EEE2012 Control Notes
description
Transcript of EEE2012 Control Notes
![Page 1: EEE2012 Control Notes](https://reader034.fdocuments.us/reader034/viewer/2022050820/5695d2061a28ab9b0298d7b3/html5/thumbnails/1.jpg)
EEE2012: Control.
1st and 2nd order systems page 1
Chapter 1. Dynamic Systems and Control
Dynamic systems and their control can be found in any branch of engineering or science; electrical, mechanical, chemical, biological, ecological and financial systems to name a few. In all cases one aims to simplify the system to allow its’ representation in a mathematical form. A set of common terms can be found, regardless of the branch of science; Controlled Variable: is the output of the system that is measured and controlled. Control Signal: is the signal or quantity that is varied by the system controller to affect the value of the output. Plant: The plant is the piece of equipment which performs a particular operation. Any physical object to be controlled is called a plant.
System: is a set of objects/elements that are connected or related to each other in such a way that they create and hence define a unity that performs a certain objective.
Disturbance: A disturbance is a signal that adversely affects the value of the output of a system. This may be internal or external.
Feedback control: is an operation that reduces the difference between the output of a system and the reference input.
Dynamics: Properties describing the behaviour of the system in response to an input.
It is our aim to study, analyse and ultimately to control a system to produce a “satisfactory” performance. We would like the system to be controlled automatically, that is without the involvement or intervention of a human beyond giving an input demand.
![Page 2: EEE2012 Control Notes](https://reader034.fdocuments.us/reader034/viewer/2022050820/5695d2061a28ab9b0298d7b3/html5/thumbnails/2.jpg)
EEE2012: Control.
1st and 2nd order systems page 2
a) System model
The model is derived from the set of Ordinary Differential Equations (ODE) describing the system. The differential equations are obtained by analyzing the physical processes of the system. This applies equally well to electrical, mechanical, thermodynamic, fluid dynamic……. systems. In this course we will concentrate on electrical and mechanical system modeling and analysis. A couple of examples are given.
Electrical system
For the electrical system below:
Applying kirchoff’s law gives the equation;
∫
And as ∫
So the electrical system is defined by a 2nd order differential equation.
V
L R
C
![Page 3: EEE2012 Control Notes](https://reader034.fdocuments.us/reader034/viewer/2022050820/5695d2061a28ab9b0298d7b3/html5/thumbnails/3.jpg)
EEE2012: Control.
1st and 2nd order systems page 3
Mechanical system
For a mass being pushed along a surface;
f(t)
x(t)
friction
The sum of forces = mass x acceleration
∑
Applied force – Force due to friction = Mass x Acceleration
and
Therefore the mechanical system can be expressed by the 2nd order differential
equation;
![Page 4: EEE2012 Control Notes](https://reader034.fdocuments.us/reader034/viewer/2022050820/5695d2061a28ab9b0298d7b3/html5/thumbnails/4.jpg)
EEE2012: Control.
1st and 2nd order systems page 4
Tutorial 1 An electrical machine is modeled by the system diagram show below;
V=voltage, i=current, R=resistance, L=inductance, E=motor EMF, T=torque, ω=angular velocity,
J=inertia, B=friction coefficient, Kt=motor torque constant, Ke=motor EMF constant.
1. Express the electrical half of the system mathematically using Kirchoff’s law.
2. Express the mechanical half of the system by summing the moments of force
(consider torque T, Friction B, inertia J)
3. If the motor back EMF E=keω and the torque produced T=Kti combine both
systems into one single differential equation
1.
2.
3. Rearrange 2:
Sub into 1:
V
i
ω T
R L
B
J E
Mechanical system Electrical system
![Page 5: EEE2012 Control Notes](https://reader034.fdocuments.us/reader034/viewer/2022050820/5695d2061a28ab9b0298d7b3/html5/thumbnails/5.jpg)
EEE2012: Control.
1st and 2nd order systems page 5
b) First order Ordinary Differential Equations
First derivative of x is a function of x and time
),( txfdt
dx
General form of an ODE:
A(t)x’+b(t)x=c(t) Non autonomous (time dependent)
Ax’+bx=c Autonomous (independent of time)
Both forms can be solved using the same methods.
Analytic methods to study first order ODEs:
There is an explicit formula for x(t) (a solution which is obtained by the separation of variables or by the use of an integrating factor) which satisfies
),( txfdt
dx
Separation of variables;
Cattxadtdxadt
dx => INFINITE curves (for all Initial
Conditions (ICs)).
x(t) is called solution of the system which is described by adt
dx .
![Page 6: EEE2012 Control Notes](https://reader034.fdocuments.us/reader034/viewer/2022050820/5695d2061a28ab9b0298d7b3/html5/thumbnails/6.jpg)
EEE2012: Control.
1st and 2nd order systems page 6
Integrating factor:
ktconstk
kdtee
uexeuekxxe ktktktkt ''
udtedtxe ktkt '
ceudteexcudtexe ktktktktkt
or:
tktktkt udteexex
0
110
Assuming that k>0 the first part is called transient and the second is called
steady state solution.
![Page 7: EEE2012 Control Notes](https://reader034.fdocuments.us/reader034/viewer/2022050820/5695d2061a28ab9b0298d7b3/html5/thumbnails/7.jpg)
EEE2012: Control.
1st and 2nd order systems page 7
Tutorial 2a.
Solve
using the separation of variables
SOLUTION:
In standard form:
Let ⁄
⁄
Then we have:
Separating the variables: ∫
∫ and solving gives;
( ) (equ 1)
Substituting in the initial conditions, :
( )
Therefore
( ) (equ 2)
Substituting equation 2 into equation 1 and rearranging gives;
( )
( ) or (
)
Raising to the power e;
(
) so
Making i the subject of the equation;
(
) and simplifying
( )
Substitute back in ⁄
⁄ :
⁄
⁄(
⁄ )
⁄ so;
(
⁄ )
⁄
![Page 8: EEE2012 Control Notes](https://reader034.fdocuments.us/reader034/viewer/2022050820/5695d2061a28ab9b0298d7b3/html5/thumbnails/8.jpg)
EEE2012: Control.
1st and 2nd order systems page 8
Tutorial 2b.
Solve
using an integrating factor
SOLUTION;
In standard form:
Let ⁄
⁄
Using the integrating factor;
∫ where ⁄ , substitute this into the 1st ODE
⁄ (
)
⁄ (
)
as
(
⁄ )
⁄
then
(
⁄ )
⁄ (
)
and integrating both sides;
(
⁄ ) (
) ∫
⁄ then;
(
⁄ )
⁄ +C multiplied by ⁄ on both sides allows simplification to;
⁄ (equ 1)
then substitute in the initial conditions :
⁄
substituting back into equation 1;
(
)
⁄
So
(
⁄ )
⁄
![Page 9: EEE2012 Control Notes](https://reader034.fdocuments.us/reader034/viewer/2022050820/5695d2061a28ab9b0298d7b3/html5/thumbnails/9.jpg)
EEE2012: Control.
1st and 2nd order systems page 9
c) Numerical solution of a first order ODE
Numerical solution of using Simulink;
Rearrange making the highest differential the subject, i.e.
Define x’ and integrate to give x.
Connect blocks with summation and correct gains.
Tutorial 3 Using Simulink to solve each of the following;
1.
2.
3.
4.
For each sketch the output over time and describe the system behaviour (stability
and steady state value).
SOLUTIONS: (1) Stable and steady state value = 1/6 (2) Unstable. (3) Stable and
steady state value = -1/6 (4) Zero input therefore zero output
Scope
1
s
Integrator
k/A
Gain
u/A
Constant
x' x
![Page 10: EEE2012 Control Notes](https://reader034.fdocuments.us/reader034/viewer/2022050820/5695d2061a28ab9b0298d7b3/html5/thumbnails/10.jpg)
EEE2012: Control.
1st and 2nd order systems page 10
d) Response to a sinusoidal input
When a first (or 2nd) order system is exposed to a periodic input, the output will clearly not be constant. It will follow the same periodic nature. If a first order system is exposed to a periodic input;
( )
The solution to the 1st order differential equation takes the form;
ktj
t
k
e
k
yk
1
2
2
tan
2
2tancos
1
1
1
1Re~Re
1
Where the periodic input is still evident
But altered by a magnitude change and a phase shift
![Page 11: EEE2012 Control Notes](https://reader034.fdocuments.us/reader034/viewer/2022050820/5695d2061a28ab9b0298d7b3/html5/thumbnails/11.jpg)
EEE2012: Control.
1st and 2nd order systems page 11
Tutorial 4
Using Simulink show the effect of a periodic input on a first order system;
( )
Compare the numerical solution of the above to a Simulink model of the equation
kt
k
1
2
2tancos
1
1
Let k=1 and y(0)=0.1
The model below will help you
Sine Wave
Scope3
Scope2
Scope1
Scope
Product1Product
eu
Math
Function1
eu
Math
Function
1
s
Integrator1
1
s
Integrator
0.1
Gain3
k
Gain2
-k
Gain1
k
Gain
Clock
total
total
steady state
Transient
0 2 4 6 8 10-0.2
-0.15
-0.1
-0.05
0
0.05
0.1
0.15
0.2
0.25
TransientSteadystate Overall
![Page 12: EEE2012 Control Notes](https://reader034.fdocuments.us/reader034/viewer/2022050820/5695d2061a28ab9b0298d7b3/html5/thumbnails/12.jpg)
EEE2012: Control.
1st and 2nd order systems page 12
0 1 2 3 40
0.02
0.04
0.06
0.08
0.1
e) First order ODE summary
First order systems have an underlying exponential growth or decay
The system is considered stable if the output reaches a steady state.
Dynamics described by time to reach a percentage of steady state and stability or lack of.
Output follows input after transient with change in magnitude (and phase)
Steady state value
↑
50% rise time
![Page 13: EEE2012 Control Notes](https://reader034.fdocuments.us/reader034/viewer/2022050820/5695d2061a28ab9b0298d7b3/html5/thumbnails/13.jpg)
EEE2012: Control.
1st and 2nd order systems page 13
f) Second order Ordinary Differential Equations
Second derivative of x is a function of the first derivative of x, x and time
),,'(2
2
txxfdt
xd
Second order linear ODEs with constant coefficients: uBxAxx '''
u=0 => Homogeneous ODE; I need two “representative solutions”
0''' BxAxx , assume rtex => rtrex ' & rterx 2'' =>
00''' 2 rtrtrt BeAreerBxAxx
02 BArr ; Characteristic equation => Check its roots.
2
42 BAAr
![Page 14: EEE2012 Control Notes](https://reader034.fdocuments.us/reader034/viewer/2022050820/5695d2061a28ab9b0298d7b3/html5/thumbnails/14.jpg)
EEE2012: Control.
1st and 2nd order systems page 14
Roots are real and unequal roots r1 and r2 (A2>4B Overdamped system)
trex 1
1 and trex 2
2 are solutions of the ODE =>
trtreCeCxCxCx 21
212211 . If r1 and r2<0 then 0x .
Example 01303403'4'' 2 rrrrxxx
tt eCeCx 23
1 . Assume that 10 x and 00' x :
10 21 CCx and tt eCeCx 23
13' 030' 21 CCx => C1=-0.5,
C2=3/2 => tt eex 2
35.0 3 :
Scope3
Scope2
Scope1
Scope
eu
Math
Function1
eu
Math
Function
1
s
Integrator1
1
s
Integrator
3/2
Gain5
-1
Gain4
-0.5
Gain3
-3
Gain2
-4
Gain1
-3
Gain
0
Constant
Clock1
Clock
xDxD2x
x1
x2
![Page 15: EEE2012 Control Notes](https://reader034.fdocuments.us/reader034/viewer/2022050820/5695d2061a28ab9b0298d7b3/html5/thumbnails/15.jpg)
EEE2012: Control.
1st and 2nd order systems page 15
Roots are real and equal: r1=r2 (A2=4B Critically damped system)rtex 1
and rttex 2 => rtrt teCeCxCxCx 212211
Example:
A=2, B=1, x(0)=1, x’(0)=0 => c1=c2=1
Scope3
Scope2
Scope1
Scope
Product
eu
Math
Function
1
s
Integrator1
1
s
Integrator
-1
Gain2
2
Gain1
Clock
t
t exp(-t)texp(t-)
0 1 2 3 4 5 6 -0.5
0
0.5
1
1.5 Overall solution
x 2
x 1
![Page 16: EEE2012 Control Notes](https://reader034.fdocuments.us/reader034/viewer/2022050820/5695d2061a28ab9b0298d7b3/html5/thumbnails/16.jpg)
EEE2012: Control.
1st and 2nd order systems page 16
Roots are complex: r=a+bj (A2<4B Under damped system A0)
So )sin()cos( btjbteeeeeex atjbtatbjtattbjart =Re+jIm.
Theorem: If x is a complex solution to a real ODE then Re(x) and Im(x) are the real
solutions of the ODE:
)sin(),cos( 21 btexbtex atat =>
btGebtcbtce
btecbtecxcxcx
atat
atat
cos)sin()cos(
)sin()cos(
21
212211
where
1
21
1
21
1 tan&,
tancosc
c
cc
cG
A=1, B=1, x(0)=1, x’(0)=0 => c1=1, c2=1/sqrt(3)
0 2 4 6 8 100
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1
e-t
te-t
overall
![Page 17: EEE2012 Control Notes](https://reader034.fdocuments.us/reader034/viewer/2022050820/5695d2061a28ab9b0298d7b3/html5/thumbnails/17.jpg)
EEE2012: Control.
