagram

As the gain is varied, we can visualize the Nyquist diagram is expanding (increased gain) or shrinking
(decreasedgain)likeaballoon.ThismotioncouldmovetheNyquistdiagrampastthe1point,changing
thestabilitypicture.Forthissystem,sinceP=2,thecriticalpointmustbeencircledbytheNyquistdiagram
toyieldN=2andastablesystem.AreductioningainwouldplacethecriticalpointoutsidetheNyquist
diagramwhereN=0,yieldingZ=2,anunstablesystem.
IftheNyquistdiagramintersectstherealaxisat1,then
.Fromrootlocusconcepts,when
G(s)H(s)=1,thevariablesisaclosedlooppoleofthesystem.Thus,thefrequencyatwhichtheNyquist
diagram intersects 1 is the same frequency at which the root locus crosses the /coaxis. Hence, the
systemismarginallystableiftheNyquistdiagramintersectstherealaxisat1.
In summary, then, if the openloop system contains a variable gain, K, set K = 1 and sketch the Nyquist
diagram.Considerthecriticalpointtobeat1/Kratherthanat1.AdjustthevalueofKtoyieldstability,
basedupontheNyquistcriterion.

BPUT

Page92

ControlSystems

PROBLEM: Fortheunityfeedbacksystem,whereG(s)=K/[s(s+3)(s+5)],findtherangeofgain,K,for
stability,instability,andthevalueofgainformarginalstability.Formarginalstabilityalsofindthefrequency
ofoscillation.UsetheNyquistcriterion.
SOLUTION: FirstsetK=1andsketchtheNyquistdiagramforthesystem

Figure4.Nyquistcontureandstabilitydiagram

Forallpointsontheimaginaryaxis,

,
At
NextfindthepointwheretheNyquistdiagramintersectsthenegativerealaxis.Settingtheimaginarypart
ofEq.(1)equaltozero,wefind
.
back into Eq. (1) yields the real part of 0.0083. Finally, at
,
Substituting this value of

From the contour of Figure, P= 0; for stabilityN must then beequalto zero. From Figure, the system is
stableifthecriticalpointliesoutsidethecontour(N=0),sothatZ=PN=0.Thus,Kcanbeincreasedby
1/0.0083=120.5beforetheNyquistdiagramencircles1.
Hence,forstability,K<120.5.FormarginalstabilityK=120.5.AtthisgaintheNyquistdiagramintersects
1,andthefrequencyofoscillationis
rad/s

BPUT

Page93

ControlSystems

Axis
StabilityviaMappingOnlythePositive
OncethestabilityofasystemisdeterminedbytheNyquistcriterion,continuedevaluationofthesystem
canbesimplifiedbyusingjustthemappingofthepositive
axis.

Figure5.Contourandrootlocusofsystemthatisstableforsmallgainandunstableforlargegain

ConsiderthesystemshowninaboveFigure,whichisstableatlowvaluesofgainandunstableathighvalues
of gain. Since the contour does not encircle openloop poles, the Nyquist criterion tells us that we must
havenoencirclementsof1forthesystemtobestable.WecanseefromtheNyquistdiagramthatthe
encirclementsofthecriticalpointcanbedeterminedfromthemappingofthepositive
axisalone.If
thegainissmall,themappingwillpasstotherightof1,andthesystemwillbestable.Ifthegainishigh,
themappingwillpasstotheleftof1,andthesystemwillbeunstable.Thus,thissystemisstableforthe
rangeofloopgain,K,thatensuresthatthe openloopmagnitudeislessthanunityatthatfrequencywhere
the phase angle is 180 (or, equivalently, 180). This statement is thus an alternative to the Nyquist
criterionforthissystem.

