Control Systems Overview and Analysis
Control Systems Overview and Analysis
Reference books: “Control Systems Engineering,” N. S. Nise, John Wiley & Sons
“Modern Control Engineering,” K. Ogata, Prentice Hall
Course Contents:
* Mathematical Backgrounds
Laplace Transform (Chap. 1)
Linear Algebra
Complex Functions (Chap. 9)
* Control System Modeling
Input-Output Models (Chaps. 2, 3)
State-Space Models (Chap. 4)
* Analysis & Design Objectives
Transient Response (Chap. 5)
Stability (Chap. 6)
Steady-State Response (Chap. 7)
* Analysis & Design Methods
Root Locus Method (Chap. 8)
Frequency Response Method (Chap. 9)
1
1. Introduction
1.1. Introduction
* System, inputs and outputs
f(t) F(s)
(t), impulse 1
1, step 1
s
t, ramp 1
s2
e−at 1
s +a
sin(t) ω
s +ω 2
2
cos(t) s
s +ω 2
2
2
* Laplace transform theorems
Differentiation df (t )
sF(s)- f(0)
dt
Differentiation d 2 f (t )
s2 F (s)−sf (0)−ḟ (0)
dt 2
⃗ −1
* Partial fraction expansion ( F ( s ) L f (t ))
– Case 1: Real and distinct roots
2 K K K
F (s )= = 1+ 2 + 3
s( s+1 )(s +2) s s+1 s +2 ⇒ 2=K 1 ( s+1 )(s +2)+K 2 s(s +2)+K 3 s(s+1)
(i)
s=0⇒ 2=2 K 1 ⇒ K 1 =1
(iii)
s=−2⇒ 2=2 K 3 ⇒ K 3 =1
1 K1 K2 K3
F ( s )= = + + 2
s( s+1 )2 s ( s+ 1)2 s +1 ⇒1=K 1 ( s+1 ) + K 2 s+ K 3 s (s +1)
(i)
s=0⇒ 1=K 1 ⇒ K 1=1
2 2
(iii) 1=K 1 (s +1) + K 2 s + K 3 s( s+1 )⇒ s :0=K 1 + K 3 ⇒ K 3 =−K 1 =−1
– Case 3: Complex roots
3
5 K1 K 2s+K3
F ( s )= = + 2
s( s2 + 2 s+5 ) s s2 +2 s+ 5 ⇒5=K 1 ( s +2 s +5)+( K 2 s + K 3 ) s
2
(ii) 5=K 1 ( s + 2 s+5 )+( K 2 s+ K 3 )s
s2 : 0=K 1 + K 2 ⇒ K 2 =−1
s :0=2 K 1 + K 3 ⇒ K 3 =−2
d 2 y dy dy
2
+2 +5 y =5 x , y (0 )= (0 )=0
* Example: dt dt dt .
x(t) is a unit step input(x(t)=1), determine the output, y(t)
Lapalce transform
⇒L { d2 y
dt 2
dy
dt }
+2 +5 y=5 x , ⇒ ( s 2 +2 s +5 ) Y ( s )=5 X ( s )=
5
s
5 1 −s−2 1 s+1 1 2
⇒Y ( s )= = + = − −
s ( s2 + 2 s+5 ) s s 2 +2 s +5 s ( s+1 )2 + 22 2 ( s+ 1 )2 +22
1
⇒ y (t )=L−1 { Y ( s ) }=1−e−t cos2 t− e−t sin 2 t
Inverse Laplace transform 2
1.4. Problems
Problem 1.1
A 2nd-order system is described by the following differential equations and x(t)=1(a unit step input).
d2 y dy dy
2
+ 4 +5 y =15 x , y (0 )= (0 )=0
dt dt dt
Apply Lapalce transform method to determine the output, y(t).
Problem 1.2
A 2nd-order system is described by the following set of differential equations and u(t)=1 (a unit step
input).
{dx
dt
+x=2 u(t ) ¿ ¿ ¿ ¿
Apply Lapalce transform method to determine the output, y(t).
