0% found this document useful (0 votes)
163 views23 pages

EE3331C Feedback Control System L6: Root Locus Analysis: Arthur TAY

The document outlines root locus analysis, which plots how closed-loop poles vary with changes in system gain. It reviews the key idea, which is that closed-loop poles start at open-loop poles and end at open-loop zeros or infinity as gain increases. Several examples are provided to demonstrate this concept, including simple first-order systems and motor position control systems. The document concludes with a summary and practice problems.
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
0% found this document useful (0 votes)
163 views23 pages

EE3331C Feedback Control System L6: Root Locus Analysis: Arthur TAY

The document outlines root locus analysis, which plots how closed-loop poles vary with changes in system gain. It reviews the key idea, which is that closed-loop poles start at open-loop poles and end at open-loop zeros or infinity as gain increases. Several examples are provided to demonstrate this concept, including simple first-order systems and motor position control systems. The document concludes with a summary and practice problems.
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 23

Outline Motivations Examples A final note Summary

EE3331C Feedback Control System


L6: Root Locus Analysis

Arthur TAY

ECE, NUS

6–1 / 23
Outline Motivations Examples A final note Summary

Outline
Motivations
Review
Key idea
Examples
Example 1: Simple first-order system
Example 2: Motor position control
Example 3: Root locus with respect to open-loop pole
Example 4: Unstable open-loop system
Example 5: Unstable system with extra zero
Example 6: 3rd order loop transfer function
Example 7: Two complex poles and one zero
Example 8: Conditionally stable system
Example 9: Unstable system
A final note
Summary
Summary
Practice Problems 6–2 / 23
Outline Motivations Examples A final note Summary
Review Key idea

Review:
I In our earlier heater example, we show a plot of how the closed-loop
poles varies by solving the closed-loop c.e. for different values of
gain, k.
Closed-loop poles locations as k varies
1.5
k=6
1
k = 3.96
k=1
k = 0.05
0.5
Imag
0

0.5

k=0
1

1.5
1.5 1 0.5 0 0.5
Real

Also known as the root locus plot.


6–3 / 23
Outline Motivations Examples A final note Summary
Review Key idea

I Key Idea: Consider the feedback system in Figure 6.1, the closed-loop
transfer function is
Y (s) G(s)K(s)
=
R(s) 1 + G(s)K(s)Gs (s)
with the closed-loop poles given by

1 + G(s)K(s)Gs (s) = 0

d
r e + y
+
K(s) G(s)
- +
n
Gs(s)

Figure 6.1: Basic closed-loop block diagram

6–4 / 23
Outline Motivations Examples A final note Summary
Review Key idea

I Assuming that we are interested in a parameter which we called K here


(note: does not need to be the controller gain k). The closed-loop c.e. can
be written as
1 + KL(s) = 0
b(s)
and if we let L(s) = , we then have
a(s)

b(s)
1+K = 0
a(s)
a(s) + Kb(s) = 0 (6.1)
a(s)
+ b(s) = 0
K
where a(s) and b(s) are of order n and m respectively and n ≥ m.
I As K → 0, the poles of the closed-loop system are a(s) = 0 or the
poles of L(s)
I As K → ∞, the poles of the closed-loop system are b(s) = 0 or the
zeros of L(s).
6–5 / 23
Outline Motivations Examples A final note Summary
Review Key idea

I Regardless of the value of K, the closed-loop system must always have n


poles:
I the root locus must have n branches, each branch starts at a pole of
L(s) and goes to a zero of L(s)
I if L(s) more poles than zeros (as is often the case), n > m, and we
say that L(s) has zeros at infinity; the number of zeros at infinity is
n − m, and is the number of branches of the root locus that go to
infinity (asymptotes).
I We have shown previously that by varying the gain, K, we can plot out the
locus of all possible roots (closed-loop poles). The resulting plot can then
aid us in selecting the best value of K to achieve the desired performance
specifications. We next demonstrate the idea via a few examples.

I In Matlab,
1. Describe L(s) in Matlab as a transfer function.
2. Plot the root locus with ‘rlocus(sys)’ where ‘sys’ is L(s).

6–6 / 23
Outline Motivations Examples A final note Summary
Example 1 Example 2 Example 3 Example 4 Example 5 Example 6 Example 7 Example 8 Example 9

Example 1: Root locus of first-order system


I Consider unity feedback control system with
kp
G(s) = , K(s) = k
sτ + 1
Closed-loop poles given by
kkp 1 + kkp
1 + G(s)K(s) = 1 + =0 ⇒ s=−
sτ + 1 τ
for kp and τ ≥ 0, the pole is always on the left-half plane.

