systems analysis and controlcontrol.asu.edu/classes/mmae443/443lecture23.pdf · the nyquist plot...
TRANSCRIPT
Systems Analysis and Control
Matthew M. PeetIllinois Institute of Technology
Lecture 23: Drawing The Nyquist Plot
Overview
In this Lecture, you will learn:
Review of Nyquist
Drawing the Nyquist Plot
• Using the Bode Plot
• What happens at r =∞• Poles on the imaginary axis
Phase Margin and Gain Margin
• Reading Stability Margins off the Nyquist Plot
M. Peet Lecture 23: Control Systems 2 / 30
ReviewSystems in Feedback
The closed loop iskG(s)
1 + kG(s)
We want to know when1 + kG(s) = 0
Question: Does 1k +G(s) have any zeros in the RHP?
G(s)k+
-
y(s)u(s)
M. Peet Lecture 23: Control Systems 3 / 30
ReviewThe Nyquist Contour
Definition 1.
The Nyquist Contour, CN is a contour which contains the imaginary axis andencloses the right half-place. The Nyquist contour is clockwise.
A Clockwise Curve
• Starts at the origin.
• Travels along imaginary axis till r =∞.
• At r =∞, loops around clockwise.
• Returns to the origin along imaginary axis.
We want to know if
1
k+G(s)
has any zeros in the Nyquist Contour
r = ∞
M. Peet Lecture 23: Control Systems 4 / 30
ReviewContour Mapping Principle
Key Point: For a point on the mapped contour, s∗ = G(s),
∠s∗ = ∠G(s)
• We measure θ, not phase.
To measure the 360◦ resets in ∠G(s)
• We count the number of +360◦ resets in θ!
• We count the number of times CG encircles the origin Clockwise.
The number of clockwise encirclementsof 0 is
• The #poles −#zeros in the RHPs
s*= G(s)
θ = < G(s)
M. Peet Lecture 23: Control Systems 5 / 30
The Nyquist ContourClosed Loop
The number of unstable closed-loop poles isN + P , where
• N is the number of clockwiseencirclements of −1k .
• P is the number of unstable open-looppoles.
If we get our data from Bode, typically P = 0
-1/k
How to Plot the Nyquist Curve?
M. Peet Lecture 23: Control Systems 6 / 30
Plotting the Nyquist DiagramExample
How are we to plot the Nyquist diagram for
G(s) =1
(τ1s+ 1)(τ2s+ 1)
• τ1 = 1• τ2 = 1
10
First lets take a look at the root locus.
Obviously stable for any k > 0.
−12 −10 −8 −6 −4 −2 0 2−6
−4
−2
0
2
4
6
Root Locus
Real Axis
Imag
inar
y A
xis
M. Peet Lecture 23: Control Systems 7 / 30
The Nyquist Plot
Bode Plot: Lets look at the Frequency Response.
The Bode plot can give us informationon |G| at different frequencies.
Point ω ∠G |G|A .1 0◦ 1B 1 −45◦ .7C 3 −90◦ .3D 10 −135◦ .07E 100 −175◦ .001
The last two columns give us points onthe Nyquist diagram.
-120
-100
-80
-60
-40
-20
0
20
Ma
gn
itu
de
(d
B)
10-2
10-1
100
101
102
103
-180
-135
-90
-45
0
Ph
ase
(d
eg
)
Bode Diagram
Frequency (rad/sec)
A BC
D
E
A
B
C
D
E
M. Peet Lecture 23: Control Systems 8 / 30
The Nyquist Plot
Plot the points from the Bode Diagram.
Point ω ∠G |G|A .1 0◦ 1B 1 −45◦ .7C 3 −90◦ .3D 10 −135◦ .07E 100 −175◦ .001
We get
the upper half of the Nyquist diagramfrom symmetry.