1st and 2nd order systems page 17
cos
Trigonometric
Function1
sin
Trigonometric
Function
Scope5
Scope4
Scope3
Scope2
Scope1
Scope
Product
eu
Math
Function
1
s
Integrator1
1
s
Integrator
1/sqrt(3)
Gain5
sqrt(3)/2
Gain4
1
Gain3
-0.5
Gain2
1
Gain1
Clockt
bt sin(bt)
cos(bt)
exp(-at)
0 2 4 6 8 10-1.5
-1
-0.5
0
0.5
1
1.5
C1cos(bt)
C2sin(bt)
C1cos(bt)+C
2sin(bt)
0 2 4 6 8 10-1.5
-1
-0.5
0
0.5
1
1.5
C1cos(bt)+C
2sin(bt)
eat
Overall
![Page 18: EEE2012 Control Notes](https://reader034.fdocuments.us/reader034/viewer/2022050820/5695d2061a28ab9b0298d7b3/html5/thumbnails/18.jpg)
EEE2012: Control.
1st and 2nd order systems page 18
Undamped system (A=0)
BrBeerBxx rtrt 22 0000'' => Imaginary roots (If B<0 then I
would have two equal real roots).
So jbr => btGbtcbtcx cos)sin()cos( 21
A=0, B=1, x(0)=1, x’(0)=0 =>c1=1, c2=0:
cos
Trigonometric
Function1
sin
Trigonometric
Function
Scope5
Scope4
Scope3
Scope2
Scope1
Scope
Product
eu
Math
Function
1
s
Integrator1
1
s
Integrator
0
Gain5
1
Gain4
1
Gain3
0
Gain2
0
Gain1
Clockt
bt sin(bt)
cos(bt)
exp(-at)
0 2 4 6 8 10-1
-0.8
-0.6
-0.4
-0.2
0
0.2
0.4
0.6
0.8
1
Overall
![Page 19: EEE2012 Control Notes](https://reader034.fdocuments.us/reader034/viewer/2022050820/5695d2061a28ab9b0298d7b3/html5/thumbnails/19.jpg)
EEE2012: Control.
1st and 2nd order systems page 19
In all previous cases if the real part is positive then the solution will diverge to
infinity and the ODE (and hence the system) is called unstable.
Space of roots
jb
a
Critical or
overdamped
underdamped
jb
a
Stable
Unstable
![Page 20: EEE2012 Control Notes](https://reader034.fdocuments.us/reader034/viewer/2022050820/5695d2061a28ab9b0298d7b3/html5/thumbnails/20.jpg)
EEE2012: Control.
1st and 2nd order systems page 20
g) Natural frequency, damping frequency, damping factor
2nd order systems very important with rich dynamic behaviour
So 2,2 nn BA => 0'2'' 2 xxx nn
is the damping factor and n is the natural frequency of the system.
2
422
2
4222nnnBAA
r
222nnnr
1. Real and unequal 110
2222
nn => Overdamped system
implies that 1 ; 2222,1 nnnr => replace at
trtreCeCx 21
21 .
2. Real and equal 112222 nn => Critically damped system
implies that 1 ; nr => tt nn teCeCx
21 .
3. Complex 110
2222
nn => Underdamped systems implies
1 ; 22222,1 1 nnnnn jjr
dn jnd
21
=>
tGex d
tn cos ; d is called damped
frequency or pseudo-frequency.
4. Imaginary roots 0 and therefore the solution is njr =>
tGx ncos ; so when there is no damping the frequency of the
oscillations = natural frequency. tGx ncos
5. In all the previous cases if 0 then the transient part tends to zero. If
0 then the system will diverge to infinity with or without oscillations.
![Page 21: EEE2012 Control Notes](https://reader034.fdocuments.us/reader034/viewer/2022050820/5695d2061a28ab9b0298d7b3/html5/thumbnails/21.jpg)
EEE2012: Control.
1st and 2nd order systems page 21
Oscillations? Name Components of solution
1 No Overdamped Two exponentials:
tktkee 21 , , 0, 21 kk
1 No Critically
damped
Two exponentials:
ktkt tee , , 0k
1 Yes Underdamped One exponential and one
cosine tekt cos, , 0k
0 Yes Undamped one cosine tcos
If 0 then cases 1-3 are the same but with 0k
dj
n
Critical or
overdamped
underdamped
Stable
Unstable
dj
n
![Page 22: EEE2012 Control Notes](https://reader034.fdocuments.us/reader034/viewer/2022050820/5695d2061a28ab9b0298d7b3/html5/thumbnails/22.jpg)
EEE2012: Control.
1st and 2nd order systems page 22
h) NonHomogeneous (NH) differential equations
uBxAxx '''
u=0 => Homogeneous => x1 & x2.
Assume a particular solution of the nonhomogeneous ODE: xp
o If u(t)=R=cosnt => B
RxP
Then all the solutions of the NHODE are 2211 xcxcxx P
So we have all the previous cases for under/over/un/critically damped
systems plus a constant R/B.
If complementary solution is stable then the particular solution is called
steady state.
Example:
22''' Pxxxx
)sin()cos(22 212211 btcbtcexcxcx at
x(0)=1, x’(0)=0 => c1=-1, c2=-1/sqrt(3)
![Page 23: EEE2012 Control Notes](https://reader034.fdocuments.us/reader034/viewer/2022050820/5695d2061a28ab9b0298d7b3/html5/thumbnails/23.jpg)
EEE2012: Control.
1st and 2nd order systems page 23
cos
Trigonometric
Function1
sin
Trigonometric
Function
Scope5
Scope4
Scope3
Scope2
Scope1
Scope
Product
eu
Math
Function
1
s
Integrator1
1
s
Integrator
-1/sqrt(3)
Gain5
sqrt(3)/2
Gain4
-1
Gain3
-0.5
Gain2
1
Gain12
Constant
Clockt
bt sin(bt)
cos(bt)
exp(-at)
0 2 4 6 8 10-1
-0.5
0
0.5
1
1.5
2
2.5
Overall
Homogeneous solution
Particular solution
![Page 24: EEE2012 Control Notes](https://reader034.fdocuments.us/reader034/viewer/2022050820/5695d2061a28ab9b0298d7b3/html5/thumbnails/24.jpg)
EEE2012: Control.
1st and 2nd order systems page 24
Tutorial 5
State whether each of the following systems are stable/unstable, oscillatory/non-oscillatory.
1.
+3
+2y=1
2.
+2
+2y=1
3.
-5
+6y=1
4.
-2
+104y=1
5.
y=1
Using Simulink numerically solve each and sketch the output. Solutions
1. Stable, non-oscillatory
2. Stable, oscillatory
3. Unstable
4. Unstable, oscillatory
5. Marginally stable
0 1 2 3 4 5 6 70
0.05
0.1
0.15
0.2
0.25
0.3
0.35
0.4
0.45
0.5
Step Response
Time (sec)
Am
plit
ude
0 1 2 3 4 5 6 70
0.1
0.2
0.3
0.4
0.5
0.6
0.7
Step Response
Time (sec)
Am
plit
ude
0 2 4 6 8 10 12 14 16 18 200
1
2
3
x 1025 Step Response
Time (sec)
Am
plit
ude
0 10 20 30 40 50 60-10
-8
-6
-4
-2
0
2
4
6
8x 10
23 Step Response
Time (sec)
Am
plit
ude
10 20 30 40 50 60 70-0.5
0
0.5
1
1.5
2
Step Response
Time (sec)
Am
plit
ude
![Page 25: EEE2012 Control Notes](https://reader034.fdocuments.us/reader034/viewer/2022050820/5695d2061a28ab9b0298d7b3/html5/thumbnails/25.jpg)
EEE2012: Control
Transfer functions and the s-plane Page 25
Chapter 2. Transfer functions
A Transfer function is the ratio of the Laplace transform of the output over the Laplace transform of the input.
The Laplace Transform
Used only on Linear Time Invariant systems Convert the Differential expression into a Polynomial expression;
{ ( )
( ) } ( ) ( )
The LT is transforming a DE from the time domain (domain is a set of values that describe a function, in that case the variable is the time) to another complex domain (i.e. the variable has a real and imaginary part).
dttfetfLsF st )()()( , js
Can perform this from first principles, but the formula tables are easier.
Differentiation
{ ( )
} ( ) ( )
{ ( )
} ( ) ( )
( )
Where f(0) and df(0)/dt are the initial conditions.
![Page 26: EEE2012 Control Notes](https://reader034.fdocuments.us/reader034/viewer/2022050820/5695d2061a28ab9b0298d7b3/html5/thumbnails/26.jpg)
EEE2012: Control
Transfer functions and the s-plane Page 26
Example 1
A mechanical system with force as the input and position as the output:
xout
(t)xin(t)
Define the system equations;
2
2
dt
xdmfmaf out
Spring
2
2
dt
xdmxxK out
outin
Perform the Laplace Transform;
KmssXsKX
sXmssKXsKX
dt
xdmxxK
outin
outoutin
LT
IC
out
outin
2
2
02
2
And express the transfer function;
Kms
K
sX
sX
in
out
2
This result is compared to the Characteristic Equation (CE) of differential equation;
outout
inout
outin Kxdt
xdmKx
dt
xdmKxKx
2
2
2
2
The homogeneous system is: outout Kx
dt
xdm
2
2
0 and therefore the CE is
02 Kmr i.e. exactly the same as the denominator of TF.
![Page 27: EEE2012 Control Notes](https://reader034.fdocuments.us/reader034/viewer/2022050820/5695d2061a28ab9b0298d7b3/html5/thumbnails/27.jpg)
EEE2012: Control
Transfer functions and the s-plane Page 27
Example 2 An electromechanical system with a voltage input, an intermediate force and position as the output:
Define the system equations. The force of attraction of the mass to the coil is non-linear, therefore cannot be transformed to the Laplace domain, therefore linearise this over a small region of operation;
2
2
x
ikf A => ikf A
Define the remaining linear system equations – both electrical and mechanical and perform the Laplace transform on each set;
)()()()(
)()(1
)(1
2 sXmssBsXskXsIk
RsIsVL
ssI
xmxBkxik
iRvLdt
di
AA
Separate both sets with Output=input – (Current and voltage, forces and acceleration);
)()()()(
)()()(
2 sXmssBsXskXsIk
sVRsIssLI
A
As current is in both the electrical and mechanical equations isolate and substitute to simplify;
)()()()(
)()(
2 sXmssBsXskXRsL
sVk
RsL
sVsI
A
v(t)
R
L
f
Km
x
B
![Page 28: EEE2012 Control Notes](https://reader034.fdocuments.us/reader034/viewer/2022050820/5695d2061a28ab9b0298d7b3/html5/thumbnails/28.jpg)
EEE2012: Control
Transfer functions and the s-plane Page 28
Then simplify the combined equation;
)()( 2 sXkBsmsRsL
sVkA
And finally express as a transfer function;
kBsmsRsL
k
sV
sX A
2)(
)(
Example 3 A motor connected via an ideal gear box to an inertial load with friction.
Define the physical equations;
0
1
2
0
1
20
1
20
000
n
nK
dt
diLRivK
dt
diLRiv
BiKn
nJ
iKT
Tn
nT
BTJ
Ta
aaaamTa
aaaa
aT
aTm
m
Combine and simplify;
0
1
2
0
2
10
n
nK
dt
diLRiv
BiKn
nJ
T
a
aaaa
aT
v(t)
R L
EMF
Tm,θ
m
r1
r2
N1
N2
θo
J2
B
Tm,θ
m
To,θ
o
![Page 29: EEE2012 Control Notes](https://reader034.fdocuments.us/reader034/viewer/2022050820/5695d2061a28ab9b0298d7b3/html5/thumbnails/29.jpg)
EEE2012: Control
Transfer functions and the s-plane Page 29
Now perfume the Laplace transform;
sKssILRsIsV
sBsIKsJs
aaaaa
aLT
n
nKK
n
nKK TT 01
020
2
,2
12
1
21
Isolate the component in both equations;
sKsIsLRsV
sK
BJssI
aaaa
a
01
0
2
2
And simplify;
1
2
2
0
010
2
2
1
KK
BJssLR
sV
s
sKsK
BJssLRsV
aa
a
aaa
Now express the transfer function in its most simple form;
21
2
20
KKBJssLR
K
sV
s
aaa
![Page 30: EEE2012 Control Notes](https://reader034.fdocuments.us/reader034/viewer/2022050820/5695d2061a28ab9b0298d7b3/html5/thumbnails/30.jpg)
EEE2012: Control
Transfer functions and the s-plane Page 30
a) Characteristic Equation and Order
The denominator of Transfer Function is the characteristic Equation of ODE, this is the characteristic Equation (CE) of the system. The order of the ODE is the order of the system. 1. Perform the Laplace transform on the following system equation and rearrange to give the transfer function. What is the system order?
( )
( )
( )
2. What is the order of this system?
( )
( )
( )(
)
4th order
![Page 31: EEE2012 Control Notes](https://reader034.fdocuments.us/reader034/viewer/2022050820/5695d2061a28ab9b0298d7b3/html5/thumbnails/31.jpg)
EEE2012: Control
Transfer functions and the s-plane Page 31
b) Pole location in the s-plane
sD
sNsG )( : the roots of the numerator are called zeros whilst the roots of
the denominator are called poles.
Points on the S plane where G(s) is analytic are called ordinary points.
Points where G(s) is not analytic are called singular points.
Singular points approaching infinity are called poles.
Singular points approaching zero are called…………………..zeros!