Figure6.Contourandrootlocusofsystemthatisunstableforsmallgainandstableforlargegain

NowconsiderthesystemshowninaboveFigure,whichisunstableatlowvaluesofgainandstableathigh
valuesofgain.Sincethecontourenclosestwoopenlooppoles,twocounterclockwiseencirclementsofthe
criticalpointarerequiredforstability.Thus,forthiscasethesystemisstableiftheopenloopmagnitudeis
greaterthanunityatthatfrequencywherethephaseangleis180(or,equivalently,180).
Insummary,firstdeterminestabilityfromtheNyquistcriterionandtheNyquistdiagram.Nextinterpretthe
Nyquistcriterionanddeterminewhetherthemappingofjustthepositiveimaginaryaxisshouldhaveagain
BPUT

Page94

ControlSystems
of less than or greater than unity at 180. If the Nyquist diagram crosses 180 at multiple frequencies,
determinetheinterpretationfromtheNyquistcriterion.
PROBLEM: Findtherangeofgainforstabilityandinstability,andthegainformarginalstability,forthe
unityfeedbacksystem,whereG(s)=K/[(s2+2s+2)(s+2)].Formarginalstabilityfindtheradianfrequency
ofoscillation.UsetheNyquistcriterionandthemappingofonlythepositiveimaginaryaxis.
SOLUTION: Sincetheopenlooppolesareonlyinthelefthalfplane,theNyquistcriteriontellsusthatwe
wantnoencirclementsof1forstability.Hence,againlessthanunityat180isrequired.Beginbyletting
K=1anddrawtheportionofthecontouralongthepositiveimaginaryaxisasshowninFigure.

Figure7.NyquistdiagramofMappingofpositiveimaginaryaxis

In Figure, the intersection with the negative real axis is found by letting 5 = jco in G(s)H(s), setting the
imaginarypartequaltozerotofindthefrequency,andthensubstitutingthefrequencyintotherealpartof
G(jco)H{jco).Thus,foranypointonthepositiveimaginaryaxis,

Settingtheimaginarypartequaltozero,wefind

.Substitutingthisvaluebackintoequationyields

therealpart,
This closedloop system is stable if the magnitude of the frequency response is less than unity at 180.
Hence, the system is stable for K < 20, unstable for K > 20, and marginally stable for K = 20. When the
systemismarginallystable,theradianfrequencyofoscillationis .

Example:1
Considerthefollowingtransferfunction

Puttingthevalueof

inaboveequation,weobtain

Themagnitudeandphaseangleequations:

BPUT

Page95

ControlSystems
Evaluatingmagnitudeandphaseresponseat
At

At

and

PHASEANDGAINMARGINTHROUGHNYQUISTPLOTS:

BPUT

Page96

ControlSystems

BPUT

Page97

ControlSystems

Lecturer - 35, 36
Closedloopfrequencyresponse:ConstantMcircles.ConstantNCircles,Nicholschart

Theclosedloopfrequencyresponseisthelocusoftheclosedloopmagnitudefrequencyresponseforunity
feedback system. If the frequency response of an open loop system is plotted in polar coordinates, and
superimposedonthetopofMcircles,thentheclosedloopmagnitudefrequencyresponseisdetermined
byeachintersectionofthispolarplotwiththeconstantMcircles.

McirclesarecontoursofconstantclosedloopmagnitudeonNyquistplane.
Let

.Then

.Hence,

Thentwocasesarepossible:

,soweget:

Constant Mcircles are the circles in the complex plane with radius

centered at

BPUT

Page98

ControlSystems

Figure8.ConstantMcircles

ConstantNcircles

Therefore

Foraconstantvalueof
isalsoconstant.
Rearrangingtheequationweget,

centeredat(1/2,1/2N)(see
constantNcirclesarethecirclesinthecomplexplanewithradius
figure 2). Constant Ncircles are the locus of the closedloop phase frequency response. Similarly to M
circles,ifthefrequencyresponseofanopenloopsystemisplottedinpolarcoordinates,andsuperimposed
onthetopofNcircles,thentheclosedloopphasefrequencyresponseisdeterminedbyeachintersection
ofthispolarplotwiththeconstantNcircles.AlltheconstantNcirclespassthroughtheoriginand(1+j0)
pointregardlessofthevalueofN.

BPUT

Page99

ControlSystems

Figure9.ConstantNcircles

Example 1- Closed-loop frequency response from open-loop frequency response

Find the closedloop frequency response of the unity feedback system with openloop transfer function

usingtheopenlooppolarfrequencyresponsecurve,constantMcircles,andconstantNcircles.