4
2. Input-Output Models
c (n−1)(0 )=¿⋅¿=ċ(0)=c(0)=0
m m−1
C ( s ) bm s +bm−1 s + ¿⋅¿⋅+ b0
⇒G ( s )= =
R( s ) an sn +an−1 s n−1 +¿⋅¿⋅+a0
resistor R
v (t )=Ri(t )
inductor di(t )
v (t )=L Ls
dt
1
Ls [ I 2 ( s )−I 1 (s )]+R 2 I 2 (s )+ I ( s )=0
Cs 2
5
1
V o (s )= I (s )
Cs 2
(2) by nodal analysis (current law):
V in ( s )−V 1 ( s ) V 1 (s ) V 1 ( s )−V o ( s )
− − =0
R1 Ls R2
V 1 ( s )−V o (s )
−CsV o ( s )=0
R2
(3) by mesh inspection:
[ Ls+R 1 - Ls ¿ ] ¿
¿¿
¿
1 |Ls+R1 V in (s) ¿|¿
V o (s)= I 2(s)= ¿¿¿¿
Cs ¿
Transfer function:
L
V o ( s) C Ls
T (s)= = =
V in (s) 1 2
( R1 +R 2 )LCs +( R1 R2 C +L)s+R1
( Ls+R 1 )( Ls+R 2 + )−L2 s2
Cs
V o (s ) Z 2( s )
=−
* Operational amplifier: V in ( s ) Z 1( s )
1
1
Cs+
V o (s ) Z 2( s ) R2 R2
=− =− =−
Example: V in ( s ) Z 1( s ) Ls+ R1 ( Ls + R1 )(R 2 Cs+1)
6
2.3. Translational Mechanical Systems
force F (s )
Z (s )= =
position X ( s )
* Impedance :
mass dx 2 (t )
f (t )=ma=m ms 2
dt 2
F (s )−k 2 [ X 2 (s )− X 1 ( s )]=m2 s2 X 2 ( s)
(2) by inspection
[ m1 s2+bs+k 1+k 2 - k 2 ¿ ] ¿ ¿¿
¿
7
|0 -k2 ¿|¿ X (s)=|m1 s2+bs+k 1+k2 0¿|¿ ¿¿¿¿
⇒X 1(s)= ¿¿¿¿ 2
¿ ¿ ,
X1 (s ) k2 X 2( s ) m 1 s 2 +bs +k 1 + k 2
⇒ = =
F (s) (m1 s2 +bs+ k 1 + k 2 )( m2 s 2 + k 2 )−k 22 , F( s ) ( m 1 s 2 + bs+k 1 +k 2 )( m2 s 2 +k 2 )−k
22
torque τ ( s )
Z (s )= =
angle θ( s )
* Impedance :
damper dθ(t )
τ (t )=Dω=D Ds
dt
inertia d 2 θ(t )
τ (t )=Jα=J Js 2
dt 2
* Example:
[ 1 +D1 s+k
J s 2
-k ¿ ] ¿ ¿¿
¿
8
⇒θ1(s)=| 0 -k ¿|¿ ¿¿¿¿ θ (s)=|J 1 s2+D1 s+k 0 ¿|¿ ¿¿¿¿
2
¿ , ¿
θ1 ( s ) k θ2 ( s ) J 1 s2 + D 1 s +k
⇒ = =
τ ( s ) ( J s2 + D s +k )( J s2 + D s +k )−k 2 τ ( s ) ( J 1 s 2 + D 1 s+ k )( J 2 s 2 + D 2 s+ k )−k 2
1 1 2 2 ,
( )
2
n1
σ=
n2
* Equivalent elements by transfer ratio:
* Example:
n1 n1
()θ1 ( s )
θ (s) n n2
T (s )= 2 = 2 =
τ 1( s ) τ 1 (s ) 2
J e s + De s+ K e
Transfer function: ,
( ) ( ) ( )
2 2 2
n1 n1 n1
J e =J 1 + J 2 D e =D 1 + D 2 K e =K 2
n2 n2 n2
where , ,
9
dθm
v b =K b
* Assumptions: dt , τ m=K t i a , La ≈0
Kt
θm ( s ) ( Ra J m )
=
[ ( )] ; ( ) ( )
2 2
Ea ( s ) 1 K K nm nm
s s+ D m+ t b J m=J a + J L D m=D a + D L
Jm Ra nL nL
where ,
10
2.7. Problems
Problem 2.1
Find the transfer function, T(s) = Eout (s)/Ein(s), for the following electrical system.