Example 1 ( )
( )
plot the real and imaginary part of the poles and zeros on the
s-plane; ( )
( )
( )
( )( )
( )( ) {
0
j
-1-2 3
![Page 32: EEE2012 Control Notes](https://reader034.fdocuments.us/reader034/viewer/2022050820/5695d2061a28ab9b0298d7b3/html5/thumbnails/32.jpg)
EEE2012: Control
Transfer functions and the s-plane Page 32
Example 2
1
1)(
2
sssG
Poles at -0.5+/- 0.866j ;
To find the roots of a polynomial use the command roots( ):
>> num=[1 4];
>> den=[1 3 -6];
>> roots(num)
ans =
-4
>> roots(den)
ans =
-4.3723
1.3723
0
j
-0.5
+0.866j
-0.866j
![Page 33: EEE2012 Control Notes](https://reader034.fdocuments.us/reader034/viewer/2022050820/5695d2061a28ab9b0298d7b3/html5/thumbnails/33.jpg)
EEE2012: Control
Transfer functions and the s-plane Page 33
To find and plot directly the poles/zeros use the command pzmap( ):
>> num=[1 4];
>> den=[1 3 -6];
>> g=tf(num,den);
>> pzmap(g)
>> [p,z]=pzmap(g)
p =
-4.3723
1.3723
z =
-4
Useful commands: zgrid, sgrid.
-5 -4 -3 -2 -1 0 1 2-1
-0.8
-0.6
-0.4
-0.2
0
0.2
0.4
0.6
0.8
1 Pole-Zero Map
Real Axis
Imagin
ary
Axis
![Page 34: EEE2012 Control Notes](https://reader034.fdocuments.us/reader034/viewer/2022050820/5695d2061a28ab9b0298d7b3/html5/thumbnails/34.jpg)
EEE2012: Control
Transfer functions and the s-plane Page 34
c) Final value theorem
The final value theorem is used to calculate the steady state value – If
one exists.
( ) where fss is the value of f(t) after infinite time.
Performing the Laplace transform gives;
( ) ( ) ( ) ( )
As a transfer function this is;
( )
( )
( )
Rearranging to express the output, X(s) as a function of the input;
( ) ( )
( )
And applying the final value theorem ( ( )
( )) with
step input of magnitude r; ( )
gives
(
( ))
ONLY VALID WHEN A STEADY STATE EXISTS.
![Page 35: EEE2012 Control Notes](https://reader034.fdocuments.us/reader034/viewer/2022050820/5695d2061a28ab9b0298d7b3/html5/thumbnails/35.jpg)
EEE2012: Control
Transfer functions and the s-plane Page 35
d) Transfer functions in Matlab
Define the numerator
Define the denominator
Use the command tf( numerator, denominator )
Example 1
2
5
ssG
>> num=5;
>> den=[1 2];
>> g=tf(num,den)
Transfer function:
5
-----
s + 2
Example 2
63
42
ss
ssG
>> num=[1 2];
>> num=[1 4];
>> den=[1 3 -6];
>> g=tf(num,den)
Transfer function:
s + 4
-------------
s^2 + 3 s - 6
![Page 36: EEE2012 Control Notes](https://reader034.fdocuments.us/reader034/viewer/2022050820/5695d2061a28ab9b0298d7b3/html5/thumbnails/36.jpg)
EEE2012: Control
Transfer functions and the s-plane Page 36
e) Block diagrams
System
X(t) Y(t)
Input OutputIn the time domain
G(s)
X(s) Y(s)
And in the s domain
Block Diagram Algebra
1. To sum (or subtract) two signals, we use a summing point:
2. To “distribute” a signal, we use a branch point:
X1(s)
X2(s)
X1(s)+X
2(s) X
1(s)
X2(s)
X1(s)-X
2(s)
X1(s) X
1(s)
X1(s)
![Page 37: EEE2012 Control Notes](https://reader034.fdocuments.us/reader034/viewer/2022050820/5695d2061a28ab9b0298d7b3/html5/thumbnails/37.jpg)
EEE2012: Control
Transfer functions and the s-plane Page 37
3. Series connection:
4. Parallel connection
G1(s)
Y(s)
X(s)
G2(s)
Z(s)
X(s)
X(s)
C(s)
G1(s)+G
2(s)
X(s) C(s)
G1(s)
Y(s)X(s)G
2(s)
Z(s)
G1(s) G
2(s)
Z(s)X(s)
![Page 38: EEE2012 Control Notes](https://reader034.fdocuments.us/reader034/viewer/2022050820/5695d2061a28ab9b0298d7b3/html5/thumbnails/38.jpg)
EEE2012: Control
Transfer functions and the s-plane Page 38
Block diagrams in Matlab:
>> num1=1;
>> den1=[1 1];
>> num2=[1 0];
>> den2=[1 5 6];
>> g1=tf(num1,den1)
Transfer function:
1
-----
s + 1
>> g2=tf(num2,den2)
Transfer function:
s
-------------
s^2 + 5 s + 6
>> g=series(g1,g2)
Transfer function:
s
----------------------
s^3 + 6 s^2 + 11 s + 6
>> g=parallel(g1,g2)
Transfer function:
2 s^2 + 6 s + 6
----------------------
s^3 + 6 s^2 + 11 s + 6
![Page 39: EEE2012 Control Notes](https://reader034.fdocuments.us/reader034/viewer/2022050820/5695d2061a28ab9b0298d7b3/html5/thumbnails/39.jpg)
EEE2012: Control
Transfer functions and the s-plane Page 39
Tutorial
1. Find the order and plot the poles and zeros of;
( )
, ( )
( ) ( )
( )( ) ( )
2. If Y(s)=R(s)G(s), for the stable systems in Q2 find the final value Y
ss in
response to a unit step input ( )
3. Using Matlab find the transfer function of the following;
a) G1
and G2
are in parallel connection
b) G1
and G2
are in series connection
c) G1
and G2
are in series and this is in parallel with G2
connection
d) G1
and G2
are in parallel and this is in series with G1
connection
Where ( )
( )
4. Simplify the system below:
G2(s)X(s)
G1(s) G3(s)
G4(s)
G5(s)
Y(s)
![Page 40: EEE2012 Control Notes](https://reader034.fdocuments.us/reader034/viewer/2022050820/5695d2061a28ab9b0298d7b3/html5/thumbnails/40.jpg)
EEE2012: Control
Transfer functions and the s-plane Page 40
Q1. G1, 1st order, No zeros, Pole@s=-4 G2, 2nd order, No zeros, Pole@4±j8 G3, 3rd order, zero@s=-4, pole@s= 2±j6 G4, 2nd order, no zeros, pole @ s=±j6 Q2.
Q3.
>> ga=parallel(g1,g2) Transfer function: 2 s^2 + 9 s + 11 --------------------- s^3 + 4 s^2 + 9 s + 6
>> gb=series(g1,g2) Transfer function: s + 5 --------------------- s^3 + 4 s^2 + 9 s + 6
>> gc=parallel(gb,g2) Transfer function: s^4 + 10 s^3 + 37 s^2 + 72 s + 60 ----------------------------------------- s^5 + 7 s^4 + 27 s^3 + 57 s^2 + 72 s + 36
>> gd=series(ga,g1) Transfer function: 2 s^2 + 9 s + 11 ------------------------------- s^4 + 5 s^3 + 13 s^2 + 15 s + 6
Q4. Done in class
( ( )
) ( )
(
)
![Page 41: EEE2012 Control Notes](https://reader034.fdocuments.us/reader034/viewer/2022050820/5695d2061a28ab9b0298d7b3/html5/thumbnails/41.jpg)
EEE 2002: Control
Chapter 3 – Time domain characteristics 41/132
Chapter 3. Time domain characteristics
You have derived the mathematical model of the system. Now apply various methods to analyse the performance of the system. All systems have a forcing signal which we want to control, four main types of input are given below, all are simple functions of time (and s)
;
The Dirac function The step or the pulse function
The Ramp function The parabolic function
“Once a control system is designed on the basis of test signals, the performance of the system in response to actual inputs is generally satisfactory” [Ogata p.220] Gradually changing function of time → Ramp Sudden disturbance → Step Shock → Impulse Acceleration → Parabolic Allows for comparison of all systems on the same basis
0
r(t)
t
ttr
1sR
0
r(t)
t
Atr
s
AsR
0
r(t)
t
Attr
2s
AsR
0
r(t)
t
2Attr
3s
AsR
0
r(t)
t
ttr
1sR
0
r(t)
t
Atr
s
AsR
0
r(t)
t
Attr
2s
AsR
0
r(t)
t
2Attr
3s
AsR
![Page 42: EEE2012 Control Notes](https://reader034.fdocuments.us/reader034/viewer/2022050820/5695d2061a28ab9b0298d7b3/html5/thumbnails/42.jpg)
EEE 2002: Control
Chapter 3 – Time domain characteristics 42/132
a) First order systems
Remember: baxx ' ;
Homogeneous equation: 0' axx
Characteristic equation; 0 am A solution of the HE: ateCx 1 hence this is a stable solution if a<0 (or the
pole is at the left hand side)
The same in the s-plane: LsRsV
sI
1
s
K
sV
sI
1, K=1/R and τ=L/R. =>
1
s
KsVsI
Step response: 1
1
ss
VKsI
=>
t
eR
Vti 1
Steady state value;
R
V
R
Ve
R
Vi
t
tF
01lim
And the same in the s domain;
R
VVK
s
K
s
VsI
sss
1lim
0
V(s) I(s)
LsR
10
j
LR
![Page 43: EEE2012 Control Notes](https://reader034.fdocuments.us/reader034/viewer/2022050820/5695d2061a28ab9b0298d7b3/html5/thumbnails/43.jpg)
EEE 2002: Control
Chapter 3 – Time domain characteristics 43/132
To find the step response directly in Matlab (apart from using Simulink), use
the command:
step( , ):
>> num=1;
>> den=[1 1];
>> step(num,den)
Step Response
Time (sec)
Am
plit
ude
0 T 2T 3T 4T 5T 6T0
0.1(V/R)
0.2(V/R)
0.3(V/R)
0.4(V/R)
0.5(V/R)
0.6(V/R)
0.7(V/R)
0.8(V/R)
0.9(V/R)
1 (V/R)
Step Response
Time (sec)
Am
plit
ude
0 T 2T 3T 4T 5T 6T0
0.1(V/R)
0.2(V/R)
0.3(V/R)
0.4(V/R)
0.5(V/R)
0.6(V/R)
0.7(V/R)
0.8(V/R)
0.9(V/R)
1 (V/R)
0.632 (V/R)
![Page 44: EEE2012 Control Notes](https://reader034.fdocuments.us/reader034/viewer/2022050820/5695d2061a28ab9b0298d7b3/html5/thumbnails/44.jpg)
EEE 2002: Control
Chapter 3 – Time domain characteristics 44/132
Or to extend the plot to 15 seconds; >> step(num,den,15)
0 1 2 3 4 5 60
0.2
0.4
0.6
0.8
1Step Response
Time (sec)
Am
plitu
de
0 5 10 150
0.2
0.4
0.6
0.8
1Step Response
Time (sec)
Am
plitu
de
![Page 45: EEE2012 Control Notes](https://reader034.fdocuments.us/reader034/viewer/2022050820/5695d2061a28ab9b0298d7b3/html5/thumbnails/45.jpg)
EEE 2002: Control
Chapter 3 – Time domain characteristics 45/132
Ramp response;
1
12
ss
VKsC
=>
t
etR
Vtc
The final value of the error is;
R
Ve
R
Vive
t
xF
1lim
Step Response
Time (sec)
Am
plit
ude
0 T 2T 3T 4T 5T 6T 0 T 2T 3T0
0.1(V/R)
0.2(V/R)
0.3(V/R)
0.4(V/R)
0.5(V/R)
0.6(V/R)
0.7(V/R)
0.8(V/R)
0.9(V/R)
1 (V/R) VT----- R
![Page 46: EEE2012 Control Notes](https://reader034.fdocuments.us/reader034/viewer/2022050820/5695d2061a28ab9b0298d7b3/html5/thumbnails/46.jpg)
EEE 2002: Control
Chapter 3 – Time domain characteristics 46/132
Tutorial
A first order system is given by ( )
( )
1. Write the transfer function as ( )
( )
and hence find the value
of the DC gain ( ) and of the system’s time constant ( ). 2. For k=1 and a=5 repeat the previous step. 3. Plot a pole zero map for this transfer function, and based on this
predict the dynamic behaviour of the system in response to a (unit) step input.
4. Using the final value theorem calculate the steady state value Yss in response to a (unit) step input.
5. Confirm your findings using Matlab or Simulink 6. Repeat the previous steps and comment on your results for:
1. k=0.1 and k=10 (keep a fixed to 5). 2. a=0.5 and a=50 (keep k fixed to 1).
7. Using the appropriate Matlab commands find and plot the error for the system in question 2. What is the steady state error?
![Page 47: EEE2012 Control Notes](https://reader034.fdocuments.us/reader034/viewer/2022050820/5695d2061a28ab9b0298d7b3/html5/thumbnails/47.jpg)
EEE 2002: Control
Chapter 3 – Time domain characteristics 47/132
b) Second order systems
Physical equations;
xout(t)xin(t)
Friction
LT
IC
outoutoutin
dt
xdm
dt
dxBxxK
02
2
In the s domain; sXmssBsXsKXsKX outoutoutin
2
sXmsBsKsKX outin
2
And simplified to give a transfer function;
KBsms
K
sX
sX
in
out
2
mKs
mBs
mK
sX
sX
in
out
2
Expressed in terms of damping factor and natural frequency n; 22
2
2 nn
n
sssR
sC
,
mK
mB
n 2,2
The roots of the characteristic equation are given by;
12 nns
![Page 48: EEE2012 Control Notes](https://reader034.fdocuments.us/reader034/viewer/2022050820/5695d2061a28ab9b0298d7b3/html5/thumbnails/48.jpg)
EEE 2002: Control
Chapter 3 – Time domain characteristics 48/132
Case 1 >1:
Then the system has two negative real roots and is called overdamped:
212
21
121
s
e
s
etc
tsts
n
1 => tn
etc
12
1
An overdamped system response is similar in form to a slow first order system (long decaying exponential).