Solution
Openloopfrequencyresponseis

Polarplotof

BPUT

isshownsuperimposedovertheMandNcirclesinfigure3.

Page100

ControlSystems

Figure10.ConstantNandMcircles

Theclosedloopmagnitudefrequencyresponsecanbeobtainedbyfindingtheintersectionofeachpointof
with the Mcircles,and the closedloop phase frequency response can be obtained by finding
the
theintersectionofeachpointofthe
withtheNcircles.

BPUT

Page101

ControlSystems

Figure11.Closedloopfrequencyresponse

NicholsCharts
Sinceitiseasiertoconstructabodeplotthanapolarplot,itispreferabletohaveconstantMandconstant
contoursconstructedonlogarithmicgainandphasecoordinates.N.B.Nicholstransformedtheconstant
Mandconstant contoursconstructedonlogarithmicgainandphasecoordinateandtheresultingchart
is known as the Nichols chart. It displays magnitude response in decibels, so that changes in gain are as
simple to handle as in the Bode plot. Nichols chart is a plot of openloop magnitude in dB vs. openloop
phase.EverypointontheconstantMandNcirclesistransferredtotheNicholschart(seefigure4).The
intersectionofthe
withtheNicholschartyieldsthefrequencyresponseoftheclosedloopsystem.

BPUT

Page102

ControlSystems
Example2ClosedloopfrequencyresponsefromopenloopfrequencyresponseusingNicholschart
Consideraunityfeedbacksystemwiththefollowingopenlooptransferfunction

FindtheclosedloopfrequencyresponseusingNicholschart.

Solution
SuperimposingtheopenloopfrequencyresponseforK=1ontheNicholschart,weobtaintheplotshown
infigure5.

Figure 5: Nichols chart for Example 2

Theintersectionoftheplotof
withtheNicholschartyieldsthefrequencyresponseoftheclosedloop
system
Ifthegainisincreasedby10dB,oneshouldsimplyraisethecurveforK=1by10dBtoobtainthecurvefor
K=3.16(10dB)(seefigure5).

BPUT

Page103

ControlSystems
Nichols Chart for elementary systems
Bode

Nyquist

Nichols

BPUT

Page104

ControlSystems

Lecture - 37, 38
Controllers:ConceptofProportional,Derivative,IntegralControlactions,P,PD

11.Controllers:ConceptofProportional,DerivativeandIntegralControlactions:

Avarietyofcontrolsareusedtomanipulateprocesses,howeverthemostsimpleandoftenmosteffective
isthePIDcontroller.Muchmorepracticalthanthetypicalon/offcontroller,PIDcontrollersallowformuch
better adjustments to be made in the system. While this is true, there are some advantages to using an
on/offcontroller:
Relativelysimpletodesignandexecute
Binary sensors and actuators (such as an on/off controller) are generally more reliable and less
expensive
Althoughtherearesomeadvantages,therearelargedisadvantagestousinganon/offcontrollerscheme:
Inefficient(usingthiscontrolislikedrivingwithfullgasandfullbreaks)
Cangeneratenoisewhenseekingstability(candramaticallyovershootorundershootasetpoint)
Physicallywearingonvalvesandswitches(continuouslyturningvalves/switchesfullyonandfullyoff
causesthemtobecomewornoutmuchquicker)

To allow for much better control and finetuning adjustments, most industrial processes use a PID
controllerscheme.

Thecontrollerattemptstocorrecttheerrorbetweenameasuredprocessvariableanddesiredsetpointby
calculatingthedifferenceandthenperformingacorrectiveactiontoadjusttheprocessaccordingly.APID
controller controls a process through three parameters: Proportional (P), Integral (I), and Derivative (D).
Theseparameterscanbeweighted,ortuned,toadjusttheireffectontheprocess.Thefollowingsection
willprovideabriefintroductiononPIDcontrollers.
TheProcessGain(K)istheratioofchangeoftheoutputvariable(respondingvariable)tothechangeofthe
input variable (forcing function). It specifically defines the sensitivity of the output variable to a given
changeintheinputvariable.