1
Imaginary Axis

increasing k
0

− τ1
1
0
Real Axis

6–7 / 23
Outline Motivations Examples A final note Summary
Example 1 Example 2 Example 3 Example 4 Example 5 Example 6 Example 7 Example 8 Example 9

Example 2: Root locus of motor position control


I Consider the following normalized transfer function of a DC motor
Y (s) A
= G(s) =
U (s) s(s + c)

I we are interested how the closed-loop varies as the system gain A


varies
I referring to Figure 6.1, we have K(s) = 1, Gs (s) = 1 and we also
assumed c = 1
1
I We then have L(s) = s(s+1) , and K = A
K
1 + KL(s) = 1 + = 0
s(s + 1)
s2 + s + K = 0
solving, we have √
1 1 − 4K
r1 , r2 = − ±
2 2

6–8 / 23
Outline Motivations Examples A final note Summary
Example 1 Example 2 Example 3 Example 4 Example 5 Example 6 Example 7 Example 8 Example 9

I Two branches of the root locus: the locus starts from the open-loop poles
and ends at the open-loop zeros or infinity.
I As K increases, the poles moved towards each other along the real-axis,
meeting at s = −0.5 before breaking away from the real-axis towards
infinity with real-part of the poles at s = −0.5.
I By varying the gain K, we can have any closed-loop poles along the locus
to meet design specifications. e.g. plot shows ζ > 0.5.
1.5
u sin21 z 30!
1

0.5
Imag axis

20.5

21

21.5
22 21 0 1 2
Real axis

6–9 / 23
Outline Motivations Examples A final note Summary
Example 1 Example 2 Example 3 Example 4 Example 5 Example 6 Example 7 Example 8 Example 9

Example 3: Root locus of motor position control (cont)


I If we are now interested to find out how the closed-loop poles varies as c
varies, assuming A = 1, we have
1
1 + G(s) = 1 +
s(s + c)
The closed-loop poles is given by
s2 + cs + 1 = 0
Rearranging, we have
s
1+c =0
s2 + 1
The roots are given by

c c2 − 4
r1 , r2 = − ±
2 2

6–10 / 23
Outline Motivations Examples A final note Summary
Example 1 Example 2 Example 3 Example 4 Example 5 Example 6 Example 7 Example 8 Example 9

I At c = 0, the poles are at the open-loop pole locations.


I At c = 2, the two poles are both at s = −1.
I For c > 2, the two locus segments change direction and move in opposite
directions; one towards the open-loop zero and the other towards infinity.

1.5

0.5
Imag axis

20.5

21

21.5
22 21 0 1 2
Real axis

6–11 / 23
Outline Motivations Examples A final note Summary
Example 1 Example 2 Example 3 Example 4 Example 5 Example 6 Example 7 Example 8 Example 9

Example 4: Unstable open-loop system.


1 √
I Consider L(s) = −1± 5
2
, the open-loop poles are at s = 2 .
s +s−1

0.8
>> sys = tf(1, [1 1 −1]);
0.6 >> rlocus(sys)

0.4
Imaginary Axis

0.2

0.2

0.4 stable for K > 1

0.6

0.8
2 1.5 1 0.5 0 0.5 1
Real Axis

I At K = 1, CL poles crossed over to the RHP, giving the CL only marginal


stability.
6–12 / 23
Outline Motivations Examples A final note Summary
Example 1 Example 2 Example 3 Example 4 Example 5 Example 6 Example 7 Example 8 Example 9

Example 5: Unstable open-loop system with open-loop zero.


s+1
L(s) = 2
s +s−1
same open-loop poles as in previous example.

0.5
>> sys = tf([1 1], [1 1 −1]);
>> rlocus(sys)
0.3
Imaginary Axis

0.1

0.1

K = K0
0.3

0.5
4 3 2 1 0 1
Real Axis

I Closed-loop is only stable when K > K0 , the closed-loop c.e. is


s2 + s(K + 1) + K − 1 = 0
6–13 / 23
Outline Motivations Examples A final note Summary
Example 1 Example 2 Example 3 Example 4 Example 5 Example 6 Example 7 Example 8 Example 9

Example 6: loop transfer function with three poles, unity feedback


1 K
G(s) = 2 , K(s) =
s +s+1 s

2
>> sys = tf([1], [1 1 1 0]); At K = K0 , s = jω
1.5 >> rlocus(sys)