0
j
s1
s2
0 2 4 6 8 10 120
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1Step Response
Time (sec)
Am
plit
ude
Overdamped system
![Page 49: EEE2012 Control Notes](https://reader034.fdocuments.us/reader034/viewer/2022050820/5695d2061a28ab9b0298d7b3/html5/thumbnails/49.jpg)
EEE 2002: Control
Chapter 3 – Time domain characteristics 49/132
Case 2 =1:
The system has two equal real roots at ns :
0
j
n
tetc n
tn 11 :
Step Response
Time (sec)
Am
plit
ude
0 2 4 6 8 10 120
0.5
1
1.5
Overdamped system
Critically damped system
nnn ss 12
![Page 50: EEE2012 Control Notes](https://reader034.fdocuments.us/reader034/viewer/2022050820/5695d2061a28ab9b0298d7b3/html5/thumbnails/50.jpg)
EEE 2002: Control
Chapter 3 – Time domain characteristics 50/132
Case 3 0<<1:
21 nn js or dn js
The line between the origin and the pole is:
222
ndd nnn 2222 1
And the angle is;
1coscos
n
n
In the time domain the response expressed in terms of damping factor and
natural frequency;
21
2
1tansin
11 t
etc d
tn
0
j
n
dj
dj
![Page 51: EEE2012 Control Notes](https://reader034.fdocuments.us/reader034/viewer/2022050820/5695d2061a28ab9b0298d7b3/html5/thumbnails/51.jpg)
EEE 2002: Control
Chapter 3 – Time domain characteristics 51/132
The envelope that will be created from the exponential terms is:
0 1 2 3 4 5 6 7 8 9 10-1
-0.8
-0.6
-0.4
-0.2
0
0.2
0.4
0.6
0.8
1
Sinusoidal term
Exponential term
0 1 2 3 4 5 6 7 8 9 10-1
-0.8
-0.6
-0.4
-0.2
0
0.2
0.4
0.6
0.8
1
![Page 52: EEE2012 Control Notes](https://reader034.fdocuments.us/reader034/viewer/2022050820/5695d2061a28ab9b0298d7b3/html5/thumbnails/52.jpg)
EEE 2002: Control
Chapter 3 – Time domain characteristics 52/132
And their product:
And by adding the constant factor:
0 1 2 3 4 5 6 7 8 9 10-1
-0.8
-0.6
-0.4
-0.2
0
0.2
0.4
0.6
0.8
1
0 1 2 3 4 5 6 7 8 9 10-1
-0.5
0
0.5
1
1.5
2
![Page 53: EEE2012 Control Notes](https://reader034.fdocuments.us/reader034/viewer/2022050820/5695d2061a28ab9b0298d7b3/html5/thumbnails/53.jpg)
EEE 2002: Control
Chapter 3 – Time domain characteristics 53/132
The three stable second order system responses are compared below;
And with reducing further toward zero the damping reduces and the system oscillates for longer;
Step Response
Time (sec)
Am
plit
ude
0 2 4 6 8 10 120
0.5
1
1.5
Overdamped system
Critically damped systemUnderdamped system
0 2 4 6 8 100
0.2
0.4
0.6
0.8
1
1.2
1.4
1.6
1.8Step Response
Time (sec)
Am
plit
ude
![Page 54: EEE2012 Control Notes](https://reader034.fdocuments.us/reader034/viewer/2022050820/5695d2061a28ab9b0298d7b3/html5/thumbnails/54.jpg)
EEE 2002: Control
Chapter 3 – Time domain characteristics 54/132
Case 4 =0:
Note: The system is called marginally stable because the solutions do not diverge to infinity. Hence if the previous four cases are combined to one graph:
0
j
nj
nj
0 5 10 15 20 250
0.2
0.4
0.6
0.8
1
1.2
1.4
1.6
1.8
2Step Response
Time (sec)
Am
plit
ude
Marginally stable
Underdamped
Critically damped
Overdamped
![Page 55: EEE2012 Control Notes](https://reader034.fdocuments.us/reader034/viewer/2022050820/5695d2061a28ab9b0298d7b3/html5/thumbnails/55.jpg)
EEE 2002: Control
Chapter 3 – Time domain characteristics 55/132
Case 5 <0 (negative damping i.e. positive feedback):
21
2
1tansin
11 t
etc d
tn
Since 0 then angle defined in the s-plane
(
1coscos
n
n ) has to be greater than 90o:
By combining the previous s-planes we have:
0 1 2 3 4 5-10
-5
0
5
10
0
j
n
dj
dj
![Page 56: EEE2012 Control Notes](https://reader034.fdocuments.us/reader034/viewer/2022050820/5695d2061a28ab9b0298d7b3/html5/thumbnails/56.jpg)
EEE 2002: Control
Chapter 3 – Time domain characteristics 56/132
0
j
n
dj
dj
nj
nj
1,0cos
1
0
1
0,1cos
1
1
![Page 57: EEE2012 Control Notes](https://reader034.fdocuments.us/reader034/viewer/2022050820/5695d2061a28ab9b0298d7b3/html5/thumbnails/57.jpg)
EEE 2002: Control
Chapter 3 – Time domain characteristics 57/132
0
j
0 2 4 6 8 100
0.2
0.4
0.6
0.8
1
0 2 4 6 8 100
0.2
0.4
0.6
0.8
1
1.2
1.4
0 2 4 6 8 100
0.5
1
1.5
2
0 0.25 0.5 0.75 1 -4
-3
-2
-1
0
1
2x 10
8
0 0.2 0.4 0.6 0.8 10
2
4
6
8
10
0 0.2 0.4 0.6 0.8 10
20
40
60
80
100
120
140
0 2 4 6 8 100
0.2
0.4
0.6
0.8
1
![Page 58: EEE2012 Control Notes](https://reader034.fdocuments.us/reader034/viewer/2022050820/5695d2061a28ab9b0298d7b3/html5/thumbnails/58.jpg)
EEE 2002: Control
Chapter 3 – Time domain characteristics 58/132
A general response in the time domain is:
This response can be described by a number of parameters;
Time that the system needs to reach half of its final value:
Rise time (10%-90% or 5%-95% or 0%-100%) d
rt
Peak time: d
pt
Maximum overshoot:
"1
eMp
Settling time: n
st
3%5 and
n
st
4%2
0 1 2 3 4 5 60
0.2
0.4
0.6
0.8
1
1.2
1.4Step Response
Time (sec)
Am
plit
ud
e
![Page 59: EEE2012 Control Notes](https://reader034.fdocuments.us/reader034/viewer/2022050820/5695d2061a28ab9b0298d7b3/html5/thumbnails/59.jpg)
EEE 2002: Control
Chapter 3 – Time domain characteristics 59/132
0 1 2 3 4 5 60
0.2
0.4
0.6
0.8
1
1.2
1.4Step Response
Time (sec)
Am
plit
ude
td
tr
tp
Mp
1.05
0.95
ts
![Page 60: EEE2012 Control Notes](https://reader034.fdocuments.us/reader034/viewer/2022050820/5695d2061a28ab9b0298d7b3/html5/thumbnails/60.jpg)
EEE 2002: Control
Chapter 3 – Time domain characteristics 60/132
c) Time domain characteristics in Matlab
>> num=1;
>> den=[1 4 13];
>> g=tf(num,den);
>> step(g)
Right click on figure and choose the characteristics that you would like to see.
0 0.5 1 1.5 2 2.5 30
0.01
0.02
0.03
0.04
0.05
0.06
0.07
0.08
0.09Step Response
Time (sec)
Am
plitu
de
![Page 61: EEE2012 Control Notes](https://reader034.fdocuments.us/reader034/viewer/2022050820/5695d2061a28ab9b0298d7b3/html5/thumbnails/61.jpg)
EEE 2002: Control
Chapter 3 – Time domain characteristics 61/132
![Page 62: EEE2012 Control Notes](https://reader034.fdocuments.us/reader034/viewer/2022050820/5695d2061a28ab9b0298d7b3/html5/thumbnails/62.jpg)
EEE 2002: Control
Chapter 3 – Time domain characteristics 62/132
d) Extra poles and zeros
Systems with an order higher than 2; General form of a Transfer function;
nn
nn
mm
m
asasasa
bsbsb
sR
sC
1
1
10
10
...
...
)(
)(
For a step input:
r
k kkk
kkkkkkq
j j
j
ss
csb
ps
a
ssC
122
1 2
11
nrq 2
i.e. combination of first and second order systems.
Example – 3rd order system
Express the TF as a product of 1st and 2nd order systems;
edssfscbsass
223
11=>
Equate the like terms;
edssfscbsass 223
fesfdesfdscbsass 2323
Form a set of simultaneous equations and solve;
fec
fdeb
fda
11
The response of a higher order system is the sum of exponential and damped sinusoidal curves.
Assuming that all poles are at the left hand side then the final value of the output is “1” since all exponential terms will converge to 0.
Let’s assume that some poles have real parts that are far away from the imaginary axis=>
![Page 63: EEE2012 Control Notes](https://reader034.fdocuments.us/reader034/viewer/2022050820/5695d2061a28ab9b0298d7b3/html5/thumbnails/63.jpg)
EEE 2002: Control
Chapter 3 – Time domain characteristics 63/132
21
2
1tansin
11 t
etc d
tn
=> 0 tne
Overall performance is characterised by the isolated (far away from zeros) poles that are close to the imaginary axis.
If we have only one pole (or a pair for complex roots) that is closed to the real axis then we say that this pole (or pair of poles) is (are) the DOMINANT pole(s) for the system.
A simple rule is that the dominant poles must be at least five to ten times closer to the imaginary axis than the other ones.
r
k kkk
kkkkkkq
j j
j
ss
csb
ps
a
ssC
122
1 2
11
r
k
kk
t
k
r
k
kk
t
k
q
j
tp
j tectebeatc kkkkj
1
2
1
2
1
1sin1cos1
The values of b (numerator coefficients) determine the amplitude of the oscillations of the system.
![Page 64: EEE2012 Control Notes](https://reader034.fdocuments.us/reader034/viewer/2022050820/5695d2061a28ab9b0298d7b3/html5/thumbnails/64.jpg)
EEE 2002: Control
Chapter 3 – Time domain characteristics 64/132
Tutorial
1. A second order system is given by cbsas
ksG
2; write the
transfer function as: 22 2
'
nnss
ksG
.
2. For k’=1 and sradn /5,5.0 define the transfer function.
3. Enter this into Matlab and use the command “damp” to confirm the damping factor and natural frequency.
4. Based on the previous answer predict the behavior of the system for a (unit) step input.
5. Find the step response of that system using Matlab AND Simulink and hence crosscheck your previous answer. Also plot the system’s error.
6. Use the formulae to calculate overshoot, MP, time to peak, TP, settling time to within 5%, ts5%, and confirm these findings using your Matlab plot.
7. Using the final value theorem, calculate the steady state value in response to a unit step input. Confirm this finding with your Matlab plot.
8. The mechanical system in Figure 1 is described by a differential equation expressing the sum of forces;
a) Where F is the applied force, k is the spring constant, is the
coefficient of friction and m is the mass. are position, velocity and acceleration respectively.
b) Assuming zero initial conditions calculate the Laplace transform of this
system, expressing your result as a transfer function relating output X and input F.
m
←
k F→
FIGURE 1.
![Page 65: EEE2012 Control Notes](https://reader034.fdocuments.us/reader034/viewer/2022050820/5695d2061a28ab9b0298d7b3/html5/thumbnails/65.jpg)
EEE 2002: Control
Chapter 3 – Time domain characteristics 65/132
c) Express the damping factor ζ and the natural frequency ωn as a function of m, and k.
d) Assuming the mass m=0.1kg and the friction coefficient =0.1 what spring constant k will result in an under-damped system?
e) For a spring constant k ten times greater than your calculated value in part iii, calculate the position of the system poles and the damping factor. Can you confirm that you have invoked a stable, under-damped system?
9. A plant has the transfer function ( )
( )( )( ) and forms
part of an open loop system where the output Y(s)=X(s)G(s). X(s) is the system input. a) Calculate the position of any poles and zeros for this system. Plot
these on an s-plane diagram. b) Considering the location of the poles and zeros, what do you think will
be the system response to a unit step input? c) Calculate the final value Yss(s) in response to a step input. d) Use matlab to plot the system step response and confirm your answer
to 4. e) Approximate this system by considering the position of any dominant
poles. f) Take account of the final value of your approximated system and
modifying the system gain (numerator value). g) Hold your previous step response and superimpose your simplified
system step response. Has your simplification worked?
