Solution: 7.9.3.5
For the system of Figure 1 Let
Σ
G
C
G
P
+
R
C
Figure 1: Cascade Compensation with UnityFeedback
G
p
=
10
s(s+1)
: and G
c
=
K
c
(s+3)
s+18
:
K=7
The following MATLAB programwill generate the step response for K
c
=7.
g=zpk([] ,[0 -1],10)
gc = zpk([-3],[-18],7)
gcgp = series(g,gc)
h=1
tc = feedback(gcgp,h)
step(tc,3)
The output of the MATLAB program is
Zero/pole/gain:
10
-------
s (s+1)
Zero/pole/gain:
7 (s+3)
-------
(s+18)
Zero/pole/gain:
70 (s+3)
--------------
s (s+18) (s+1)
1
h=
1
Zero/pole/gain:
70 (s+3)
--------------------------------
(s+13.69) (s^2 + 5.306s + 15.34)
EDU>
Thus
T
c
(s)=
70(s+3)
(s+13:69)(s+2:65;j2:88)(s+2:65+ j2:88)
Then
c(t)=[1+Ae
;13:69t
+2jMje
;2:65t
cos(2:88t+ )]1(t):
Wenow nd the residues of the partial fraction expansion.
A =
(s+13:69)
70(s+3)
s(s+13:69)(s+2:65;j2:88)(s+2:65+ j2:88)
s=;13:69
= 0:42
M =
(s +2:65;j2:88)
70(s+3)
s(s+13:69)(s+2:65;j2:88)(s+2:65+ j2:88)
s=;2:65+j2:88
= ;0:70995;j0:34416
= 0:789
6
;2:69
Then
c(t)=[1+0:42e
;13:69t
+1:578e
;2:65t
cos(2:88t;2:69)]1(t):
K=7.8
The procedures are the same.
T
c
(s)=
78(s+3)
(s+13)(s+3;j3)(s+3+j3)
2
Then
c(t)=[1+Ae
;13t
+2jMje
;3t
cos(3t+ )]1(t):
A =
(s+13)
78(s+3)
s(s+13)(s+3;j3)(s+3+j3)
s=;13
= 0:55
M =
(s+3; j3)
78(s+3)
s(s+13)(s+3;j3)(s+3+j3)
s=;3+j3
= ;0:77523; j0:4174
= 0:8805
6
;2:648
Then
c(t)=[1+0:55e
;13t
+1:76e
;3t
cos(3t; 2:648)]1(t):
K=8.6
T
c
(s)=
86(s+3)
(s+12:22)(s+3:392; j3:1)(s+3:392+j3:1)
Then
c(t)=[1+Ae
;12:22t
+2jMje
;3:392t
cos(3:1t+)]1(t):
A =
(s+12:22)
86(s+3)
s(s+12:22)(s+3:392; j3:1)(s+3:392+ j3:1)
s=;12:22
= 0:742
M =
(s +3:392;j3:1)
86(s+3)
s(s+12:22)(s+3:392; j3:1)(s+3:392+ j3:1)
s=;3:392+j3:1
= ;0:8709; j0:5085
= 1:008
6
;2:613
Then
c(t)=[1+0:742e
;12:22t
+2:017e
;3:392t
cos(3:1t; 2:613)]1(t):
Fig. 2 shows the three responses. As can be seen they are very close.
AMATLAB program to nd all the residues and plot the responses
follows.
3
0.0
0.2
0.4
0.6
0.8
1.0
1.2
1.4
0 123
Time in Seconds
K = 8.6
K = 7.8
K = 7
Figure 2: Time Response for Nominal Gain 10%
g=zpk([] ,[0 -1],10)
gc = zpk([-3],[-18],7)
gcgp = series(g,gc)
h=1
tc = feedback(gcgp,h)
step(tc,3)
a=3
Kc = 7
Kp = 10
K=Kp*Kc
B=K*[1 a]
B=[00B]
c=[10]
d=[11]
e=[118]
A=conv(c,d)
A=conv(A,e)
roots(A)
A=A+B
roots(A)
A=conv(A,c)
roots(A)
roots(B)
[R,P,K] = residue(B,A)
4
realm = real(R(2))
imagm = imag(R(2))
m=realm + j*imagm
absm = abs(m)
absm2 =2*absm
angm = angle(m)
t0=1 + R(1) + absm2*cos(angm)
Kc = 7.8
Kp = 10
K=Kp*Kc
B=K*[1 a]
B=[00B]
c=[10]
d=[11]
e=[118]
A=conv(c,d)
A=conv(A,e)
roots(A)
A=A+B
roots(A)
A=conv(A,c)
roots(A)
roots(B)
[R,P,K] = residue(B,A)
realm = real(R(2))
imagm = imag(R(2))
m=realm + j*imagm
absm = abs(m)
absm2 =2*absm
angm = angle(m)
t0=1 + R(1) + absm2*cos(angm)
Kc = 8.6
Kp = 10
K=Kp*Kc
B=K*[1 a]
B=[00B]
c=[10]
d=[11]
e=[118]
A=conv(c,d)
A=conv(A,e)
roots(A)
A=A+B
roots(A)
A=conv(A,c)
5
roots(A)
roots(B)
[R,P,K] = residue(B,A)
realm = real(R(2))
imagm = imag(R(2))
m=realm + j*imagm
absm = abs(m)
absm2 =2*absm
angm = angle(m)
t0=1 + R(1) + absm2*cos(angm)
6