-0.2 0 0.2 0.4 0.6 0.8 1 1.2-0.8
-0.6
-0.4
-0.2
0
0.2
0.4
0.6
0.8
Nyquist Diagram
Real Axis
Ima
gin
ary
Axi
s
A
B
C
D
E
M. Peet Lecture 23: Control Systems 9 / 30
The Nyquist Plot
-0.2 0 0.2 0.4 0.6 0.8 1 1.2-0.8
-0.6
-0.4
-0.2
0
0.2
0.4
0.6
0.8
Nyquist Diagram
Real Axis
Ima
gin
ary
Axi
s
A
B
C
D
E
There are no encirclements of − 1k .
• Stable for all k > 0.• We already knew that from Root Locus.
M. Peet Lecture 23: Control Systems 10 / 30
The Nyquist PlotExample 2
G(s) =1
(s+ 1)3
First lets take a look at the root locus.
We expect instability for large k.
−2.5 −2 −1.5 −1 −0.5 0 0.5−2
−1.5
−1
−0.5
0
0.5
1
1.5
2
Root Locus
Real Axis
Imag
inar
y A
xis
M. Peet Lecture 23: Control Systems 11 / 30
The Nyquist Plot
Bode Plot: Lets look at the Frequency Response.
The Bode plot can give us informationon |G| at different frequencies.
Point ω ∠G |G|A .1 0◦ 1B .28 −45◦ .95C 1 −135◦ .35D 1.8 −180◦ .1E 10 −260◦ .001
-70
-60
-50
-40
-30
-20
-10
0
10
Ma
gn
itu
de
(d
B)
10-2
10-1
100
101
-270
-225
-180
-135
-90
-45
0
Ph
ase
(d
eg
)
Bode Diagram
Frequency (rad/sec)
A B
C
D
E
AB
C
D
E
M. Peet Lecture 23: Control Systems 12 / 30
The Nyquist Plot
Plot the points from the Bode Diagram.
Point ω ∠G |G|A .1 0◦ 1B .28 −45◦ .95C 1 −135◦ .35D 1.8 −180◦ .1E 10 −260◦ .001
Point D is especially important.
-0.4 -0.2 0 0.2 0.4 0.6 0.8 1 1.2-0.8
-0.6
-0.4
-0.2
0
0.2
0.4
0.6
0.8
Nyquist Diagram
Real Axis
Ima
gin
ary
Axi
s
A
B
C
D E
M. Peet Lecture 23: Control Systems 13 / 30
The Nyquist Plot
-0.4 -0.2 0 0.2 0.4 0.6 0.8 1 1.2-0.8
-0.6
-0.4
-0.2
0
0.2
0.4
0.6
0.8
Nyquist Diagram
Real Axis
Ima
gin
ary
Axi
s
A
B
C
D E
Point D: Two CW encirclements when − 1k < −.1 (N=2).
• Instability for − 1k < −.1
• Stable for k < 10.
• Could have used Routh Table.
M. Peet Lecture 23: Control Systems 14 / 30
The Nyquist Plot
Conclusion: We can use the Bode Plot to map theimaginary axis onto the Nyquist Diagram.
Question: What about the other part of the Nyquistcontour at r =∞?
r = ∞
Case 1: Strictly Proper.
lims→∞
|G(s)| = 0
What happens at ∞ doesn’t matter.
Case 2: Not Strictly Proper.
lims→∞
|G(s)| = c
Constant Magnitude at ∞.
Im(s)
Re(s)
< s-p1
< s-z = < s-p1 = < s-p
2 = < s-p
3
M. Peet Lecture 23: Control Systems 15 / 30
The Nyquist Plot
Case 2: Not Strictly Proper.
• Angle to all poles and zeros is the same.
• Degree of n(s) and d(s) the same.I Number of Poles and Zeros the same.
The total angle is
∠G(s) =
n∑i=1
∠(s− zi)−n∑
i=1
∠(s− pi)
= 0
The contour map at ∞ has
• Constant magnitude.
• Zero angle.
The infinite loop is mapped to a single point!