![Page 66: EEE2012 Control Notes](https://reader034.fdocuments.us/reader034/viewer/2022050820/5695d2061a28ab9b0298d7b3/html5/thumbnails/66.jpg)
EEE 2002: Control
Chapter 4 –Closed loop systems 66/132
Chapter 4. Closed loop systems
Assume the following electrical system:
R L
U
Change or control V so the output current i is equal to the demanded current i*. Assume R=2 Ohms, L=1 H and i*=2 A:
Ri* U
RLs
1 i
Expected response:
i
t
2A
Assume that you want a faster system (same steady state):
V
V
![Page 67: EEE2012 Control Notes](https://reader034.fdocuments.us/reader034/viewer/2022050820/5695d2061a28ab9b0298d7b3/html5/thumbnails/67.jpg)
EEE 2002: Control
Chapter 4 –Closed loop systems 67/132
t
2A
L & R = constant hence we can ONLY change U but then iss i* Ideally we would like:
t
2A
U
=i* x R
I
So V=f(i*,i)
As ii* Vi*R
And as i-i*0 Vi*R The difference i-i* is called ERROR (e)
V
![Page 68: EEE2012 Control Notes](https://reader034.fdocuments.us/reader034/viewer/2022050820/5695d2061a28ab9b0298d7b3/html5/thumbnails/68.jpg)
EEE 2002: Control
Chapter 4 –Closed loop systems 68/132
So we would like:
Ri* U i
e
+
+RLs
1
Where e0 ii*
Ri* U i
e
+
+
+ -
RLs
1
Crosscheck:
RLsG
1
iGGRii
RiiiRieV
VGi1*
***
11
1
1
1
*1*1
sR
LG
RG
i
iGRiGi
Which implies that 1R
Lnew
As new> the system is now faster. At the steady state:
2
11
12lim
11
1*
0
s
RLs
sis
RL
iis
ss .
Hence the input to the system sees the output. This is called Feedback and is the cornerstone of control systems. A system with feedback is called a CLOSED LOOP system.
V
V
![Page 69: EEE2012 Control Notes](https://reader034.fdocuments.us/reader034/viewer/2022050820/5695d2061a28ab9b0298d7b3/html5/thumbnails/69.jpg)
EEE 2002: Control
Chapter 4 –Closed loop systems 69/132
a) General closed loop systems
A more general form for a closed loop system:
HG
G
R
C
'1
'
where GGcG ' Gc= Transfer function of the controller G = TF of plant (or the open loop transfer function)
UC
GGc+
-
H
Feedback
Feedback
TF
Controller System
RE
Your task is to design the controller Gc and feedback H (if applicable). Assume H=1 and Gc=K (a constant value gain).
KR
K
KRLS
K
ssC
KRLS
K
KG
KG
R
C
sss
2lim
1 0
If K>>R then Css=2, also:
1
SKR
L
KRK
KRLS
K
R
C
So KRL
new
As new then we have a faster system.
A closed loop can also achieve a predefined steady state:
OLFT: 2
2
sG
Open loop response (desired current=1A): >> num=2;
>> den=[1 2];
>> g=tf(num,den);
![Page 70: EEE2012 Control Notes](https://reader034.fdocuments.us/reader034/viewer/2022050820/5695d2061a28ab9b0298d7b3/html5/thumbnails/70.jpg)
EEE 2002: Control
Chapter 4 –Closed loop systems 70/132
>> step(g)
CLTF: K
KCss
Ks
K
R
CGCL
22
Closed loop response (k=10): >> numc=10;
>> denc=1;
>> gc=tf(numc,denc);
>> gol=series(g,gc);
>> h=tf(1,1);
>> gcl=feedback(gol,h);
>> hold on
>> step(gcl)
0 1 2 30
0.2
0.4
0.6
0.8
1Step Response
Time (sec)
Am
plitu
de
![Page 71: EEE2012 Control Notes](https://reader034.fdocuments.us/reader034/viewer/2022050820/5695d2061a28ab9b0298d7b3/html5/thumbnails/71.jpg)
EEE 2002: Control
Chapter 4 –Closed loop systems 71/132
And for K=100:
Step Response
Time (sec)
Am
plitu
de
0 1 2 30
0.2
0.4
0.6
0.8
1
Step Response
Time (sec)
Am
plitu
de
0 1 2 30
0.2
0.4
0.6
0.8
1
![Page 72: EEE2012 Control Notes](https://reader034.fdocuments.us/reader034/viewer/2022050820/5695d2061a28ab9b0298d7b3/html5/thumbnails/72.jpg)
EEE 2002: Control
Chapter 4 –Closed loop systems 72/132
Ri* U i
e
+
+
+ -
RLs
1
Tutorial
1. The RL circuit described during the lectures is to be controlled by
Simulate the system for a unit step and ramp response and crosscheck the
results using the final value theorem. Monitor the value of the error and
the control signal U.
2. Change the control strategy to:
with ksGsHRLs
sG c
,1,1
For k=1, 10 and 100 simulate the system for a unit step and ramp response
and crosscheck the results using the final value theorem. Monitor the value
of the error and the control signal U.
UC
GGc+
-
H
Feedback
Feedback
TF
Controller System
RE
![Page 73: EEE2012 Control Notes](https://reader034.fdocuments.us/reader034/viewer/2022050820/5695d2061a28ab9b0298d7b3/html5/thumbnails/73.jpg)
EEE 2002: Control
Chapter 4 –Closed loop systems 73/132
b) Higher order systems
Until now we have seen only 1st order systems with feedback. In these systems, the feedback and the controller influence the steady state and the time constant. The decrease of the time constant means that we moved the pole further into “minus infinity” area. Hence we changed the s-plane of the system. What is it going to happen if we use a 2nd order system in a feedback system?
21
1
ssG
>> num=1;
>> den=conv([1 1],[1 2]);
>> g=tf(num,den);
>> step(g)
>> hold
Current plot held
>> numc=10;
>> denc=1;
>> gc=tf(numc,denc);
>> gol=series(g,gc);
>> h=tf(1,1);
>> gcl=feedback(gol,h);
A 2nd order system with feedback becomes faster, but note that oscillations are introduced.
kss
k
kss
kGCL
2321 2
CE: 0232 kss but the general from: 02 22 nnss
Step Response
Time (sec)
Am
plitu
de
0 1 2 3 4 5 60
0.2
0.4
0.6
0.8
1
1.2
1.4
![Page 74: EEE2012 Control Notes](https://reader034.fdocuments.us/reader034/viewer/2022050820/5695d2061a28ab9b0298d7b3/html5/thumbnails/74.jpg)
EEE 2002: Control
Chapter 4 –Closed loop systems 74/132
so 32 n and kn 22
for k=10 12122 nn and hence 433.03122 :
>> [wn,z]=damp(gcl)
wn =
3.46410161513775
3.46410161513775
z =
0.43301270189222
0.43301270189222
So the feedback and the controller can completely change the location of the poles in the s-plane.
Example: 321
sss
KG
>> num=10;
>> den=conv(conv([1 1],[1 2]),[1 3]);
>> g=tf(num,den);
>> step(g)
>> hold
Current plot held
>> num=50;
>> g=tf(num,den);
>> step(g)
The open loop response for various gains is:
![Page 75: EEE2012 Control Notes](https://reader034.fdocuments.us/reader034/viewer/2022050820/5695d2061a28ab9b0298d7b3/html5/thumbnails/75.jpg)
EEE 2002: Control
Chapter 4 –Closed loop systems 75/132
The open loop system will be stable for all values of K since they do not influence the poles of the system. The response of closed loop system for K=1, 10, 100 is:
Hence the feedback may introduce instability.
0 1 2 3 4 5 60
1
2
3
4
5
6
7
8
9Step Response
Time (sec)
Am
plitu
de
0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5-4
-3
-2
-1
0
1
2
3
4
5
Step Response
Time (sec)
Am
plit
ude
![Page 76: EEE2012 Control Notes](https://reader034.fdocuments.us/reader034/viewer/2022050820/5695d2061a28ab9b0298d7b3/html5/thumbnails/76.jpg)
EEE 2002: Control
Chapter 4 –Closed loop systems 76/132
To understand why we have these changes solve the characteristic equation. We will extensively study this at the “Root Locus” chapter. Properties of feedback systems:
Minimise steady state error.
Faster system.
Less sensitive to system uncertainties.
Introduce instability (even for negative feedback).
Expensive (we need to feedback the signal, i.e. use a sensor).
![Page 77: EEE2012 Control Notes](https://reader034.fdocuments.us/reader034/viewer/2022050820/5695d2061a28ab9b0298d7b3/html5/thumbnails/77.jpg)
EEE 2002: Control
Chapter 4 –Closed loop systems 77/132
c) Systems classification
Types We already have seen some of these categories:
Electrical
Mechanical
Hydraulic
1st order
2nd order
Higher order
Overdamped
Underdamped
Critically damped
Stable
Unstable
Marginally stable and others Another way to classify control systems is to use the number of the poles at the origin of the OPEN LOOP system:
m
N
nOL
cscscss
bsbsbsKsG
21
21
This implies a pole multiplicity N at the origin. This system is called type N.
43
10
ss
ssGOL => Type 0
43
51
sss
sssGOL => Type 1
43
12
sss
sGOL => Type 2
1005
122
sss
sGOL => Type 2
1
1105
ss
sGOL => Type 5
Note: THE TYPE OF THE SYSTEM IS DIFFERENT FROM THE ORDER OF THE SYSTEM!!!
![Page 78: EEE2012 Control Notes](https://reader034.fdocuments.us/reader034/viewer/2022050820/5695d2061a28ab9b0298d7b3/html5/thumbnails/78.jpg)
EEE 2002: Control
Chapter 4 –Closed loop systems 78/132
Error constants
Consider a unity feedback system with an open loop transfer function G(s), hence:
sG
sG
sR
sCsGCL
1
sCsRsE , this implies that the transfer function between the error
and the input is:
sGsG
sG
sR
sC
sR
sE
1
1
111
And therefore sG
sRsE
1
1
From the final value theorem:
sG
ssRE
sss
1lim
0
For a unit step input:
01
1
1
1
lim0 GsG
ss
Es
ss
We define as static position error constant as Kp: 0lim0
GsGKs
p
Hence p
ssK
E
1
1
For a unit ramp:
ssGssGssG
ss
Esss
ss
1lim
1lim
1
1
lim00
2
0
We define as static velocity error constant as Ku: ssGKs
u0
lim
And therefore u
ssK
E1
For a unit parabolic input:
sGssGsssG
ss
Esss
ss 20220
3
0
1lim
1lim
1
1
lim
We define as static acceleration error constant as Ka: s GsKs
a
2
0lim
![Page 79: EEE2012 Control Notes](https://reader034.fdocuments.us/reader034/viewer/2022050820/5695d2061a28ab9b0298d7b3/html5/thumbnails/79.jpg)
EEE 2002: Control
Chapter 4 –Closed loop systems 79/132
And therefore a
ssK
E1
Types and error constants
Another way to classify control systems is by their static error with a combination of their type. Error and unit step 1tr :
Type 0: n
mOL
cscscs
bsbsbsKsG
21
21 so
'limlim21
21
00K
cscscs
bsbsbsKsGK
n
m
ssp
So p
ssK
E
1
1
Type N>0: m
N
nOL
cscscss
bsbsbsKsG
21
21
m
N
n
ssp
cscscss
bsbsbsKsGK
21
21
00limlim
So 0ssE
Error and unit ramp ttr :
Type 0: m
nOL
cscscss
bsbsbsKsG
21
21 so
0limlim21
21
00
m
n
ssu
cscscs
bsbsbsKsssGK
So ssE
Type 1: m
nOL
cscscss
bsbsbsKsG
21
21
'limlim21
21
00K
cscscss
bsbsbsKsssGK
m
n
ssu
So v
ssK
E1
Type N>1: m
N
nOL
cscscss
bsbsbsKsG
21
21
m
N
n
ssa
cscscss
bsbsbsKsssGK
21
21
00limlim
So 0ssE
![Page 80: EEE2012 Control Notes](https://reader034.fdocuments.us/reader034/viewer/2022050820/5695d2061a28ab9b0298d7b3/html5/thumbnails/80.jpg)
EEE 2002: Control
Chapter 4 –Closed loop systems 80/132
Error and unit parabolic input 2
2
1ttr :
Similarly we find that: Type 0: ssE
Type 1: ssE
Type 2: a
ssK
E1
Type N>2: 0ssE
From the above the following table of steady state error can be derived:
Input STEP: r(t)=1
RAMP: r(t)=t
Parabola: r(t)= ½
t2
Type 0 pK1
1
Type 1 0 vK
1
Type 2 0 0 aK
1
Hence if we want to decrease the error we have to increase the type of the system.
![Page 81: EEE2012 Control Notes](https://reader034.fdocuments.us/reader034/viewer/2022050820/5695d2061a28ab9b0298d7b3/html5/thumbnails/81.jpg)
EEE 2002: Control
Chapter 4 –Closed loop systems 81/132
Tutorial 1. Find the Closed Loop (CL) (with unity feedback, and GC(s)=1) Transfer
Function (TF) when the Open Loop (OL) TF is:
a. 1
1
sGA
b. 1
12
ss
GB
c. 11
12
sss
GC
2. In each case find the CL pole and zero location.
3. Using the final value theorem find the steady state value of the CL
response for a unit step and ramp input.
4. Find the error TF in each case.
5. Using the final value theorem find the steady state value of the error for
a unit step and ramp input.
6. Using Simulink crosscheck your previous answer.
7. State the order and type of each system.
8. Find the error constants Kp and Kv.
9. For each closed loop system; A,B,C double the gain GC(s) and assess the
response to a unit step input for each system.
10. Using Matlab plot the pole-zero map of the CLTF and crosscheck this
with your findings in 9.
11. Repeat 9 and 10 for another doubling in gain, which systems can be
forced into instability?