Gaincanonlybedescribedasasteadystateparameterandgivenoknowledgeaboutthedynamicsofthe
processandisindependentofthedesignandoperatingvariables.Againhasthreecomponentsthatinclude
the sign, the value, and the units. The sign indicates how the output responds to the process input. A
positivesignshowsthattheoutputvariableincreaseswithanincreaseintheinputvariableandanegative
signshowsthattheoutputvariabledecreaseswithanincreaseintheinputvariable.Theunitsdependon
theprocessconsideredthatdependonthevariablesmentioned.
As previously mentioned, controllers vary in the way they correlate the controller input (error) to the
controller output (actuating signal). The most commonly used controllers are the proportional integral
BPUT

Page105

ControlSystems
derivative(PID)controllers.PIDcontrollersrelatetheerrortotheactuatingsignaleitherinaproportional
(P),integral(I),orderivative(D)manner.PIDcontrollerscanalsorelatetheerrortotheactuatingsignal
usingacombinationofthesecontrols.
Proportional(P)Control
Proportionalcontrolisthesimplestformofcontinuouscontrolthatcanbeusedinaclosedloopedsystem.
Ponlycontrolminimizesthefluctuationintheprocessvariable,butitdoesnotalwaysbringthesystemto
thedesiredsetpoint.Thisdeviationisknownastheoffset,anditisusuallynotdesiredinaprocess.The
existenceofanoffsetimpliesthatthesystemcouldnotbemaintainedatthedesiredsetpointatsteady
state.Itisanalogoustothesystematicerrorinacalibrationcurve,wherethereisalwaysaset,constant
error that prevents the line from crossing the origin. The offset can be minimized by combining Ponly
controlwithanotherformofcontrol,suchasIorDcontrol.
MathematicalEquations
Pcontrol linearly correlates the controller output (actuating signal) to the error (difference between
measuredsignalandsetpoint).ThisPcontrolbehaviorismathematicallyillustratedinEquation1.
c(t) = Kce(t)+b

(1)

Ascanbeseenfromtheaboveequation,Ponlycontrolprovidesalinearrelationshipbetweentheerrorof
asystemandthecontrolleroutputofthesystem.. Combinedwiththebias,thisalgorithmdeterminesthe
action that the controller should take. A graphical representation of the Pcontroller output for a step
increaseininputattimet0isshownbelowinFigure2.Thisgraphisexactlysimilartothestepinputgraph
itself.

BPUT

Page106

ControlSystems

Figure 2. P-controller output for step input.

Integral(I)Control
Integral control is a second form of feedback control. It is often used because it is able to remove any
deviationsthatmayexist.Thus,thesystemreturnstobothsteadystateanditsoriginalsetting.Anegative
errorwillcausethesignaltothesystemtodecrease,whileapositiveerrorwillcausethesignaltoincrease.
However,IonlycontrollersaremuchslowerintheirresponsetimethanPonlycontrollersbecausethey
are dependent on more parameters. If it is essential to have no offset in the system, then an Ionly
controller should be used, but it will require a slower response time. This slower response time can be
reducedbycombiningIonlycontrolwithanotherform,suchasPorPDcontrol.Thephilosophybehindthe
integralcontrolisthatdeviationswillbeaffectedinproportiontothecumulativesumoftheirmagnitude.
The key advantage of adding a Icontrol to your controller is that it will eliminate the offset. The
disadvantagesarethatitcandestabilizethecontroller,andthereisanintegratorwindup,whichincreases
thetimeittakesforthecontrollertomakechanges.
Mathematical Equations
Icontrolcorrelatesthecontrolleroutputtotheintegraloftheerror.Theintegraloftheerroristakenwith
respecttotime.Itisthetotalerrorassociatedoveraspecifiedamountoftime.ThisIcontrolbehavioris
mathematicallyillustratedinEquation2.

(2)

BPUT

Page107

ControlSystems

Steady state error for a Step input -

Inthisequation,theintegraltimeistheamountoftimethatittakesforthecontrollertochangeitsoutput
byavalueequaltotheerror.Thecontrolleroutputbeforeintegrationisequaltoeithertheinitialoutputat
timet=0,orthecontrolleroutputatthetimeonestepbeforethemeasurement.Asexpected,thisgraph
representstheareaunderthestepinputgraph.