Problem 2.2
Find the transfer function, G(s) = Y(s)/X(s), for the following mechanical system.
Problem 2.3
Find the transfer function, G(s) = o(s)/in(s), for the following mechanical system.
11
3. Reduction of Subsystems
3.1. Step-by-Step
* Components of a block diagram: signals, summing junction, pickoff point
* Cascade form:
* Parallel form:
* Feedback form:
12
* Example:
13
3.2. Mason's Rule
C (s )
∑ Fi Δi
i
T (s )= =
* R (s ) Δ
Δ=1−∑ Li +∑ Li L j − ∑ Li L j Lk
i i,j (two non-touching loops) i, j, k (three non-touching loops) + ….
i is formed by eliminating the components of that touch Fi.
Fi is the ith forward-path gain
Li the ith loop gain
* Example:
L1 =−G1 G2 G4 H 1
L2 =−G2 H 2
L3 =−G3 G4 H 1
F 1=G1 G2 G 4
F 2=G3 G4
Δ=1− {−G1 G2 G4 H 1−G 2 H 2−G3 G 4 H 1 }+ {G2 H 2⋅G3 G4 H 1 }
=1+G1 G2 G 4 H 1 +G2 H 2 +G3 G4 H 1 +G2 G3 G 4 H 1 H 2
Δ 1=1
Δ 2 =1+ G 2 H 2
C (s ) F1 Δ1 + F 2 Δ 2 G1 G2 G4 +G3 G4 (1+G2 H 2 )
T (s )= = =
R (s ) Δ 1+ G1 G2 G4 H 1 +G2 H 2 +G3 G 4 H 1 + G2 G3 G4 H 1 H 2
14
3.3. Problems
Problem 3.1
For the system shown in the following figure, find the transfer function, T(s) = C(s)/R(s).
Problem 3.2
For the system shown in the following figure, find the transfer function, T(s) = C(s)/R(s), and
the differential equation.
15
4. State-Space Models
4.1. Introduction
* Classical, frequency-domain, transfer function, modern, time-domain, state-space model
*
{ ẋ= Ax+Bu¿¿¿¿
where x = state vector, u = input vector, y = output vector
* Example
x=¿ [ x1 ¿] [ x2 ¿] [ x3 ¿ ] ¿
¿¿
¿ , u=f , y=x 1
{ 1
⇒¿{ẋ1=x3 ¿ ẋ3= [−(k1+k2)x1+k2 x2−bx3]¿ {ẋ2=x4 ¿ ¿¿
m1
and y=x 1
16
⇒¿ [ ẋ1 ¿ ] [ ẋ2 ¿][ ẋ3 ¿] ¿
¿¿
¿
¿
4.3. Transfer Function State-Space Form
* Phase-variable form:
d4 y d3 y d2 y dy
4
+a 3 3 + a2 2 + a1 +a 0 y=u
dt dt dt dt
{{{
2 3
dy d y d y
{x1=y¿ x2=dt =ẋ1¿ x3=dt2 =ẋ2¿ x4=dt3 =ẋ3 ¿ ¿
⇒¿ [ ẋ1 ¿ ] [ ẋ2 ¿][ ẋ3 ¿] ¿
¿¿
¿
¿
* Modified phase-variable form:
17
⇒¿ [ ẋ1 ¿ ] [ ẋ2 ¿][ ẋ3 ¿] ¿
¿¿
¿
¿
* Decomposing the block diagram:
4.5. Problems
Problem 4.1
Represent the system as shown in the following figure in the state-space form. Note that the
system has one input (f), one output (v=dx/dt), and two state variables (x1, x2) = (x, v).