12. Find the gain for which the 3rd order system is marginally stable.
![Page 82: EEE2012 Control Notes](https://reader034.fdocuments.us/reader034/viewer/2022050820/5695d2061a28ab9b0298d7b3/html5/thumbnails/82.jpg)
EEE 2002: Control
Chapter 5 –PID Control 82/132
Chapter 5. PID control
Proportional Integral Differential. A PID Controller modifies the error signal fed to the plant. From previous table, if the type is 0 then we have a steady state error for a step input. Also by increasing K, we increase the oscillations and we may cause instability to the system. Hence we need another solution.
Consider the system: 321
sss
KsG for K=10:
To increase the type of the OLTF (which is GCxG) we add an integrator:
sKKsG ipc
1)( =>
321
1)(
ssss
KsKsG
ip
OL
This is the so-called PI controller (proportional + integrating).
0 2 4 6 8 100
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1
error
output
![Page 83: EEE2012 Control Notes](https://reader034.fdocuments.us/reader034/viewer/2022050820/5695d2061a28ab9b0298d7b3/html5/thumbnails/83.jpg)
EEE 2002: Control
Chapter 5 –PID Control 83/132
Further increase of Kp, Ki:
1
s +6s +11s+63 2
Transfer Fcn
StepScope
10
Kp
1
Ki
1
s
Integrator
0 2 4 6 8 100
0.2
0.4
0.6
0.8
1
1.2
1.4
Ki=0
Ki=2
Ki=6
Proportional gain
Integrator gain
![Page 84: EEE2012 Control Notes](https://reader034.fdocuments.us/reader034/viewer/2022050820/5695d2061a28ab9b0298d7b3/html5/thumbnails/84.jpg)
EEE 2002: Control
Chapter 5 –PID Control 84/132
0 2 4 6 8 100
0.2
0.4
0.6
0.8
1
1.2
1.4
Kp=10, K
i=6
Kp=20, K
i=10
![Page 85: EEE2012 Control Notes](https://reader034.fdocuments.us/reader034/viewer/2022050820/5695d2061a28ab9b0298d7b3/html5/thumbnails/85.jpg)
EEE 2002: Control
Chapter 5 –PID Control 85/132
Check the derivative of error, i.e. the rate of change of e:
Maximum value of De just before e=0. So De can control the oscillations:
sKs
KKsG dipc 1
)( => 321
1)(
2
ssss
KsKKssG
ipd
OL
0 2 4 6 8 10-1.5
-1
-0.5
0
0.5
1
De
e
10
ki
1
kd
1
s +6s +11s+63 2
Transfer FcnStepScope
20
Kp
1/s
Integrator
du/dt
Derivative
![Page 86: EEE2012 Control Notes](https://reader034.fdocuments.us/reader034/viewer/2022050820/5695d2061a28ab9b0298d7b3/html5/thumbnails/86.jpg)
EEE 2002: Control
Chapter 5 –PID Control 86/132
Since I have no oscillations I can increase Kp a little bit more to make the system faster:
0 2 4 6 8 100
0.2
0.4
0.6
0.8
1
1.2
1.4
Kd=0
Kd=5
Kd=10
0 2 4 6 8 100
0.2
0.4
0.6
0.8
1
1.2
1.4
Kp=20
Kp=50
![Page 87: EEE2012 Control Notes](https://reader034.fdocuments.us/reader034/viewer/2022050820/5695d2061a28ab9b0298d7b3/html5/thumbnails/87.jpg)
EEE 2002: Control
Chapter 5 –PID Control 87/132
0 2 4 6 8 100
0.2
0.4
0.6
0.8
1
1.2
1.4
0 2 4 6 8 10-20
-10
0
10
20
30
40
50OutputP output
0 2 4 6 8 100
0.5
1
1.5
0 2 4 6 8 10-20
-10
0
10OutputD output
![Page 88: EEE2012 Control Notes](https://reader034.fdocuments.us/reader034/viewer/2022050820/5695d2061a28ab9b0298d7b3/html5/thumbnails/88.jpg)
EEE 2002: Control
Chapter 5 –PID Control 88/132
Another way to write the PID controller:
s
K
K
sK
KKsK
s
KKsG
p
d
p
ipd
ipCL
11)(
sT
sTKsG d
i
pCL
11)(
0 2 4 6 8 100
0.5
1
1.5
0 2 4 6 8 100
2
4
6
OutputI Output
![Page 89: EEE2012 Control Notes](https://reader034.fdocuments.us/reader034/viewer/2022050820/5695d2061a28ab9b0298d7b3/html5/thumbnails/89.jpg)
EEE 2002: Control
Chapter 5 –PID Control 89/132
a) Tuning of PID controllers
Trial and error.
Ziegler Nichols I
Ziegler Nichols II
Root locus
Frequency response
Other advanced control methods
Trial and error:
P: Faster system, in some cases reduces the error (can cause instability). I: Reduces the steady state error, increases the number of oscillations. D: Reduces the oscillations.
Ziegler Nichols I
Assume a system with no delays (we do not study these systems) and with no-complex conjugate poles.
u(t) c(t)Plant
Its open loop step response may look like (obtained experimentally of from simulations):
![Page 90: EEE2012 Control Notes](https://reader034.fdocuments.us/reader034/viewer/2022050820/5695d2061a28ab9b0298d7b3/html5/thumbnails/90.jpg)
EEE 2002: Control
Chapter 5 –PID Control 90/132
This can be modelled as:
1)(
)(
Ts
Ke
sU
sC Ls
Based on that we have the following table:
Type of controller
Kp Ti Td
P LT 0
PI LT9.0
3.0L 0
PID LT2.1 L2 L5.0
Ziegler Nichols II
1. Initially assume Ki=Kd=0. 2. Increase Kp until the system is marginally stable.
3. Record the Kp=Kcr and the frequency of oscillations: 4. Set the controller gains as follows based on the critical gain and
frequency of oscillation;
Type of controller
Kp Ti Td
P crK5.0 0
0
t,s
c(t
)
Pcr
![Page 91: EEE2012 Control Notes](https://reader034.fdocuments.us/reader034/viewer/2022050820/5695d2061a28ab9b0298d7b3/html5/thumbnails/91.jpg)
EEE 2002: Control
Chapter 5 –PID Control 91/132
PI crK45.0 crP
2.1
1 0
PID crK6.0 crP5.0
crP125.0
These methods aim at achieving an overshoot of 25%.
![Page 92: EEE2012 Control Notes](https://reader034.fdocuments.us/reader034/viewer/2022050820/5695d2061a28ab9b0298d7b3/html5/thumbnails/92.jpg)
EEE 2002: Control
Chapter 5 –PID Control 92/132
Zeigler Nichols 2 Class example
A PID controller ( )
is connected in series with a plant
whose transfer function is
( )
( )( ) and a unity feedback loop is imposed.
Using the Zeigler Nichols 2 method determine the values of Kp, Ti and Td
ZN2 STEPS
1. Set ( ) ( )
2. Find closed loop transfer function;
( )
( )
( )( )
( )
( )( )
3. Find the critical gain at which the system is marginally stable;
Marginally stable when poles lie on the imaginary axis so
substitute into C.E;
4. Separate variables in
√
![Page 93: EEE2012 Control Notes](https://reader034.fdocuments.us/reader034/viewer/2022050820/5695d2061a28ab9b0298d7b3/html5/thumbnails/93.jpg)
EEE 2002: Control
Chapter 5 –PID Control 93/132
=
As √
Critical gain for marginal stability Kcr=120
5. Substitute KCR into ZN2 table for PID controller
6. Proportional gain:
KP=0.6 x KCR = 0.6 x 120 = 72
7. Period of natural oscillation
So
√
8. Integral and differential gain
Ti=0.5 x PCR = 0.8112 seconds
Td=0.125 x PCR = 0.2028 seconds
![Page 94: EEE2012 Control Notes](https://reader034.fdocuments.us/reader034/viewer/2022050820/5695d2061a28ab9b0298d7b3/html5/thumbnails/94.jpg)
EEE 2002: Control
Chapter 5 –PID Control 94/132
Zeigler Nichols 2 Tutorial
A plant is described by the transfer function;
( )
( )( )
This system is to be controlled by a PID controller and a unity feedback loop implemented. Tune the PID controller using the Zeigler Nichols II method and state your calculated values of Kp, Ti and Td.
![Page 95: EEE2012 Control Notes](https://reader034.fdocuments.us/reader034/viewer/2022050820/5695d2061a28ab9b0298d7b3/html5/thumbnails/95.jpg)
EEE 2002: Control
Chapter 5 –PID Control 95/132
b) Root locus method
With the root locus method we specifically target a pole location at the s-plane, i.e. we target a desired damping factor, natural and damped frequency.
Root locus class example
The OLTF is 3411
1)(
2
sssG use a PI controller.
34113411
1)(
22
sss
KsK
sss
KKsG
ipipOL
ip
ip
CLKsKsss
KsKsG
3411)(
2
This is a 3rd order system = 2nd order x 1st order: CE: 222 23411 nnip ssasKsKsss
222323 223411 nnnnip asasasKKsss
2
2234
211
ni
nnp
n
aK
aK
a
Assume that the design specs are:
5.0
6
n
So:
180
100
5
36
36634
611
i
p
i
p
K
K
a
aK
aK
a
Matlab: CL response, no controller: >> num=1;
>> den=[1 11 -34];
>> g=tf(num,den);
>> gcl=feedback(g,1);
>> step(gcl)
![Page 96: EEE2012 Control Notes](https://reader034.fdocuments.us/reader034/viewer/2022050820/5695d2061a28ab9b0298d7b3/html5/thumbnails/96.jpg)
EEE 2002: Control
Chapter 5 –PID Control 96/132
CL, response, PI controller: >> kp=100;
>> ki=180;
>> gc=tf([kp ki],[1 0]);
>> gol=series(gc,g);
>> gcl=feedback(gol,1);
>> step(gcl)
Homework:
Find the PID gains : the CLTF of 166
1)(
2
sssG has
5.0
6
n and a real pole
at a=-5. Solution: Kp=50, Ki=180, Kd=5.
0 0.2 0.4 0.6 0.8 10
0.05
0.1
0.15
0.2
0.25
0.3
0.35Step Response
Time (sec)
Am
plitu
de
0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.60
0.5
1
1.5
2Step Response
Time (sec)
Am
plitu
de
![Page 97: EEE2012 Control Notes](https://reader034.fdocuments.us/reader034/viewer/2022050820/5695d2061a28ab9b0298d7b3/html5/thumbnails/97.jpg)
EEE 2002: Control
Chapter 5 –PID Control 97/132
Root Locus Tutorial
A plant is controlled by a proportional controller with gain K and a unity
gain feedback loop. The plant transfer function is ( )
( )( )
If a ramp input where r(t)=5t is used, what is the steady state error with a proportional gain K=1.5? What value of gain K will ensure an error ESS<0.1?
15.01
sss
KsGOL , Unity feedback and input: r(t)=5t
a) If K=1.5, find the steady state error b) The system must have steady state error, Ess<0.1 find the value of K
C(s)
C(s)
15.01 sss
K
2
5
s sR
Ksss
K
sR
sC
15.01
Ksss
K
ssC
15.01
52
sCsRsE
Ksss
K
ssE
15.011
52
Ksss
ss
ssE
15.01
15.015
Ksss
ss
ssE
sss
15.01
15.015lim
0
KEss
5
a) ...33.35.1
5ssE
b) 1.0ssE 1.05
K50K
![Page 98: EEE2012 Control Notes](https://reader034.fdocuments.us/reader034/viewer/2022050820/5695d2061a28ab9b0298d7b3/html5/thumbnails/98.jpg)
EEE 2012: Control
Chapter 6 – The Root Locus Method Page 98 of 132
Chapter 6. The Root Locus Method
“The Root Locus is the path (locus) of the roots of the characteristic equation traced out in the s-plane as a system parameter is changed” The dynamic response depends on system’s and controller’s parameters , therefore the closed loop pole location depends on system’s and controller’s parameters. The system designer will find the controller gain(s) such as we have a satisfactory performance, this equates to finding the controller gain(s) such as we have a satisfactory pole location. To do that we need a good understating of how poles move in the s-plane as the controller parameters are changed.
![Page 99: EEE2012 Control Notes](https://reader034.fdocuments.us/reader034/viewer/2022050820/5695d2061a28ab9b0298d7b3/html5/thumbnails/99.jpg)
EEE 2012: Control
Chapter 6 – The Root Locus Method Page 99 of 132
a) Root locus on a 1st order system
First order system with a proportional controller with gain k and a unity gain feedback loop;
( )
Gc(S)=K H(s)=1
Forward path: ( )
Closed loop: ( )
The pole location s + 1 + k = 0
then s = -1 -k As k increases the pole moves further in the negative x direction
Properties:
Root locus starts from the open loop pole location.
Ends at .
This implies that as K is increased the system becomes faster (we
already know this from before)
-10
-8
-6
-4
-2
0
2
4
6
8
10
-10 -8 -6 -4 -2 0 2 4 6 8 10
jω
σ
X
![Page 100: EEE2012 Control Notes](https://reader034.fdocuments.us/reader034/viewer/2022050820/5695d2061a28ab9b0298d7b3/html5/thumbnails/100.jpg)
EEE 2012: Control
Chapter 6 – The Root Locus Method Page 100 of 132
Example 2 Second order system with a proportional controller with gain k and a unity gain feedback loop;
( )
Gc(S)=K H(s)=1
Forward path: ( )
Closed loop: ( )
√ ( )
Two real poles until ( ( )) i.e. k > 0.25 After this we have
√ hence a complex pole
Properties:
Starts from the OL pole location.
Collides at a point.
