Chapter 2
Chapter 2
Automatic Control
ECE 306
Nguyen Thi My Thanh
Room 320.B11 (office hours)
Phone: 0907599993
Email: [email protected]
1
Chapter 2 The stability of linear systems
2.1 Definition
2.2 Algebraic stability criteria
• Necessary condition
• Routh’s stability criterion
2.3 Frequency stability criteria
• Frequency response
• Nyquist criteria
• Gain margin and phase margin
2.4 Root locus
2
2.1 Definition
Bounded-Input Bounded-Output (BIBO) stability criterion
➢A stable system produces a bounded output for a given bounded input.
3
2.1 Definition
Poles and Zeros
U(s) Y(s)
G(s)
➢ Transfer function
5
2.1 Definition
Example
1 1 1
𝑎) 𝐺 𝑠 = 𝑏) 𝐺 𝑠 = 𝑐) 𝐺 𝑠 = 2
𝑠+1 𝑠−1 𝑠 + 2𝑠 + 5
Pole: -1 Pole: 1 Pole: −1 ∓ 2𝑗
Necessary condition
➢All the system’s characteristic equation coefficients must be positive and non-zero.
(If all coefficients are negative, they can be made positive by multiplying both sides of the equation by –1).
Examples:
1
a. 𝐺 𝑠 = 𝑠3 +3𝑠2 −3𝑠−1 => Unstable
1
b. 𝐺 𝑠 = 𝑠 4 +4𝑠 3 +2𝑠+5
=> Unstable
1
c. 𝐺 𝑠 = 𝑠 4 +4𝑠 3 +3𝑠 2 +2𝑠+5
=> Move to sufficient condition
7
2.2 Algebraic stability criteria
Routh’s stability criterion
➢The necessary and sufficient condition of stability is that all roots of equation
✓All terms in the first column of the Routh’s array have positive signs.
➢The number of roots of A(s)=0 with positive real parts is equal to the number of
8
2.2 Algebraic stability criteria
Routh’s stability criterion
Routh array
9
2.2 Algebraic stability criteria
Routh’s stability criterion
Example Consider the following polynomial:
𝑠3 1 4 4∗2−1∗2
=3
𝑠2 2 2 2
0∗2−1∗0
𝑠1 3 0 2
=0
𝑠0 2 3∗2−2∗0
=2
3
There is no change in sign of the first column ⇒ All 3 roots of A(s) have negative real parts.
All terms in the first column of the Routh array have positive signs => The system is stable.
10
2.2 Algebraic stability criteria
Routh’s stability criterion
Example Consider the following polynomial:
Example:
=> Unstable
12
2.2 Algebraic stability criteria
Routh’s stability criterion
Special case 2:
➢ First-column term in any row is zero =>a very small positive number replaces the zero termr 𝜀
Example:
In both cases: ε > 0 and ε < 0, there are 2 changes in sign of the first column
⇒ A(s) has two roots in the right-hand plane.
⇒ The system is unstable. 13
2.2 Algebraic stability criteria
Application of Routh’s Stability Criterion to Control-System Analysis
14
2.2 Algebraic stability criteria
Exercises
Exercise 1: Determine the range of K for stability of a unity feedback control system whose open-
Exercise 2: Consider the following characteristic equation. Using the Routh stability criterion,
15
2.3 Frequency stability criteria
Frequency response
➢Frequency response is the response of a linear system at a steady state when the input
is a sinusoidal signal.
➢The output is also a sinusoidal signal with the same frequency as the input, but a
difference in the amplitude and phase.
➢Frequency response is the ratio between the steady-state output and the input.
16
2.3 Frequency stability criteria
Graphical Forms
▪ Nyquist plot: consists the magnitude of G(jω) versus the phase angle of G(jω)
on polar coordinates as ω is varied from zero to infinity.
▪ Bode diagram: consists of two graphs: a plot of the logarithm of the magnitude;
and a plot of the phase angle. Both are plotted on a logarithmic scale.
17
2.3 Frequency stability criteria
Nyquist plot
➢The Nyquist plot is the locus of vectors as 𝐺(𝑗𝜔) ∠G(jω) is varied from zero to infinity.