18
Problem 4.2
Represent the system as shown in the following figure in the state-space form. Note that the
system has two inputs (u1, u2), one output (y), and three state variables (x1, x2, x3).
5. Transient Response
zi = zeros, T ( zi )=0
* Step response:
K (s−z 1 )(s−z 2 )
G( s )=
(s− p1 )( s−p 2 )( s− p3 )
k k k k3
⇒C ( s )= 0 + 1 + 2 +
s s− p1 s− p2 s− p 3
p1 t p2 t p3 t
⇒ c ( t )=k 0 +k 1 e +k 2 e + k3 e (forced response + natural response)
* Classification:
2
(1) Overdamped systems: a −4 b>0 ( >1)
2
(2) Critically damped systems: a −4 b=0 ( = 1)
2
(3) Underdamped systems: a −4 b<0 (0 < < 1)
(4) Undamped systems: a=0 ( = 0)
(5) Unstable systems: a< 0 ( < 0)
p2 p t p1 p t
c (t )=1+ e 1+ e 2
For unit step input (r = 1), output p1 −p 2 p 2− p1
* Example:
20
10 10 −t 1 −10 t
G( s)= ⇒ c (t )=1− e + e
(s +1)( s+10 ) 9 9
1
G 1 ( s )= ⇒ c1 (t )=1−e−t
s +1
10
G 2 (s )=
s +10 ⇒ c 2 (t )=1−e−10 t
* Examples
X(s) 1
F(s) ms +bs+k
* Step response:
m
k
= 2 ;¿ ωn= ¿ ¿¿ {√ V o (s)
=
1
V in(s) LCs +RCs+1
2
;¿ {√
ω n =
1
LC
¿ ¿¿
21
K (static gain) = 1
( ς
)
−ςωn t
cosω n √ 1−ς t + sin ωn √1−ς t
2 2
⇒ c(t )=1−e
√1−ς 2
−σ t
e d
⇒ c (t )=1− cos ( ω d t−φ )
√1−ς 2
where
σ d =ςωn = exponential damping frequency
φ=tan−1
( √ 1−ς )
ς
2
= phase lag
* Poles,
p1 ,2 =−σ d ± jωd =−ςωn± jωn √ 1−ς 2
* Specifications:
π π π
t p= =
ωd ω √ 1−ς 2 ⇔ ωd =ωn √ 1−ς 2 =
Peak time, n tp
22
σ
d ln ( OS ) −ln ( OS )
−( σ d π /ω d ) −( ςπ / 1−ς
√ 2
) ⇔ ω =− ⇔ ς=
Overshoot, OS=e =e d π √ π 2+ln 2( OS )
4 4 4
t s= = ⇔ σ d =ςωn=
Settling time (2%),
σ d ςωn ts
1
G( s )= 2
* Example: ms +bs +k
Analysis process:
{ m , b , k } ⇒ {σ d , ωd } or { ς , ωn } ⇒ {OS , t p , t s }
Design process: {
OS , t p , t s } ⇒ {σ d , ω d } or {ς , ω n }⇒ { m, b , k }
23
5.6. System Responses with Additional Poles/Zeros
* Dominant poles:
{ { 5 25(s+20)
G1(s)= 2 ¿ G2(s)= 2
s +2s+5 (s +2s+5)( s+4)(s+25)
⇒G 2 (s )≈G1 ( s )
¿ ¿¿¿
{G 3 ( s )=
40( s +30 )
( s 2 +2 s +5)( s+ 6)( s+20 )
¿¿¿¿
24
5.7. Problems
Problem 5.1
For the system as shown in the following figure, compute the system’s peak time (tp), settling
time (ts), and overshoot (OS) for step input.
Problem 5.2
For the mechanical system as shown, the desired system specifications are: mass (m) = 20 Kg,
settling time (ts) = 0.5 sec, and overshoot (OS) = 5%. Determine the spring constant (k) and the
damper factor (b).