Moves on a straight line perpendicular to the real axis.
Symmetrical with respect to the real axis
Ends at -1.5 ±
-6
-4
-2
0
2
4
6
-2.5 -2 -1.5 -1 -0.5 0 0.5
jω
σ
X X
X X
![Page 101: EEE2012 Control Notes](https://reader034.fdocuments.us/reader034/viewer/2022050820/5695d2061a28ab9b0298d7b3/html5/thumbnails/101.jpg)
EEE 2012: Control
Chapter 6 – The Root Locus Method Page 101 of 132
Root locus tutorial
For each the following open loop systems;
( )
( )
( ) ( )
a) Input the open loop transfer function into Matlab
hint: >>g1=tf(1,[1 2 2])
b) Use the Matlab command >>rlocus(g1) to plot the open loop poles,
zeros and root locus for each.
c) Sketch each root locus and discuss the general patterns found, i.e
direction of travel of roots, asymptotes, effect of zeros, collision
points….
![Page 102: EEE2012 Control Notes](https://reader034.fdocuments.us/reader034/viewer/2022050820/5695d2061a28ab9b0298d7b3/html5/thumbnails/102.jpg)
EEE 2012: Control
Chapter 6 – The Root Locus Method Page 102 of 132
G1(s)
Second order system, unity feedback loop, proportional controller gain K;
kss
kG
ss
kG CLOL
2222 22 so pole location: 0222 kss :
Properties:
Starts from the OL pole location.
Starts already with complex poles so no collision.
Moves on a straight line perpendicular to the real axis.
Symmetrical with respect to the real axis
Ends at 1 .
-1.6 -1.4 -1.2 -1 -0.8 -0.6 -0.4 -0.2 0-5
0
5Root Locus
Real Axis
Imagin
ary
Axis
![Page 103: EEE2012 Control Notes](https://reader034.fdocuments.us/reader034/viewer/2022050820/5695d2061a28ab9b0298d7b3/html5/thumbnails/103.jpg)
EEE 2012: Control
Chapter 6 – The Root Locus Method Page 103 of 132
G2(s)
Third order system with a unity gain feedback loop;
ksss
kG
sss
kG CLOL
61166116 2323 so pole location:
06116 23 ksss :
Properties:
Starts from the OL pole location.
Two poles collide and then they follow asymptotically a line which is
not perpendicular to the real axis.
Symmetrical with respect to the real axis
The other pole goes to .
-8 -7 -6 -5 -4 -3 -2 -1 0 1-5
-4
-3
-2
-1
0
1
2
3
4
5Root Locus
Real Axis
Imag
inar
y A
xis
![Page 104: EEE2012 Control Notes](https://reader034.fdocuments.us/reader034/viewer/2022050820/5695d2061a28ab9b0298d7b3/html5/thumbnails/104.jpg)
EEE 2012: Control
Chapter 6 – The Root Locus Method Page 104 of 132
G3(s)
Third order system with a unity gain feedback loop;
226116
22
6116
22223
2
23
2
ssksss
sskG
sss
sskG CLOL so pole location:
02621160226116 23223 ksksksssksss
Properties:
Starts from the OL pole location.
Two poles collide and then they follow asymptotically a line which is
not perpendicular to the real axis.
They do not diverge to infinity but they converge to the two
complex zeros.
Symmetrical with respect to the real axis
The other pole goes to .
-4.5 -4 -3.5 -3 -2.5 -2 -1.5 -1 -0.5 0
-1
-0.5
0
0.5
1
Root Locus
Real Axis
Imagin
ary
Axis
![Page 105: EEE2012 Control Notes](https://reader034.fdocuments.us/reader034/viewer/2022050820/5695d2061a28ab9b0298d7b3/html5/thumbnails/105.jpg)
EEE 2012: Control
Chapter 6 – The Root Locus Method Page 105 of 132
b) Summary
Hence it can be seen that the root location depends on the order and the number of zeros of the system. The procedure of finding the closed loop pole location for various values of the proportional gain k is known as root locus method. We can find that location numerically or graphically (W. R. Evans). Matlab provides the command rlocus(num_OL,den_OL) but it is better to have a deeper understanding of the graphical method. This method can also be used for other parameters (apart from the proportional controller) but is rather tricky. Steps;
1. To include the effect of the feedback TF we name as OLTF the :)(sHsG
G(s)
H(s)
We always start from the location of the OL poles (i.e. poles of G(s)H(s)):
HGHG
HG
H
HGG
G
H
H
G
G
G
G
CLNkNDD
DkN
D
NkND
kN
D
N
D
Nk
D
Nk
GH
GG
11
0: HGHG NkNDDCE . For k=0 (first value): 0: HGDDCE , so poles of
G(s)H(s). We use the symbol “x” to denote OL poles. We use the symbol “o” to denote OL zeros. We use the symbol “□” to denote CL poles. No need for CL zeros as they do not move.
![Page 106: EEE2012 Control Notes](https://reader034.fdocuments.us/reader034/viewer/2022050820/5695d2061a28ab9b0298d7b3/html5/thumbnails/106.jpg)
EEE 2012: Control
Chapter 6 – The Root Locus Method Page 106 of 132
c) Angle and magnitude conditions
1)()(0)()(1:)()(1
)(
)(
)(
sHsGsHsGCE
sHsG
sG
sR
sC
This implies that 1)()( sHsG and ,12180)()(arg nsHsG ...2,1,0n
For example:
321
1
pspsps
zsksHsG
1
1p
3
2
1
3p
2p
1z
s
1B
1A
2A
3A
![Page 107: EEE2012 Control Notes](https://reader034.fdocuments.us/reader034/viewer/2022050820/5695d2061a28ab9b0298d7b3/html5/thumbnails/107.jpg)
EEE 2012: Control
Chapter 6 – The Root Locus Method Page 107 of 132
321
1
1321)arg(
AAA
kBsHsG
s
Every point that belongs onto the Root Locus (RL) must satisfy the magnitude and angle criteria.
For the system that we studied before 232
ss
kGOL
kss
kGCL
232
We will analyse the root locus for two test points; js 5.18.1 and js 5.15.1 .
The OL case is 232
ss
kGOL so I have two poles at -1 and -2 and no zeros:
1
1p
2
jxs 5.1
2A
1A
2p
1
jxs 5.1
1A
xs 1s
1x
5.1
![Page 108: EEE2012 Control Notes](https://reader034.fdocuments.us/reader034/viewer/2022050820/5695d2061a28ab9b0298d7b3/html5/thumbnails/108.jpg)
EEE 2012: Control
Chapter 6 – The Root Locus Method Page 108 of 132
1
5.1atan180
1
5.1180tan 11
xx
22
1 15.1 xA
2
xjs 2
2A
2s
xs
5.1
x2
xx 2
5.1atan
2
5.1tan 22
22
2 25.1 xA
For the point js 5.18.1 :
o
1 07.11818.1
5.1atan180
o
2 40.828.12
5.1atan
So 18047.20021 so that point does not belong to the root locus.
For the point js 5.15.1 :
o
1 43.10818.1
5.1atan180
o
2 56.718.12
5.1atan
So 18021 so that point does belong to the root locus.
By using the magnitude condition we can find the gain:
![Page 109: EEE2012 Control Notes](https://reader034.fdocuments.us/reader034/viewer/2022050820/5695d2061a28ab9b0298d7b3/html5/thumbnails/109.jpg)
EEE 2012: Control
Chapter 6 – The Root Locus Method Page 109 of 132
58.15.125.125.1
58.115.15.115.1
2222
2
2222
1
xA
xA
5.21 21
21
AAkAA
k
Analytically: For the test point s=-1.8±1.5j
023
112
ss
kGH 09.04120232 kj- .-kss
This is not valid as -2.41 + 0.9j 0 For the test point s=-1.5±1.5j
023
112
ss
kGH 05.20232 k-kss
This is a valid point, therefore s=-1.5±1.5j belongs on the root locus.
![Page 110: EEE2012 Control Notes](https://reader034.fdocuments.us/reader034/viewer/2022050820/5695d2061a28ab9b0298d7b3/html5/thumbnails/110.jpg)
EEE 2012: Control
Chapter 6 – The Root Locus Method Page 110 of 132
d) Simple root locus
For the open loop transfer function;
21
sss
ksG .
1. The number of branches equals the number of poles. The root locus is
always symmetrical with respect to the real axis and the branches start from poles and end up to zeros or infinity (infinite zero). In this specific case I have three branches, which will diverge to imaginary zeros at infinity.
2. There are three real OL poles with angles: 2arg,1arg,arg sss .
Applying the angle criteria; If s>0 then ,0arg s ,01arg s 02arg s and therefore that area
does not belong to the root locus as 02arg1argarg sss .
If s<0 and s>-1 then ,180arg s ,01arg s 02arg s and therefore
that area belongs to the root locus as 180001802arg1argarg sss .
If s<-1 and s>-2 then ,180arg s ,1801arg s 02arg s and
therefore that area does not belong to the root locus as 36001801802arg1argarg sss .
If s<-3 then ,180arg s ,1801arg s 1802arg s and therefore that
area belongs to the root locus as 1121801801801802arg1argarg sss with 1n .
So the root locus on the real axis is:
![Page 111: EEE2012 Control Notes](https://reader034.fdocuments.us/reader034/viewer/2022050820/5695d2061a28ab9b0298d7b3/html5/thumbnails/111.jpg)
EEE 2012: Control
Chapter 6 – The Root Locus Method Page 111 of 132
0s1s2s
3. We have seen that when we numerically calculate the root locus at
some point the poles from -1 and -2 will collide. To find this break-out point we rearrange the CE as:
sA
sBkskAsBsfCE 00)(: and then we find the 1st
derivative with respect to s and we equate the result with zero:
0ds
dk. Hence in this case: 21021 ssskksss and
hence
42.0
5.1026321 2 ssssss
ds
dk. Since the 2nd
point only belongs to the root locus the other one (-1.5) is ignored:
0s1s2s
Break away point
The same procedure can be used to find break-in points.
![Page 112: EEE2012 Control Notes](https://reader034.fdocuments.us/reader034/viewer/2022050820/5695d2061a28ab9b0298d7b3/html5/thumbnails/112.jpg)
EEE 2012: Control
Chapter 6 – The Root Locus Method Page 112 of 132
4. To find the asymptote that the locus of these two poles will follow:
The angle of the asymptote is:
zp nn
n
12180, where pn is the number of
finite poles and zn is the number of finite zeros. Hence in this case
0,3 zp nn and hence their angle is 180/3=60. The point of
intersection with the real axis is:
zp
n
i
i
n
i
i
nn
zp
s
zp
11 => 1
3
210
s :
0s1s2s
Asymptote
5.1
5. The point of intersection of the root locus with the imaginary axis can
be calculated by using the CE at js . So in this case
023023021 2323
kjjksssksssjs
So
4.1
0023 and 60603 2 kkk .
6. Now use this information to sketch the root locus;
![Page 113: EEE2012 Control Notes](https://reader034.fdocuments.us/reader034/viewer/2022050820/5695d2061a28ab9b0298d7b3/html5/thumbnails/113.jpg)
EEE 2012: Control
Chapter 6 – The Root Locus Method Page 113 of 132
0s1s2s
5.1
![Page 114: EEE2012 Control Notes](https://reader034.fdocuments.us/reader034/viewer/2022050820/5695d2061a28ab9b0298d7b3/html5/thumbnails/114.jpg)
EEE 2012: Control
Chapter 6 – The Root Locus Method Page 114 of 132
In Matlab: >> num=1; den=[1 3 2 0]; rlocus(num,den)
-6 -4 -2 0 2-4
-3
-2
-1
0
1
2
3
4Root Locus
Real Axis
Ima
gin
ary
Axis
![Page 115: EEE2012 Control Notes](https://reader034.fdocuments.us/reader034/viewer/2022050820/5695d2061a28ab9b0298d7b3/html5/thumbnails/115.jpg)
EEE 2012: Control
Chapter 6 – The Root Locus Method Page 115 of 132
Example: Find the root locus of: 32
)2(2
ss
sksG , H(s)=1:
1. I have two complex poles and 1 real zero.
2. Hence I have two branches that are symmetrical with respect to the
real axis.
3. Place the OL poles and zeros:
1
j4.1
j4.1
2
4. Find the loci on the real axis:
a. For s>-2 I have no poles or zeros so even number and hence
that area does not belong to the locus.
b. From to -2 I have 1 zeros, i.e. odd number and hence this
belongs to the locus:
![Page 116: EEE2012 Control Notes](https://reader034.fdocuments.us/reader034/viewer/2022050820/5695d2061a28ab9b0298d7b3/html5/thumbnails/116.jpg)
EEE 2012: Control
Chapter 6 – The Root Locus Method Page 116 of 132
1
j4.1
j4.1
2
5. Determine break in and out points:
2
320)2(32:
32
)2( 22
2 s
sskskssCE
ss
sksG
268.0
73.30142 sss
ds
dk. The acceptable value is 73.3s
![Page 117: EEE2012 Control Notes](https://reader034.fdocuments.us/reader034/viewer/2022050820/5695d2061a28ab9b0298d7b3/html5/thumbnails/117.jpg)
EEE 2012: Control
Chapter 6 – The Root Locus Method Page 117 of 132
6. Determine behaviour of complex poles:
These poles will move towards the real axis by increasing k. After the collision at -3.73 one will move towards to the zero at -2and the
other will diverge to -. The angle of departure of the complex poles is calculated by; 180 – (angle with respect to other poles) + (angle with respect to other zeros);
180 :
1
j4.1
j4.1
2
1
1
902
So 1455590180180 121 . The angle of departure for
-1-1.4j is -145 since the locus is symmetric.