1
Imaginary Axis

0.5

0.5

1.5

2
2.5 2 1.5 1 0.5 0 0.5 1
Real Axis

I 3 branches of root locus, the locus starting from s = 0 is always stable, the
other 2 locus becomes unstable for K > K0 .
I Closed-loop c.e.: s3 + s2 + s + K = 0. How to find K0 ?
6–14 / 23
Outline Motivations Examples A final note Summary
Example 1 Example 2 Example 3 Example 4 Example 5 Example 6 Example 7 Example 8 Example 9

Example 7: loop transfer function with two complex poles, one zeros
s+1
L(s) =
s2 + 2s + 4

open-loop poles at s = −1 ± j 3, zero at s = −1.
2
>> sys = tf([1 1], [1 2 4]);
1.5 >> rlocus(sys)

0.5
Imaginary Axis

−0.5

−1

−1.5

−2
−5 −4 −3 −2 −1 0
Real Axis

I Closed-loop c.e.: s2 + (2 + K)s + 4 + K = 0. Closed-loop always stable.


6–15 / 23
Outline Motivations Examples A final note Summary
Example 1 Example 2 Example 3 Example 4 Example 5 Example 6 Example 7 Example 8 Example 9

Example 8: conditionally stable system


(s + 2)2
I Given that L(s) = ,
2
s(s + 1)(s + 8)

I using Matlab, sketch the root locus for K > 0

I determine the range of K for which the closed loop system is stable

I for what values of K > 0 do purely imaginary roots exist? what are
the values of these roots?
I would the use of the dominant roots approximation for an estimate of
settling time be justified in this case if K is large (K > 50)?

6–16 / 23
Outline Motivations Examples A final note Summary
Example 1 Example 2 Example 3 Example 4 Example 5 Example 6 Example 7 Example 8 Example 9

25

20
(s+2) 2
15 L(s) = s(s2 +1)(s+8)
10

−5

−10

−15

−20

−25
−8 −7 −6 −5 −4 −3 −2 −1 0 1

I Closed-loop c.e.: s4 + 8s3 + (K + 1)s2 + (4K + 8)s + 4K = 0


I When K = 50, the roots are s = −1.02 ± j5.63; −4.66; −1.31

6–17 / 23
Outline Motivations Examples A final note Summary
Example 1 Example 2 Example 3 Example 4 Example 5 Example 6 Example 7 Example 8 Example 9

1.8

1.6 Step response for K = 50


1.4

1.2

0.8

0.6

0.4

0.2

0
0 1 2 3 4 5 6
Time (sec)

6–18 / 23
Outline Motivations Examples A final note Summary
Example 1 Example 2 Example 3 Example 4 Example 5 Example 6 Example 7 Example 8 Example 9

Example 9: Unstable system


I Consider an OL plant, G(s), being controlled in a unity negative
feedback configuration by a controller, K(s),

s−1 K0 (s + 2)
G(s) = , K(s) =
s3 + s2 + s + 1 s
I Using Matlab, plot the root locus and determine if the CL system is
stable for any values of K0 > 0.

6–19 / 23
Outline Motivations Examples A final note Summary
Example 1 Example 2 Example 3 Example 4 Example 5 Example 6 Example 7 Example 8 Example 9

3
s−1
G(s) = s3 +s2 +s+1
2
K0 (s+2)
K(s) = s
1
Imaginary Axis

3
2 1.5 1 0.5 0 0.5 1
Real Axis

6–20 / 23
Outline Motivations Examples A final note Summary

A final note:
I What we have covered: the use of Matlab for root locus analysis,
very powerful tool for controller design.
I What we have not covered:
I Without Matlab, how to sketch the root locus?
I In control system design, we are interested to know how to modify
the dynamics in such a way as to achieve the desired performance
specifications
→ useful to know how to sketch root locus for controller design
→ help to check if computer program is right
I For the interested reader, please check Section 5.2 of Franklin’s book!

6–21 / 23
Outline Motivations Examples A final note Summary
Summary Practice Problems

Summary

I A root locus is a graph of the values of s that are solutions to the


equation 1 + KL(s) = 0 with respect to a real parameter K.
I A powerful method for controller design. Useful to be able to sketch
simple systems.
Review Questions
I How do you find the maximum k before a closed-loop system
becomes unstable?
I Discuss the stability of the following systems.

Reading: FPE: section 5.1


6–22 / 23
Outline Motivations Examples A final note Summary
Summary Practice Problems

Practice Problems
Figure shows the root locus for 1 + k s(s+2)(s12 +4s+5) .
1. What is the maximum k before the system becomes unstable?
2. At k = 6.5, the roots are s1,2 = −2.65 ± j1.23 and s3,4 = −0.35 ± j0.8.
What is the settling time and percentage overshoot of its step response?
Justify your answer.

6–23 / 23

You might also like