25
6. Stability
6.1. Introduction
* Unstable: there is any pole on the right-hand plane or repeated poles on the imaginary axis
s4 a4 a2 a0
s3 a3 a1 0
s2 a4 a4 0
b 2=a2 −a1 ( ) b 1=a0 −0( )
a3 a3
s1 a 0 0
c 1 =a1 −b 1 ( 3 )
b2
s0 b 0 0
d 0 =b 1−0( 2 )
c1
* Example:
s2 6 106
s1 1 0
11−106× =−6 .667
6
26
s0 6 0
106−0× =106
−6 . 667
s5 1 3 8
s4 2 6 4
s3 0(⇒ ) 6 0
s2 12 4 0
6−
ε
s1 4 ε2 0 0
6−
6 ε−12
s0 4 0 0
+
{ε=0 :+,+,+,−,+,+ }or{ε=0− :+,+,−,+,+,+ }⇒ 2 sign changes in the first column ⇒ 2
poles on the RHP ⇒ unstable
(closed-loop poles: 0.53 + j1.56, -1.03 + j 0.65, -1)
6.3. Problems
Problem 6.1
For the system as shown below, determine the range of the parameter K so that the system is
stable.
27
Problem 6.2
For the system as shown below, determine the range of the parameter so that the system is
stable.
28
7. Steady-State Errors
7.1. Introduction
* Unity Feedback Systems:
e ss =e (∞ )=lim e (t )
Steady-state error: t →∞
1 1
e ss = =
lim sG( s ) K v
Unit ramp input: s→0 , Kv = velocity constant
1 1
e ss = =
lim s G( s ) K a
2
29
)⋅¿⋅¿⋅¿
G(s)=K(s+z1 )(s+z2 n
¿¿
* System type: s (s+ p1 )(s+ p2 )⋅¿⋅¿⋅¿ (Type n system)
1
∞
unit ramp Kv 0
1
∞ ∞
unit parabola Ka
10( s+5 )
G( s )=
* Example: s (s 2 +2 s +10 )
3
e ss = =0. 6
ramp input, r = 3t
Kv
* Steady-state errors:
s→0
R D
e ss =lim [ sE( s ) ] =e ss +e ss =lim
s →0 [ sR ( s )
] [
−lim
sG2 ( s ) D( s )
1+G 1 ( s )G 2 ( s ) s→0 1+G 1 ( s ) G 2 ( s ) ]
10 G (s )=20
G 1 ( s )=
* Example: s +2 , 2 s ( s+5 )
30
ramp input (r = t) and step disturbance (d = -0.2)
e ss =0 . 05+0 . 04=0 . 09
7.4. Problems
Problem 7.1
Shown in the following figure is a system with a ramp input, r(t) = t, and a step disturbance, d(t)
= -0.3. Compute the steady-state error (ess).
Problem 7.2
Shown in the following figure is a system with a ramp input, r(t) = t, and a step disturbance, d(t)
= -0.3. Determine the range(範圍) of K so that ess < 0.05.
31
8. Root Locus Methods
8.1. Introduction
C (s ) kG( s )
T (s )= =
* Closed-loop transfer function: R (s ) 1+kG( s )H ( s )
* Root locus: the paths of the closed-loop poles (i.e., roots of Δ(s)=1+kG( s)H ( s)=0 )
σ a=
∑ pi − ∑ z i
n−m
(2 k +1 )π
θ a= ; k =0, 1, 2, ⋅¿⋅¿ ¿
n−m
* j-axis crossing: stability limit (by Routh table)
* Real-axis break-away and break-in points:
dD( s ) dN (s ) 1 D( s)
N ( s) − D (s )=0⇒ s=si k =− =−
ds ds ; where G(s i ) H (s i ) N ( s)
8.3. Examples
k ( s+6 )
kG (s ) H (s )=
* Example 1: ( s+ 2)(s + 4 )
n = 2, m = 1;
p1 ,2 =−2 ,−4 ; z 1=−6