Figure 3. I-controller output for step input.

Derivative(D)Control
Unlike Ponly and Ionly controls, Dcontrol is a form of feed forward control. Dcontrol anticipates the
process conditions by analyzing the change in error. It functions to minimize the change of error, thus
keepingthesystemataconsistentsetting.TheprimarybenefitofDcontrollersistoresistchangeinthe
system,themostimportantofthesebeingoscillations.Thecontroloutputiscalculatedbasedontherate
of change of the error with time. The larger the rate of the change in error, the more pronounced the
controllerresponsewillbe.
Unlike proportional and integral controllers, derivative controllers do not guide the system to a steady
state.Becauseofthisproperty,DcontrollersmustbecoupledwithP,IorPIcontrollerstoproperlycontrol
thesystem.
Mathematical Equations

BPUT

Page108

ControlSystems
Dcontrolcorrelatesthecontrolleroutputtothederivativeoftheerror.Thederivativeoftheerroristaken
with respect to time. It is the change in errorassociated with change in time. This Dcontrol behavior is
mathematicallyillustratedinEquation3.
(3)

AgraphicalrepresentationoftheDcontrolleroutputforastepincreaseininputattimet0isshownbelow
inFigure4.Asexpected,thisgraphrepresentsthederivativeofthestepinputgraph.

Figure 4. D-controller output for step input.

BPUT

Page109

ControlSystems

Lecture - 39, 40
PI,PIDcontrollers,ZeiglerNicholsmethodoftuningPIDcontrollers
Proportional-Integral (PI) Control
One combination is the PIcontrol, which lacks the Dcontrol of the PID system. PI control is a form of
feedback control. It provides a faster response time than Ionly control due to the addition of the
proportionalaction.PIcontrolstopsthesystemfromfluctuating,anditisalsoabletoreturnthesystemto
its set point. Although the response time for PIcontrol is faster than Ionly control, it is still up to 50%
slower than Ponly control. Therefore, in order to increase response time, PI control is often combined
withDonlycontrol.
Mathematical Equations
PIcontrol correlates the controller output to the error and the integral of the error. This PIcontrol
behaviorismathematicallyillustratedinEquation4.

(4)

Inthisequation,theintegraltimeisthetimerequiredfortheIonlyportionofthecontrollertomatchthe
controlprovidedbythePonlypartofthecontroller.
BPUT

Page110

ControlSystems
The PIcontroller can also be seen as a combination of the Ponly and Ionly control equations. The bias
terminthePonlycontrolisequaltotheintegralactionoftheIonlycontrol.ThePonlycontrolisonlyin
actionwhenthesystemisnotatthesetpoint.Whenthesystemisatthesetpoint,theerrorisequalto
zero,andthefirsttermdropsoutoftheequation.ThesystemisthenbeingcontrolledonlybytheIonly
portion of the controller. Should the system deviate from the set point again, Ponly control will be
enacted. A graphical representation of the PIcontroller output for a step increase in input at time is
shownbelowinFigure5.

Figure 5. PI-controller output for step input.

Effects of Kc and Ti
With a PI control system, controller activity (aggressiveness) increases as Kc and Ti decreases, however
theycanactindividuallyontheaggressivenessofacontrollersresponse.ConsiderFigure6belowwiththe
centergraphbeingalinearsecondordersystembasecase.

BPUT

Page111

ControlSystems

Figure 6. Effects of Kc and Ti [2]

TheplotdepictshowTiandKcbothaffecttheperformanceofasystem,whethertheyarebothaffectingit
oreachoneisindependentlydoingso.Regardlessofintegraltime,increasingcontrollergain(movingform
bottomtotopontheplot)willincreasecontrolleractivity.Similarly,decreasingintegraltime(movingright
toleftontheplot)willincreasecontrolleractivityindependentofcontrollergain.Asexpected,increasing
KcanddecreasingTiwouldcompoundsensitivityandcreatethemostaggressivecontrollerscenario.
Another noteworthy observation is the plot with a normal Kc and double Ti. The plot depicts how the
proportional term is practical but the integral is not receiving enough weight initially, causing the slight
oscillationbeforetheintegraltermcanfinallycatchupandhelpthesystemtowardsthesetpoint.