18
2.3 Frequency stability criteria
Nyquist plot Im
1
Example Draw a Nyquist plot of 𝐺 𝑠 =
𝑠+1
1 1
𝐺 𝑗𝜔 = = ∠ − arctg(ω) 𝜔 = +∞ 1/2 𝜔=0
𝑗𝜔 + 1 1+𝜔 2
Re
1
1 1 − 𝑗𝜔 1 −𝜔
𝐺 𝑗𝜔 = . = 2
+j
𝑗𝜔 + 1 1 − 𝑗𝜔 1 + 𝜔 1 + 𝜔2
1
-2
𝜔=1
𝜔 = 0: 𝐺 𝑗0 = 1
1 1 1
𝜔 = 1: 𝐺 𝑗 = − 𝑗 = − ∠ − 45𝑜
2 2 2
𝜔 → +∞: 𝐺 𝑗∞ = 0∠ −90°
19
2.3 Frequency stability criteria
Nyquist plot Im
1
Example Draw a Nyquist plot of 𝐺 𝑠 =
(𝑠 + 1)2
1 1
𝐺 𝑗𝜔 = = ∠ − 2arctg(ω) 𝜔=0
(𝑗𝜔 + 1)2 1 + 𝜔 2 𝜔 = +∞
1
Re
1 (1 − 𝑗𝜔)2 1 − 𝜔2 −2𝜔
𝐺 𝑗𝜔 = . = +j
(𝑗𝜔 + 1)2 (1 − 𝑗𝜔)2 (1 + 𝜔 2 )2 (1 + 𝜔 2 )2
𝜔=1
1
2
𝜔 = 0: 𝐺 𝑗0 = 1
𝜔 → +∞: 𝐺 𝑗∞ = 0∠ −180°
Crossover point: 𝜔 = 1 (real part = 0)
1
=> 𝐺 𝑗1 = −𝑗
2
20
2.3 Frequency stability criteria
Nyquist plot Im
1 𝜔 → 0−
Example Draw a Nyquist plot of 𝐺 𝑠 =
𝑠(𝑠 + 1)2
1 1 ° − 2arctg(ω)
𝐺 𝑗𝜔 = = ∠ − 90
𝑗𝜔(𝑗𝜔 + 1)2 𝜔(1 + 𝜔)2
1 (1 − 𝑗𝜔)2 1 − 𝜔2 − 2𝑗𝜔 −2
𝐺 𝑗𝜔 = 2
. 2
= 𝜔 = +∞
Re
𝑗𝜔(𝑗𝜔 + 1) (1 − 𝑗𝜔) 𝑗𝜔(1 + 𝜔 2 )2 1 𝜔 = −∞
−
2
−2 −1 + 𝜔2
𝐺 𝑗𝜔 = +j
(1 + 𝜔 2 )2 𝜔(1 + 𝜔 2 )2
𝜔 → 0: 𝐺 𝑗0 = −2 − 𝑗
𝜔 → +∞: 𝐺 𝑗∞ = 0∠ −270° 𝜔 → 0+
1
Crossover point: 𝜔 = 1 => 𝐺 𝑗1 = −
2
21
2.3 Frequency stability criteria
Nyquist plot
Exercise: Draw the Nyquist plot of the following transfer functions
a) b)
c) d)
22
2.3 Frequency stability criteria
Nyquist criteria
R(s) U(s) Y(s)
+
-
G(s)
➢For a closed-loop system to be stable, the G(j𝜔) plot must encircle the (-1, j0)
point m times the clockwise direction when 𝜔 changes from −∞ 𝑡𝑜 + ∞. (m is the
number of poles with the positive real part of G(s)).
➢If G(s) is stable then H(s) is stable <=> G(j 𝜔) doesn’t encircle the point -1 as
𝜔 = −∞ to +∞.