32
σ a =0 θ a=π
; ;
s2 1 8+6k
s1 6+k 0
s0 8+6k 0
{ 6+k>0 ¿ ¿¿¿
dD( s ) dN (s )
N ( s) − D (s )=(s +6 )(2 s+ 6)−( s2 + 6 s+ 8)=s 2 +12 s +28=0
ds ds
k ( s+4 )
kG (s ) H (s )=
* Example 2: s( s+2 )(s +6 )(s +8 )
n = 4, m = 1;
p1 ,2 ,3 ,4 =0,−2,−6,−8 ; z 1=−4
−12 π 5π
σ a= =−4 θ a= , π ,
3 3 3
; ;
s4 1 76 4k
s3 16 96+k 0
s2 k 4k 0
70−
16
33
s1 −k 2 /16+6720 0 0
70−k /16
s0 4k 0 0
{ {
2
k −k /16+6720
70− >0 ¿ >0 ¿¿¿¿
16 70−k/16
s+2
⇒ 1+k =0 ⇒ ¿ { N (s)=s+2 ¿ ¿ ¿
s 2 +3 s+6
n = 2, m = 1;
p1 ,2 =−1 .5± j1 .94 ; z 1=−2
dD( s ) dN (s )
N ( s) − D (s )=(s +2)(2 s+3 )−( s 2 +3 s +6 )=s2 +4 s=0 ⇒ s 1 ,2 =0 ,−4
ds ds
34
(at s2 =−4 , k =5 )
35
8.5. Problems
Problem 8.1 For the system shown as follows, draw the root locus as K varies from 0 to infinity.
Problem 8.2 For the system shown as follows, draw the root locus as K varies from 0 to infinity.
Problem 8.3 For the system shown as follows, draw the root locus and determine the range(範圍) of K
to satisfy the following specifications: (1) ts < 1.0(s), (3) tp < 0.5(s), (4) OS < 10%.
36
9. Frequency Response Techniques
z=a+jb⇔¿ {a=Re(z)¿¿¿
* magnitude and phase angle:
M ( z)=|z|=√ a2 +b2
φ ( z)=∠ z=tan −1
b
a ()
* Multiplication and division:
z=z1 z2 ⇒¿ {|z|=|z1||z2|¿ ¿¿
z1
{
|z1|
z= ⇒ ¿ |z|= ¿ ¿¿
z2 |z 2|
2 s +5
G( s )=
* Example: ( s+ 2 ) (s +10) , s=0+ j 3
(1)
5+ j6
G( j 3 )=
( 2+ j 3 ) (10+ j3 )
5+ j6 (5+ j 6 ) ( 11− j36 ) 271− j 114
= = = =0 . 1912− j0 . 0805
11+ j36 (11+ j 36 ) ( 11− j36 ) 1417
(2)
M =|G ( j3 )|=
√5 2+62 =0 . 2075
√ 22+3 2× √102+ 32
φ =∠G( j3 )=tan
−1
( 65 )−(tan ( 32 )+ tan ( 103 ))=−22. 8
−1 −1 o
37
9.2. Frequency Response
* Sinusoidal frequency response: magnitude(M) + phase angle()
* Analytical expression:
{M o=M i M G ¿ ¿¿¿
where
G( jω )=G( s)|s→ jω ,
{ M G (ω)=|G( jω )|¿ ¿ ¿ ¿
9.3. Bode Plots
38
s+ a
G( s )=
* a
ω
n2
G ( s )= 2
s +2 ςωn s+ ω
* n2
39
aω
n2
G ( s )=
( )
2
( s +a ) s +2 ςωn s+ ω 2
* n
40
* Example:
30
G( s)H ( s)=
( s+1 )(s+2)( s+3 )
* Example:
K
KG(s) H (s)=
s( s+1)(s+2) ⇒ K < 6 for stability
41
9.5. Relative Stability
* Gain Margin and Phase Margin
9.6. Problems
Problem 9.1
As shown below, a mechanical system is actuated by a positional input, x(t)=sin(15t), and the
system’s parameters are: m=10kg, b=150 Ns/m, and k=1500N/m. Determine the magnitude(My) and
phase(y) of the output position at steady state, yss(t)= My sin(15t+y).
Problem 9.2
500( s+1 )
G( s )=
For the system, s 2 +10 s +100 , draw the Bode plot and Nyquist plot.
42