Proportional-Derivative (PD) Control

BPUT

Page112

ControlSystems
AnothercombinationofcontrolsisthePDcontrol,whichlackstheIcontrolofthePIDsystem.PDcontrol
is combination of feedforward and feedback control, because it operates on both the current process
conditionsandpredictedprocessconditions.InPDcontrol,thecontroloutputisalinearcombinationof
the error signal and its derivative. PDcontrol contains the proportional controls damping of the
fluctuationandthederivativecontrolspredictionofprocesserror.
Mathematical Equations
Asmentioned,PDcontrolcorrelatesthecontrolleroutputtotheerrorandthederivativeoftheerror.This
PDcontrolbehaviorismathematicallyillustratedinEquation5.

(5)

TheequationindicatesthatthePDcontrolleroperateslikeasimplifiedPIDcontrollerwithazerointegral
term.Alternatively,thePDcontrollercanalsobeseenasacombinationofthePonlyandDonlycontrol
equations. In this control, the purpose of the Donly control is to predict the error in order to increase
stabilityoftheclosedloopsystem.PDcontrolisnotcommonlyusedbecauseofthelackoftheintegral
term.Withouttheintegralterm,theerrorinsteadystateoperationisnotminimized.PDcontrolisusually
usedinbatchpHcontrolloops,whereerrorinsteadystateoperationdoesnotneedtobeminimized.In
thisapplication,theerrorisrelatedtotheactuatingsignalboththroughtheproportionalandderivative
term.AgraphicalrepresentationofthePDcontrolleroutputforastepincreaseininputattimet0isshown
belowinFigure6.Again,thisgraphisacombinationofthePonlyandDonlygraphs,asexpected.

BPUT

Page113

ControlSystems

Figure 12. PD-controller output for step input.

ProportionalIntegralDerivative(PID)Control
Proportionalintegralderivativecontrolisacombinationofallthreetypesofcontrolmethods.PIDcontrol
ismostcommonlyusedbecauseitcombinestheadvantagesofeachtypeofcontrol.Thisincludesaquicker
response time because of the Ponly control, along with the decreased/zero offset from the combined
derivativeandintegralcontrollers.ThisoffsetwasremovedbyadditionallyusingtheIcontrol.Theaddition
of Dcontrol greatly increases the controller's response when used in combination because it predicts
disturbancestothesystembymeasuringthechangeinerror.Onthecontrary,asmentionedpreviously,
whenusedindividually,ithasaslowerresponsetimecomparedtothequickerPonlycontrol.However,
although the PID controller seems to be the most adequate controller, it is also the most expensive
controller.Therefore,itisnotusedunlesstheprocessrequirestheaccuracyandstabilityprovidedbythe
PIDcontroller.
Mathematical Equations
PIDcontrolcorrelatesthecontrolleroutputtotheerror,integraloftheerror,andderivativeoftheerror.
ThisPIDcontrolbehaviorismathematicallyillustratedinEquation6.

(6)

BPUT

Page114

ControlSystems

As shown in the above equation, PID control is the combination of all three types of control. In this
equation,thegainismultipliedwiththeintegralandderivativeterms,alongwiththeproportionalterm,
because in PID combination control, the gain affects the I and D actions as well. Because of the use of
derivativecontrol,PIDcontrolcannotbeusedinprocesseswherethereisalotofnoise,sincethenoise
wouldinterferewiththepredictive,feedforwardaspect.AgraphicalrepresentationofthePIDcontroller
output for a step increase in input at time t0 is shown below in Figure 7. This graph resembles the
qualitativecombinationofthePonly,Ionly,andDonlygraphs.

Figure 7. PID-controller output for step input.