![Page 118: EEE2012 Control Notes](https://reader034.fdocuments.us/reader034/viewer/2022050820/5695d2061a28ab9b0298d7b3/html5/thumbnails/118.jpg)
EEE 2012: Control
Chapter 6 – The Root Locus Method Page 118 of 132
1
j4.1
j4.1
2
Now we simply have to draw the locus from the complex poles to the real axis:
1
j4.1
j4.1
2
With Matlab: >> num=[1 2]; den=[1 2 3]; rlocus(num,den)
145°
-145°
![Page 119: EEE2012 Control Notes](https://reader034.fdocuments.us/reader034/viewer/2022050820/5695d2061a28ab9b0298d7b3/html5/thumbnails/119.jpg)
EEE 2012: Control
Chapter 6 – The Root Locus Method Page 119 of 132
-4 -3 -2 -1 0-2
-1.5
-1
-0.5
0
0.5
1
1.5
2Root Locus
Real Axis
Ima
gin
ary
Axis
![Page 120: EEE2012 Control Notes](https://reader034.fdocuments.us/reader034/viewer/2022050820/5695d2061a28ab9b0298d7b3/html5/thumbnails/120.jpg)
EEE 2012: Control
Chapter 6 – The Root Locus Method Page 120 of 132
Design procedure summary
1. The branches are always symmetric with respect to the real axis.
2. The number of the branches equals the number of the OL poles.
3. Every branch starts from an OL pole (k=0) and ends at an OL zero
(k=). If we have more poles than zeros then we assume that there
are imaginary zeros at .
4. There are zp nn asymptotes where they intersect on the real axis at:
∑ ∑
5. The angle between them is ( )
6. According the angle criterion, a root locus exists on the real axis if the number of poles and zeros to the right is odd. Complex poles do not contribute.
7. The break in and break out points occur when
8. The intersection with the imaginary axis occurs at . Substitute this into the CE.
9. Departure angle from a complex pole is found by using the angle
criterion: ∑ ∑ In matlab the command rlocus(g) is very useful!
![Page 121: EEE2012 Control Notes](https://reader034.fdocuments.us/reader034/viewer/2022050820/5695d2061a28ab9b0298d7b3/html5/thumbnails/121.jpg)
EEE 2012: Control
Chapter 7 – Design Based Root Locus Page 121 of 132
Chapter 7. Design Based Root Locus
Lead Control (equivalent to PD control)
Used when the steady state properties of the system are ok but there is poor performance, i.e. we want to move the poles of the system without greatly altering the gain.
General form: abbs
asKsG cc
,
Other forms of Lead controller: 1,1
1
1
1
Ts
Ts
KTsa
TsKsG ccc , so:
bTTa /1,/1
Design strategy: 1. Express the design requirements as a desired pole location. 2. Plot the uncompensated root locus and check if there is a gain that
brings the system to that location, i.e. if a P controller is enough. 3. If not, place the zero at the real part of the desired pole location, i.e.
na .
4. Find the value of the pole by using the angle criterion:, i.e. zerosrealpolesreal argarg
5. Find the value of Kc by using the magnitude criterion, i.e. 1GH .
6. Check to see if the new compensated system satisfied the design specifications.
7. If not repeat by placing the zero at a different location.
![Page 122: EEE2012 Control Notes](https://reader034.fdocuments.us/reader034/viewer/2022050820/5695d2061a28ab9b0298d7b3/html5/thumbnails/122.jpg)
EEE 2012: Control
Chapter 7 – Design Based Root Locus Page 122 of 132
Example – Lead Control
Assume the system:
2
4
sssG
Create a closed loop (unity feedback strategy) such as the closed loop system will have sradn /4 and 5.0 ;
EGc(s)
RG(s)
U C
If a simple proportional controller is used then the CLTF is:
'2
'
42
42 Kss
K
Kss
KsGCL
0'2
21lim
'2
2
'2
22
2
02
2
2
2
Kss
ss
ssE
Kss
sssRsE
Kss
ss
sR
sE
sss
And the root locus is: >>num=[1]
>>den=[1 2 0]
>>rlocus(num,den)
![Page 123: EEE2012 Control Notes](https://reader034.fdocuments.us/reader034/viewer/2022050820/5695d2061a28ab9b0298d7b3/html5/thumbnails/123.jpg)
EEE 2012: Control
Chapter 7 – Design Based Root Locus Page 123 of 132
Since the required damping factor is 0.5 this implies an angle of 60o which
will give 73.15.0122 2 dn and this is achieved for K’=2:
-2.5 -2 -1.5 -1 -0.5 0-1.5
-1
-0.5
0
0.5
1
1.5Root Locus
Real Axis
Ima
gin
ary
Axis
Root Locus
Real Axis
Ima
gin
ary
Axis
-2.5 -2 -1.5 -1 -0.5 0-2
-1.5
-1
-0.5
0
0.5
1
1.5
2
System: sysGain: 4
Pole: -1 + 1.73iDamping: 0.5
Overshoot (%): 16.3Frequency (rad/sec): 2
![Page 124: EEE2012 Control Notes](https://reader034.fdocuments.us/reader034/viewer/2022050820/5695d2061a28ab9b0298d7b3/html5/thumbnails/124.jpg)
EEE 2012: Control
Chapter 7 – Design Based Root Locus Page 124 of 132
But for this gain and damping factor the natural frequency is not the required, i.e. the system is slower. So we replace the proportional controller with a more general:
The new open loop TF is: 2
4
ssbs
asKG
The location of the open loop poles and zeros and the desired closed loop
poles 322 j is:
0-2
3.464
-3.464
One strategy is to place the open loop zero such as it will be equal to the real part of the desired closed loop, i.e. a=-2 (in this example it just happens to coincide with the open loop pole). This will leave us with the location of the pole. Let’s assume that the controller pole is placed at x=-b and create an angle with the real axis:
![Page 125: EEE2012 Control Notes](https://reader034.fdocuments.us/reader034/viewer/2022050820/5695d2061a28ab9b0298d7b3/html5/thumbnails/125.jpg)
EEE 2012: Control
Chapter 7 – Design Based Root Locus Page 125 of 132
0-2
3.464
-3.464
-b
o90
In order for the closed loop pole to be on the desired point the new characteristic equation must satisfy the angle condition:
301801201809090 and hence b=4.
So the OLTF is:
4
4
2
4
4
2
ssK
sss
sKG
Now we can use the magnitude criterion to find the gain:
1
4
4
322
jsss
K (as 01 GH )
![Page 126: EEE2012 Control Notes](https://reader034.fdocuments.us/reader034/viewer/2022050820/5695d2061a28ab9b0298d7b3/html5/thumbnails/126.jpg)
EEE 2012: Control
Chapter 7 – Design Based Root Locus Page 126 of 132
4116
4
1644
4
1322322
4
KK
K
jj
K
Root Locus
Real Axis
Ima
gin
ary
Axis
-4 -3 -2 -1 0-5
0
5
System: gol
Gain: 4
Pole: -2 + 3.46i
Damping: 0.5
Overshoot (%): 16.3
Frequency (rad/sec): 4
![Page 127: EEE2012 Control Notes](https://reader034.fdocuments.us/reader034/viewer/2022050820/5695d2061a28ab9b0298d7b3/html5/thumbnails/127.jpg)
EEE 2012: Control
Chapter 7 – Design Based Root Locus Page 127 of 132
a) Lag Control
Used when the steady state properties of the system are not satisfactory but the pole location is ok.
General form: babs
asKsG cc
,
Other forms of Lead controller: 1,1
1
1
1
Ts
Ts
KTs
TsKsG ccc , so:
bTTa /1,/1
Design strategy: 1. Express the design requirements as gain (usually as Kv), pole location
(damping factor and natural frequency). 2. Check if it is possible to achieve these with a simple proportional
controller. 3. If the location is possible but at that point Kv is too small (i.e. big
error for a ramp input) then add a lag controller such as the controller pole will be close to the origin (for example at -0.001) and the zero to the left of the pole as many times as it is required to achieve the desired Kv.
4. Use the magnitude criterion to find the necessary controller gain. 5. Check to see if you have achieved the design specifications.
![Page 128: EEE2012 Control Notes](https://reader034.fdocuments.us/reader034/viewer/2022050820/5695d2061a28ab9b0298d7b3/html5/thumbnails/128.jpg)
EEE 2012: Control
Chapter 7 – Design Based Root Locus Page 128 of 132
Example – Lag Control
An plant transfer function is 21
06.1
ssssG . Create a control structure
that will achieve the following: 673.0,491.0 n
5vK
EGc(s)
RG(s)
U C
If a simple proportional controller is used then the CLTF is:
'21
'
06.121
06.1
Ksss
K
Ksss
KsGCL
Hence the CE is:
0'23
0'21
23
Ksss
Ksss
As this is a 3rd order system the generic CE ca be written as:
022
02
2223
22
nnnn
nn
asasas
ssas
By equating the corresponding coefficients:
34.2,06.1'
'
22
32673.0
491.02
2
aK
Ka
a
an
n
nn
n
Hence remarkably the desired pole location is achieved even without any extra proportional controller. But:
53.0
21
06.1lim
0
ssssK
sv
Hence we need to increase Kv 10 times (more or less). We use a Lag controller: We place the pole at 005.0b and hence the zero at 05.0a (as we need to increase the gain 10 times).:
![Page 129: EEE2012 Control Notes](https://reader034.fdocuments.us/reader034/viewer/2022050820/5695d2061a28ab9b0298d7b3/html5/thumbnails/129.jpg)
EEE 2012: Control
Chapter 7 – Design Based Root Locus Page 129 of 132
0181.1
121
06.1
005.0
05.0
21
06.1
005.0
05.0
58.033.0
c
js
c
cOL
K
ssss
sK
ssss
sKsG
The new Kv is:
396.5
21
0792.1
005.0
05.0lim
0
ssss
ssK
sv
And the new pole location is: >>num=1.0792*[1 0.05]
>>den=conv(conv(conv([1 0.005],[1 0]),[1 1]),[1 2])
>>gcl=feedback(g,1); damp(gcl)
Eigenvalue Damping Freq.
(rad/s)
-5.46e-002 1.00e+000 5.46e-
002
-3.06e-001 + 5.73e-001i 4.71e-001 6.50e-
001
-3.06e-001 - 5.73e-001i 4.71e-001 6.50e-
001
-2.34e+000 1.00e+000
2.34e+000
![Page 130: EEE2012 Control Notes](https://reader034.fdocuments.us/reader034/viewer/2022050820/5695d2061a28ab9b0298d7b3/html5/thumbnails/130.jpg)
EEE 2012: Control
Chapter 7 – Design Based Root Locus Page 130 of 132
b) Lead - Lag Control
This is a combination of a Lead and a lag controller (similar to PID). General form:
2
2
1
1
/1
/1
/
/1
cTs
Ts
Tcs
TsKsG cc
1
1
/
/1
Tcs
Ts
is the lead controller and
2
2
/1
/1
cTs
Ts
is the lag controller.
Example – Lead - Lag Control
An plant transfer function is 5.0
4
sssG . Create a control structure
that will achieve the following: 5,5.0 n
80vK
EGc(s)
RG(s)
U C
0-0.5
Desired pole
location
4.33
-2.5
![Page 131: EEE2012 Control Notes](https://reader034.fdocuments.us/reader034/viewer/2022050820/5695d2061a28ab9b0298d7b3/html5/thumbnails/131.jpg)
EEE 2012: Control
Chapter 7 – Design Based Root Locus Page 131 of 132
By using the lead controller design steps:
0-0.5
Desired pole
location
4.33
-2.5
o120
o90
2.5
4.33
o652
33.4tan 1
o351809012065180
35
4.33
X-2.5
68.85.2
33.435tan
x
x
Hence 47.368.8
1 c
Tc
By choosing now T2=10 so that the second pole will be close to the origin:
10
805.0
4
805.0
4
/1
/1
/
/1lim
2
2
1
1
0
c
c
cs
v
K
K
sscTs
Ts
Tcs
TssKK
And hence:
![Page 132: EEE2012 Control Notes](https://reader034.fdocuments.us/reader034/viewer/2022050820/5695d2061a28ab9b0298d7b3/html5/thumbnails/132.jpg)
EEE 2012: Control
Chapter 7 – Design Based Root Locus Page 132 of 132
5.0
1
0288.0
1.0
68.8
5.240,
0288.0
1.0
68.8
5.210
sss
s
s
ssG
s
s
s
ssG OLc
>>num=40*conv([1 2.5],[1 0.1]);
den=conv(conv(conv([1 8.68], [1 .0288]),[1 0]),[1
0.5]);
>> g=tf(num,den)
Transfer function:
40 s^2 + 104 s + 10
-------------------------------------
s^4 + 9.209 s^3 + 4.604 s^2 + 0.125 s
>> gcl=feedback(g,1)
Transfer function:
40 s^2 + 104 s + 10
-----------------------------------------
s^4 + 9.209 s^3 + 44.6 s^2 + 104.1 s + 10
>> damp(gcl)
Eigenvalue Damping Freq. (rad/s)
-1.00e-001 1.00e+000 1.00e-001
-4.34e+000 1.00e+000 4.34e+000
-2.39e+000+4.16e+000i 4.97e-001 4.80e+000
-2.39e+000-4.16e+000i 4.97e-001 4.80e+000
80
5.0
1
0288.0
1.0
68.8
5.240
5.0
1
288.0
1.0
68.8
5.240lim
0
sss
s
s
ssKv
s