Either (0, 0) or (c, 0).
Im(s)
Re(s)
< s-p1
< s-z = < s-p1 = < s-p
2 = < s-p
3
0.6 0.8 1 1.2 1.4 1.6 1.8 2 2.2−0.8
−0.6
−0.4
−0.2
0
0.2
0.4
0.6
0.8
Nyquist Diagram
Real Axis
Imag
inar
y A
xis
M. Peet Lecture 23: Control Systems 16 / 30
The Nyquist Plot
Another Problem: Recall the non-inverted pendulumwith PD feedback.
G(s) =s+ 1
s2 +√
gl
Magnitude goes to ∞ at ω =√
gl .
Question How do we plot the NyquistDiagram?
−2 −1.5 −1 −0.5 0 0.5 1 1.5 2 2.5
−6
−4
−2
0
2
4
6
Nyquist Diagram
Real Axis
Imag
inar
y A
xis
Matlab can’t help us this time.
−40
−20
0
20
40
60
80
100
120
140
160
Mag
nitu
de (
dB)
10−2
10−1
100
101
102
−135
−90
−45
0
45
Pha
se (
deg)
Bode Diagram
Frequency (rad/sec)
M. Peet Lecture 23: Control Systems 17 / 30
The Nyquist Plot
Problem: The Nyquist Contour passes through a pole.
Because of the pole, the argument principle is invalid.
What to do? r = ∞
We Modify the Nyquist Contour.
• We detour around the poles.• Can detour to the right or left.
If we detour to the left, then the poles count as unstable open loop poles.
• P=2
Assume we detour to the right.
• P=0M. Peet Lecture 23: Control Systems 18 / 30
The Nyquist Plot
Look at the detours at small radius.
• Obviously, magnitude →∞
• Before the Detour, the phase from the pole is
−∠(s− p) = 90◦
• In the middle of the Detour, the phase from thepole is
−∠(s− p) = 0◦
• At the end of the Detour, the phase from the poleis
−∠(s− p) = −90◦
< (s - p) = 80o
< (s - p) = -5o
The total phase change through the detour is −180◦.
• Corresponds to a CW loop at large radius.
• If there are two or more poles, there is a -180 loop for each pole.
M. Peet Lecture 23: Control Systems 19 / 30
The Nyquist Plot
Look at the following example:
G(s) =s+ 2
s2
There are 2 poles at the origin.• At ω = 0,
I ∠G(0) = −180◦I |G(0)| =∞
• 2 poles means −360◦ loop at ω = 0M. Peet Lecture 23: Control Systems 20 / 30
The Nyquist Plot
Lets re-examine the pendulum problem with derivative feedback.
−10 −5 0 5 10
−10
−5
0
5
10
Nyquist Diagram
Real Axis
Ima
gin
ary
Axi
s
___ω=√(g/l)
Now we can figure out what goes on at ∞.
• There is a −180◦ loop at each ω =√
gl .
M. Peet Lecture 23: Control Systems 21 / 30
The Nyquist Plot
Conclusion: The loops connect in a non-obvious way!
−10 −5 0 5 10
−10
−5
0
5
10
Nyquist Diagram
Real Axis
Ima
gin
ary
Axi
s
___ω=√(g/l)
For 0 < −1k < 1, we have N = 1
−400 −200 0 200 400 600 800 1000 1200 1400−1500
−1000
−500
0
500
1000
1500
Nyquist Diagram
Real Axis
Imag
inar
y A
xis
M. Peet Lecture 23: Control Systems 22 / 30
Stability Margins
Recall the definitions of Gain Margin.
Definition 2.
The Gain Margin, Km = 1/|G(ıω)|when ∠G(ıω) = 180◦
Let Km is the maximum stable gain inclosed loop.
• KmG(s) is unstable in closed loop
• Sometimes expressed in dB
It is easy to find the maximum stablegain from the Nyquist Plot.
• Find the point −1Kmwhich
destabilizes
M. Peet Lecture 23: Control Systems 23 / 30
Stability MarginsExample
Recall
G(s) =1
(s+ 1)3
Stability: Stable for k < 10.
Km = 10 or 20dB.
-0.4 -0.2 0 0.2 0.4 0.6 0.8 1 1.2-0.8
-0.6
-0.4
-0.2
0
0.2
0.4
0.6
0.8
Nyquist Diagram
Real Axis
Ima
gin
ary
Axi
s
A
B
C
D E
M. Peet Lecture 23: Control Systems 24 / 30
Stability MarginsExample
Suspension System with integral feedback
There is a pole at the origin.
• CW loop at ∞.
−20 0 20 40 60 80 100
−50
−40
−30
−20
−10
0
10
20
30
40
50
Nyquist Diagram
Real Axis
Imag
inar
y A
xis
Conclusion: Stable for −1k > −9.5.
• Stable for k < .105
Km = .105 or − 19.5dB
-14 -12 -10 -8 -6 -4 -2 0 2-8
-6
-4
-2
0
2
4
6
8
Nyquist Diagram
Real Axis
Ima
gin
ary
Axi
s
}9.5
M. Peet Lecture 23: Control Systems 25 / 30
Stability Margins
Question: What is the effect of a phase change on the Nyquist Diagram.
• A shift in phase changes the angle of all points.
• A Rotation about the origin.
• Will we rotate into instability?
-0.2 0 0.2 0.4 0.6 0.8 1 1.2-0.8
-0.6
-0.4
-0.2
0
0.2
0.4
0.6
0.8
Nyquist Diagram
Real Axis
Imag
inary
Axi
s
-0.2 0 0.2 0.4 0.6 0.8 1 1.2-0.8
-0.6
-0.4
-0.2
0
0.2
0.4
0.6
0.8
Nyquist Diagram
Real Axis
Imag
inary
Axi
s
M. Peet Lecture 23: Control Systems 26 / 30
Stability Margins
Recall the definitions of Phase Margin.
Definition 3.
The Phase Margin, ΦM is the uniformphase change required to destabilize thesystem under unitary feedback.
−3 −2 −1 0 1 2 3−3
−2
−1
0
1
2
3
Nyquist Diagram
Real Axis
Ima
gin
ary
Axi
s ΦM
−3 −2 −1 0 1 2 3−3
−2
−1
0
1
2
3
Nyquist Diagram
Real Axis
Ima
gin
ary
Axi
s
M. Peet Lecture 23: Control Systems 27 / 30
Stability MarginsExample
The Suspension Problem
-3 -2.5 -2 -1.5 -1 -0.5 0 0.5 1-2
-1.5
-1
-0.5
0
0.5
1
1.5
2
Nyquist Diagram
Real Axis
Ima
gin
ary
Axi
s
ΦM
Looking at the intersection with the circle:• Phase Margin: ΦM
∼= 40◦
Gain Margin is infinite.M. Peet Lecture 23: Control Systems 28 / 30
Stability MarginsExample
The Inverted Pendulum Problem
-2 -1.5 -1 -0.5 0 0.5 1
-1.5
-1
-0.5
0
0.5
1
1.5
Nyquist Diagram
Real Axis
Ima
gin
ary
Axi
s
ΦM
Even though open-loop is unstable, we can still find the phase margin:
• Phase Margin: ΦM∼= 35◦
Gain Margin is technically undefined because open loop is unstable.
• There is a minimum gain, not a maximum.
M. Peet Lecture 23: Control Systems 29 / 30
Summary
What have we learned today?
Review of Nyquist
Drawing the Nyquist Plot
• Using the Bode Plot
• What happens at r =∞• Poles on the imaginary axis
Phase Margin and Gain Margin
• Reading Stability Margins off the Nyquist Plot
Next Lecture: Controller Design in the Frequency Domain
M. Peet Lecture 23: Control Systems 30 / 30