In addition to PIDcontrol, the P, I, and D controls can be combined in other ways. These alternative
combinationsaresimplificationsofthePIDcontrol.
SummaryTables
A summary of the advantages and disadvantages of the three controls is shown below is shown in Table 1.
Table 1. Advantages and disadvantages of controls
Proportional (P)
Integral (I)
BPUT

Derivative (D)
Page115

ControlSystems
Advantages

Fast response time

Reduces
error.

steady

state Improves stability

Minimizes fluctuation
Disadvantages Contains large offset

Slow response time

Does not bring system to Reduces stability margins

desired set point

Controls
process
with
rapidly changing outputs
Highly sensitive to noise
Requires combined use
with another controller

EffectsofCoefficients:
Parameter SpeedofResponse Stability Accuracy
IncreasingK
Increases
Deteriorate Improves
IncreasingKi
Decreases
Deteriorate Improves
IncreasingKd
Increases
Improves Noimpact

Whatistuning?
Tuning is adjustment of control parameters to the optimum values for the desired control response.
Stabilityisabasicrequirement.However,differentsystemshavedifferentbehavior,differentapplications
havedifferentrequirements,andrequirementsmayconflictwithoneanother.
ZieglerNichols tuning method:
This method was introduced by John G. Ziegler and Nathaniel B. Nichols in the 1940s. The Ziegler
Nichols closed loop method is based on experiments executed on an established control loop (a real
systemorasimulatedsystem).

ClosedLoop(FeedbackLoop)
1. Removeintegralandderivativeaction.Setintegraltime(Ti)to999oritslargestvalueandsetthe
derivativecontroller(Td)tozero.
2. Createasmalldisturbanceintheloopbychangingthesetpoint.Adjusttheproportional,increasing
and/ordecreasing,thegainuntiltheoscillationshaveconstantamplitude.
3. Recordthegainvalue(Ku)andperiodofoscillation(Pu).
4. PlugthesevaluesintotheZieglerNicholsclosedloopequationsanddeterminethenecessary
settingsforthecontroller.

BPUT

Page116

ControlSystems

Figure 1. System tuned using the Ziegler-Nichols closed-loop tuning method

P
PI
PID

Advantages
1. Easyexperiment;onlyneedtochangethePcontroller
2. Includesdynamicsofwholeprocess,whichgivesamoreaccuratepictureofhowthesystemis
behaving
Disadvantages
1. Experimentcanbetimeconsuming
2. CanventureintounstableregionswhiletestingthePcontroller,whichcouldcausethesystemto
becomeoutofcontrol

BPUT

Page117

ControlSystems
ZieglerNicholsOpenLoopTuningMethodorProcessReactionMethod
This method remains a popular technique for tuning controllers that use proportional, integral, and
derivativeactions.TheZieglerNicholsopenloopmethodisalsoreferredtoasaprocessreactionmethod,
becauseitteststheopenloopreactionoftheprocesstoachangeinthecontrolvariableoutput.Thisbasic
test requires that the response of the system be recorded, preferably by a plotter or computer. Once
certain process response values are found, they can be plugged into the ZieglerNichols equation with
specificmultiplierconstantsforthegainsofacontrollerwitheitherP,PI,orPIDactions.
Inthismethod,thevariablesbeingmeasuredarethoseofasystemthatisalreadyinplace.Adisturbanceis
introducedintothesystemanddatacanthenbeobtainedfromthiscurve.Firstthesystemisallowedto
reach steady state, and then a disturbance, Xo, is introduced to it. The percentage of disturbance to the
systemcanbeintroducedbyachangeineitherthesetpointorprocessvariable.Forexample,ifyouhavea
thermometerinwhichyoucanonlyturnitupordownby10degrees,thenraisingthetemperatureby1
degree would be a 10% disturbance to the system. These types of curves are obtained in open loop
systems when there is no control of the system, allowing the disturbance to be recorded. The process
reactioncurvemethodusuallyproducesaresponsetoastepfunctionchangeforwhichseveralparameters
may be measured which include: transportation lag or dead time, dead, the time for the response to
change,,andtheultimatevaluethattheresponsereachesatsteadystate,Mu.

dead=transportationlagordeadtime:thetimetakenfromthemomentthedisturbancewasintroduced
tothefirstsignofchangeintheoutputsignal

=thetimefortheresponsetooccur
Xo=thesizeofthestepchange
Mu=thevaluethattheresponsegoestoasthesystemreturnstosteadystate

An example for determining these parameters for a typical process response curve to a step change is
shownbelow.
Inordertofindthevaluesfordeadand,alineisdrawnatthepointofinflectionthatistangenttothe
responsecurveandthenthesevaluesarefoundfromthegraph.

BPUT

Page118

ControlSystems

TousetheZieglerNicholsopenlooptuningmethod,youmustperformthefollowingsteps:
1. Makeanopenloopsteptest
2. From the process reaction curve determine the transportation lag or dead time, dead, the time
constantortimefortheresponsetochange,,andtheultimatevaluethattheresponsereachesat
steadystate,Mu,forastepchangeofXo.

3. Determinethelooptuningconstants.PluginthereactionrateandlagtimevaluestotheZiegler
NicholsopenlooptuningequationsfortheappropriatecontrollerP,PI,orPIDtocalculatethe
controllerconstants.Usethetablebelow.
Table 2. Open-Loop Calculations of Kc, Ti, Td
P
PI
PID

BPUT

Page119

ControlSystems
Advantages
1. Quickandeasiertousethanothermethods
2. Itisarobustandpopularmethod
3. Ofthesetwotechniques,theProcessReactionMethodistheeasiestandleastdisruptiveto
implement
Disadvantages
1. ItdependsuponpurelyproportionalmeasurementtoestimateIandDcontrollers.
2. ApproximationsfortheKc,Ti,andTdvaluesmightnotbeentirelyaccuratefordifferentsystems.
3. ItdoesnotholdforI,DandPDcontrollers

Example1
Problem
You're a controls engineer working for Flawless Design company when your optimal controller breaks
down.Asabackup,youfigurethatbyusingcoarseknowledgeofaclassicalmethod,youmaybeableto
sustaindevelopmentoftheproduct.Afteradjustingthegaintoonesetofdatatakenfromacontroller,
youfindthatyourultimategainis4.3289.
Fromtheadjustedplotbelow,determinethetypeofloopthisgraphrepresents;then,pleasecalculateKc,
Ti,andTdforallthreetypesofcontrollers.

Solution
Fromthefactthatthisgraphoscillatesandisnotastepfunction,weseethatthisisaclosedloop.Thus,
thevalueswillbecalculatedaccordingly.
BPUT

Page120

ControlSystems
We'regiventheUltimategain,Ku=4.3289.Fromthegraphbelow,weseethattheultimateperiodatthis
gainisPu=6.28

From this, we can calculate the Kc, Ti, and Td for all three types of controllers. The results are tabulated
below.(ResultswerecalculatedfromtheZieglerNicholsclosedloopequations.)
4.3289
6.28

P 2.1645
PI 1.9677 5.2333
PID 2.5464 3.14 0.785

Example2
Problem
YourpartnerfindsanothersetofdataafterthecontrollerbreaksdownanddecidestousetheCohenCoon
methodbecauseoftheslowresponsetimeforthesystem.Theyalsonoticedthatthecontroldial,which
goesfrom08,wassetat3insteadof1.Luckilytheresponsecurvewasobtainedearlierandisillustrated
below.FromthisdatahewantedtocalculateKc,TiandTd.Helphimtodeterminethesevalues.Notethat
theyaxisispercentchangeintheprocessvariable.

BPUT

Page121

ControlSystems

Solution
In order to solve for Kc, Ti and Td, you must first determine L, Cp, and T. All of these values may be
calculatedbythereactioncurvegiven.

Fromtheprocessreactioncurvewecanfindthat:
L=3
T=11
Cp=0.55(55%)

NowthatthesethreevalueshavebeenfoundNandRmaybecalculatedusingtheequationsbelow.

BPUT

Page122

ControlSystems
Usingtheseequationsyoufindthat
N=.05
R=0.27
Wealsoknowthatsincethecontrollerwasmovedfrom1to3,soa200%change.
P=2.00
WeusethesevaluestocalculateKc,Ti,andTd,forthethreetypesofcontrollers
P 14.53
PI 12.3 6.42
PID 18.68 6.65 1.04

BPUT

Page123