23
2.3 Frequency stability criteria
Nyquist criteria Im
𝐺𝑜
Example 𝐺 𝑠 = , 𝐺𝑜 > 0
𝑠(𝑠+1)2
𝐺𝑜 𝐺𝑜
𝐺 𝑗𝜔 = = ∠ −90° − 2𝑎𝑟𝑡𝑔(𝜔)
𝑗𝜔(𝑗𝜔+1)2 𝜔(1+𝜔)2 -1 𝜔 = +∞ Re
𝐺𝑜 (1−𝑗𝜔)2 −𝐺𝑜
𝐺 𝑗𝜔 = . 2
𝑗𝜔(𝑗𝜔+1)2 (1−𝑗𝜔)2
(1−𝜔2 −2𝑗𝜔) −2𝐺𝑜 (−1+𝜔2 )𝐺𝑜
= 𝐺𝑜 = + j
𝑗𝜔(1+𝜔2 )2 (1+𝜔2 )2 𝜔(1+𝜔)2
𝜔 = 0+ : 𝐺 𝑗𝜔 → (−2 − 𝑗∞)𝐺𝑜
𝜔 → +∞: 𝐺 𝑗𝜔 →0 ∠ −2700
Pole of G(s): -1, 0
𝜔 → 0+
H(s) is stable if G(s) does not encircle the point (-1,j0)
−𝐺𝑜
−1 < => 𝐺𝑜 < 2 24
2
2.3 Frequency stability criteria
Nyquist criteria
Compare to the Routh’s crierion
Routh’s Stability Criterion
𝐺(𝑠) 𝐺𝑜 𝐺𝑜
𝐻 𝑠 = = =
1+𝐺(𝑠) 𝑠(𝑠+1)2 +𝐺𝑜 𝑠 3 +2𝑠 2 +𝑠+𝐺𝑜
𝑠3 1 1
𝑠2 2 𝐺𝑜
2 − 𝐺𝑜
𝑠1 0
2
𝑠0 𝐺𝑜
2−𝐺𝑜
>0
H(s) is stable =>ቐ 2 ⇒ 0 < 𝐺𝑜 < 2
𝐺𝑜 > 0
25
2.3 Frequency stability criteria
Bode diagram
(in degrees)
➢ The curves use the log scale for frequency and the linear scale for either
magnitude (but in decibels) or phase angle (in degrees).
26
2.3 Frequency stability criteria
Bode diagram
Gian K: 𝐺 𝑗𝜔 = K
L ω = 20𝑙𝑜𝑔𝐾
⇒ቐ
𝜑 𝜔 = 0
27
2.3 Frequency stability criteria
Bode diagram
Integral factors: Derivative factors:
1 1
𝐺 𝑗𝜔 = = −j 𝐺 𝑗𝜔 = 𝑗𝜔
𝑗𝜔 𝜔
1
L ω = 20𝑙𝑜𝑔 = −20𝑙𝑜𝑔𝜔
⇒൞ 𝑗𝜔 L ω = 20𝑙𝑜𝑔 𝑗𝜔 = 20𝑙𝑜𝑔𝜔
𝜑 𝜔 = −90° ⇒൝
𝜑 𝜔 = 90°
28
2.3 Frequency stability criteria
Bode diagram
1
➢First-order factors : G jω =
Tjω+1
1
𝐺 𝑗𝜔 = ⇒ L ω = −20𝑙𝑜𝑔 1 + 𝑇 2 𝜔 2
1+𝑇 2 𝜔2
1
▪ 𝜔< ⇒ L ω = −20𝑙𝑜𝑔1 = 0
𝑇
1
▪ 𝜔> ⇒ L ω = −20𝑙𝑜𝑔𝜔𝑇
𝑇
1
▪ Break (Corner) frequency: 𝜔 =
𝑇
29
2.3 Frequency stability criteria
Bode diagram
➢Quadratic factors:
30
2.3 Frequency stability criteria
Bode diagram
𝐺 𝑗𝜔 = 𝑐𝑜𝑠𝜔𝑇 − 𝑗𝑠𝑖𝑛𝜔𝑇 = 1 ⇒ L ω = 0
𝜑 𝜔 = −ω𝑇
31
2.3 Frequency stability criteria
Procedure for Plotting Bode Diagrams
➢Step 1 Rewrite the sinusoidal transfer function G(jω) as a product of basic factors.
➢Step 2 Identify the corner frequencies associated with these basic factors
1
𝜔𝑖 =
𝑇𝑖
➢Step 3 Draw the asymptotic log-magnitude curves with proper slopes between the corner
frequencies. The phase-angle curve of G(jω) can be drawn by adding the phase-angle
curves of individual factors.
32
2.3 Frequency stability criteria
Bode diagram
100(0.1𝑠 + 1)
Exercise 1: Plot the Bode diagram of the following transfer funtion 𝐺 𝑠 =
𝑠(0.01𝑠 + 1)
Solution:
𝐿 𝜔
100(0.1𝑗𝜔 + 1) 60 -20dB
𝐺 𝑗𝜔 = 40
𝑗𝜔(0.01𝑗𝜔 + 1) 20
0dB
-20dB
100 𝜑 𝜔
, 𝜔 < 10
𝑗𝜔
10, 10 < 𝜔 < 100
𝐺 𝑗𝜔 ≈ 1 𝜔
10 0.1 10 100 1000
, 100 < 𝜔 -45
𝑗𝜔
-90
33
2.3 Frequency stability criteria
Bode diagram
10(0.01𝑠 + 1)
Example2 : Plot the block diagram for the transfer function 𝐺 𝑠 =
𝑠(𝑠 + 1)(0.1𝑠 + 1)2
Solution: 𝐿 𝜔
40 -20dB
20
10(0.01𝑗𝜔 + 1) 0.1 1
-40dB
10 100 1000 𝜔
𝐺 𝑗𝜔 = -20
𝑗𝜔(𝑗𝜔 + 1)(0.1𝑗𝜔 + 1)2 -80dB
-100
Corner frequencies: 1, 10 (double), 100 -60dB
-160
10
, 𝜔<1 𝜑 𝜔
𝑗𝜔
10
, 1 < 𝜔 < 10 0.1 1 10 100 1000 𝜔
(𝑗𝜔)2
𝐺 𝑗𝜔 ≈ -90
10
, 10 < 𝜔 < 100 -180
(𝑗𝜔)4
10 -270
, 100 < 𝜔
(𝑗𝜔)3 -360
34
2.3 Frequency stability criteria
Margin of stability
35
2.3 Frequency stability criteria
Gain margin and Phase margin
37
2.3 Frequency stability criteria
Gain margin
➢Positive gain margin (dB) Stable. ➢Negative gain margin (dB) Unstable
38
2.3 Frequency stability criteria
Phase margin
➢Phase margin is the difference between the actual phase angle 𝜑 and –180°.
Փ = 180° + 𝜑
➢For a minimum phase system to be stable, the phase margin must be positive.
39
2.3 Frequency stability criteria
Gain margin and Phase margin 1 Im
𝐾𝑔
𝐺𝑜
Example 𝐺 𝑠 = , 𝐺𝑜 > 0
𝑠(𝑠+1)2
−𝐺𝑜
2 Re
Gain margin:
1 2 −1 Փ
𝐾𝑔 = =
𝐺(𝑗𝜔−𝜋 ) 𝐺𝑜
Phase margin:
Փ = 180° + 𝜑 = 180° − 90° − 2𝑎𝑟𝑐𝑡𝑔(𝜔)
Փ = 90° − 2𝑎𝑟𝑐𝑡𝑔(𝜔)
𝜔 → 0+
40
2.3 Frequency stability criteria
𝐼𝑚
Gain margin and Phase margin
1
𝐺𝑜 𝐾𝑔
Example 𝐺 𝑠 = , 𝐺𝑜 > 0
(𝑠+1)3
𝐺𝑜
𝐺 𝑗𝜔 =
(𝑗𝜔 + 1)3
𝜔 = 0 ⇒ 𝐺 𝑗𝜔 = 𝐺𝑜 𝜔= 3 𝜔 = +∞ 𝜔=0
𝑅𝑒
𝐺𝑜 𝐺𝑜 𝐺𝑜 𝐺𝑜
𝜔 = 0 ⇒ 𝐺 𝑗𝜔 → = ∠ −270° −
8
𝑗𝜔3 𝜔3
Gian margin
1 𝐺𝑜 8 𝜔= 3
= ⇒ 𝐾𝑔 = 𝜔 = +∞ 𝜔=0
𝑅𝑒
𝐾𝑔 8 𝐺𝑜 𝐺𝑜 𝐺𝑜
−
8
Phase margin
Փ = 180° + 𝜑 = 180° − 3𝑎𝑟𝑐𝑡𝑔(𝜔)
To be stable Փ= 60°
⇒ 60° = 180° − 3𝑎𝑟𝑐𝑡𝑔 𝜔 ⇒ 𝜔 = 𝑡𝑔40°
42
2.3 Frequency stability criteria
Gain margin and Phase margin
Exercises
43
2.4 Root locus
➢The root locus is the path of the roots of the characteristic equation traced out in the s-
plane as a system parameter is changed.
➢The root loci for the system are the loci of the closed-loop poles as the gain K is varied
from zero to infinity
▪ Closed-loop system
44
2.4 Root locus
The root locus procedure
2. K=0: P(s)=A(s)
The root Loci end at the zeros of G(s) (The roots of B(s))
4. The root loci are symmetrical with respect to the real axis of the s-plane.
45
2.4 Root locus
The root locus procedure
5. The angles of the asymptotes
6. The intersection of the 2 𝑛 − 𝑚 asymptotes lies on the real axis of the s-plane, at
σ 𝑝𝑜𝑙𝑒𝑠 𝑜𝑓 𝐺 𝑠 − σ 𝑧𝑒𝑟𝑜𝑠 𝑜𝑓 𝐺 𝑠
𝜎=
𝑛−𝑚
The intersection of the asymptotes represents the center of gravity of the root loci, and
is always a real number.
46
2.4 Root locus
Example:
Characteristic equation: