F I V E

Reduction of Multiple
Subsystems
SOLUTIONS TO CASE STUDIES CHALLENGES
Antenna Control: Designing a Closed-Loop Response
a. Drawing the block diagram of the system:

Motor,
Power load and
Pots Pre amp amp gears

ui + 150 0.16 uo
10 K
 s+150 s (s+1.32)
-

Thus, T(s) =

b. Drawing the signal flow-diagram for each subsystem and then interconnecting them yields:

10 1 1 1
 K 150 s x3 0.8 s x2 s x1 0.2
ui uo
pre gears
pot
amp -150 -1.32

power motor
amp and
10 load
-


pot
5-2 Chapter 5: Reduction of Multiple Subsystems

x1 = x2
x2 = - 1.32x2 + 0.8x3
10
( qi −0 .2x 1 ))
x3 = -150x3 +150K( p = -95.49Kx1 - 150x3 + 477.46Ki
o = 0.2x1

In vector-matrix notation,

0 1 0 0

x = 0 -1.32 0.8 x + 0 
i
-95.49K 0 -150 477.46K

o = 0.2 0 0 x

c.
T 1= (10p )( K )( 150 )( 1s )(0 . 8)( 1s )( 1s )(0 . 2)=76s.39
3

G L 1=
−150
s
; GL 2 =
−1. 32
s
; G L 3 =( K )(150)
1
s ()
(0 . 8)
1
s ( )( 1s ) (0. 2)( −10p )= −76s.39 K
3

Nontouching loops:

GL1GL2 =

 = 1 - [GL1 + GL2 + GL3] + [GL1GL2] = 1 + + + +

1 = 1

T(s) = =
d. The equivalent forward path transfer function is G(s) = .
Therefore,

T(s) =

The poles are located at -0.66 ± j1.454. n = = 1.597 rad/s; 2n = 1.32, therefore,  = 0.413.

% OS=e−zp / √1−z x 100=24 % ; Ts = = = 6.06 seconds; Tp = = = 2.16 seconds; Using Figure

2

4.16, the normalized rise time is 1.486. Dividing by the natural frequency, Tr = = 0.93 seconds.
e.
 Solutions to Case Studies Challenges 5-3

f. Since G(s) = , T(s) = . Also,  = = 0.517 for 15% overshoot; n = ; and 2n = 1.32. Therefore, n

= = = 1.277 = .

Solving for K, K=3.2.

UFSS Vehicle: Pitch-Angle Control Representation

a. Use the observer canonical form for the vehicle dynamics so that the output yaw rate is a state

variable.
1

1 1 1 1
1 -1 2 s -0.125 0.437 s 1 s s 1
u y
x4 x3 x2 x1

-2 -1.483

-0.24897

-1

b. Using the signal flow graph to write the state equations:

In vector-matrix form:
5-4 Chapter 5: Reduction of Multiple Subsystems

c.
Program:
numg1=-0.25*[1 0.437];
deng1=poly([-2 -1.29 -0.193 0]);
'G(s)'
G=tf(numg1,deng1)
numh1=[-1 0];
denh1=[0 1];
'H(s)'
H=tf(numh1,denh1)
'Ge(s)'
Ge=feedback(G,H)
'T(s)'
T=feedback(-1*Ge,1)
[numt,dent]=tfdata(T,'V');
[Acc,Bcc,Ccc,Dcc]=tf2ss(numt,dent)

Computer response:
ans =

G(s)

Transfer function:
-0.25 s - 0.1093
--------------------------------------
s^4 + 3.483 s^3 + 3.215 s^2 + 0.4979 s

ans =

H(s)

Transfer function:
-s

ans =

Ge(s)

Transfer function:
-0.25 s - 0.1093
--------------------------------------
s^4 + 3.483 s^3 + 3.465 s^2 + 0.6072 s

ans =

T(s)

Transfer function:
0.25 s + 0.1093
-----------------------------------------------
s^4 + 3.483 s^3 + 3.465 s^2 + 0.8572 s + 0.1093
 Solutions to Case Studies Challenges 5-5

Acc =

-3.4830 -3.4650 -0.8572 -0.1093

1.0000 0 0 0
0 1.0000 0 0
0 0 1.0000 0

Bcc =

1
0
0
0

Ccc =

0 0 0.2500 0.1093

Dcc =

ANSWERS TO REVIEW QUESTIONS

1. Signals, systems, summing junctions, pickoff points
2. Cascade, parallel, feedback
3. Product of individual transfer functions, sum of individual transfer functions, forward gain divided by
one plus the product of the forward gain times the feedback gain
4. Equivalent forms for moving blocks across summing junctions and pickoff points
5. As K is varied from 0 to ∞, the system goes from overdamped to critically damped to underdamped.
When the system is underdamped, the settling time remains constant.
6. Since the real part remains constant and the imaginary part increases, the radial distance from the origin
is increasing. Thus the angle  is increasing. Since = cos  the damping ratio is decreasing.
7. Nodes (signals), branches (systems)
8. Signals flowing into a node are added together. Signals flowing out of a node are the sum of signals
flowing into a node.
9. One
10. Phase-variable form, cascaded form, parallel form, Jordan canonical form, observer canonical form
11. The Jordan canonical form and the parallel form result from a partial fraction expansion.
12. Parallel form
13. The system poles, or eigenvalues
14. The system poles including all repetitions of the repeated roots
15. Solution of the state variables are achieved through decoupled equations. i.e. the equations are solvable
individually and not simultaneously.
16. State variables can be identified with physical parameters; ease of solution of some representations
5-6 Chapter 5: Reduction of Multiple Subsystems

17. Systems with zeros

18. State-vector transformations are the transformation of the state vector from one basis system to another.
i.e. the same vector represented in another basis.
19. A vector which under a matrix transformation is collinear with the original. In other words, the length
of the vector has changed, but not its angle.
20. An eigenvalue is that multiple of the original vector that is the transformed vector.
21. Resulting system matrix is diagonal.

SOLUTIONS TO PROBLEMS
1.
a. Combine the inner feedback and the parallel pair.

Multiply the blocks in the forward path and apply the feedback formula to get,

T(s) = .
b.
Program:
'G1(s)'
G1=tf(1,[1 0 0])
'G2(s)'
G2=tf(50,[1 1])
'G3(s)'
G3=tf(2,[1 0])
'G4(s)'
G4=tf([1 0],1)
'G5(s)'
G5=2
'Ge1(s)=G2(s)/(1+G2(s)G3(s))'
Ge1=G2/(1+G2*G3)
'Ge2(s)=G4(s)-G5(s)'
Ge2=G4-G5
'Ge3(s)=G1(s)Ge1(s)Ge2(s)'
Ge3=G1*Ge1*Ge2
'T(s)=Ge3(s)/(1+Ge3(s))'
T=feedback(Ge3,1);
T=minreal(T)

Computer response:
ans =
 Solutions to Case Studies Challenges 5-7

G1(s)

Transfer function:
1
---
s^2

ans =

G2(s)

Transfer function:
50
-----
s+1

ans =

G3(s)

Transfer function:
2
-
s

ans =

G4(s)

Transfer function:
s

ans =

G5(s)

G5 =

ans =

Ge1(s)=G2(s)/(1+G2(s)G3(s))

Transfer function:
50 s^2 + 50 s
-------------------------
s^3 + 2 s^2 + 101 s + 100

ans =

Ge2(s)=G4(s)-G5(s)

Transfer function:
s-2

ans =

Ge3(s)=G1(s)Ge1(s)Ge2(s)
5-8 Chapter 5: Reduction of Multiple Subsystems

Transfer function:
50 s^3 - 50 s^2 - 100 s
-------------------------------
s^5 + 2 s^4 + 101 s^3 + 100 s^2

ans =

T(s)=Ge3(s)/(1+Ge3(s))

Transfer function:
50 s - 100
-----------------------
s^3 + s^2 + 150 s - 100

2.
Push G2(s) to the left past the summing junction.

Collapse the summing junctions and add the parallel transfer functions.
 Solutions to Case Studies Challenges 5-9

Push G1(s)G2(s) + G5(s) to the right past the summing junction.

Collapse summing junctions and add feedback paths.

Applying the feedback formula,

5-10 Chapter 5: Reduction of Multiple Subsystems

3.
Combine G6 and G7 yielding G6G7. Add G4 and obtain the following diagram:

Next combine G3 and G4+G6G7.

Push G5 to the left past the pickoff point.

 Solutions to Case Studies Challenges 5-11

Notice that the feedback is in parallel form. Thus the equivalent feedback, H eq(s) = + G3(G4+G6G7)

+ G8. Since the forward path transfer function is G(s) = Geq(s) = G1G5, the closed-loop transfer

function is

T(s) = .
Hence,

4.

Push 3s to the right past the pickoff point.

Combine summing junctions.

5-12 Chapter 5: Reduction of Multiple Subsystems

Combine parallel 3s and s. Apply feedback formula to unity feedback with G(s) = s.

Combine cascade pair and add feedback around 1/(s+1).

 Solutions to Case Studies Challenges 5-13

Combine parallel pair and feedback in forward path.

2
( ) 7 s +3 s
G s=
The cascade pair gives (s+1)(s+5) and

5.
Combine the feedback with G6 and combine the parallel G2 and G3.

Move G2+G3 to the left past the pickoff point.

5-14 Chapter 5: Reduction of Multiple Subsystems

Combine feedback and parallel pair in the forward path yielding an equivalent forward-path transfer

function of

Ge(s) =

But, T(s) = . Thus,

6.

2 ζ ω n=11 ωn ζ
. Therefore, and = 11. Hence, = 0.5<1 so the system is
underdamped.
2
%OS e   / 1 
x100 16.3% ; Ts = =0.7273 sec; Tp = =1.0938 sec.

7.
10 5 2
s (s +7) 1 10 1 3 3
C ( s )= = = − +
10 s s( s +2)(s+ 5) s s+ 2 s +5
1+
s(s+7)

−2 t
5 2 −5 t
c ( t ) =1− e + e
3 3

8.
−4
The forward paths are T 1=2 and T 2=6 s . The loops are L1=−2, L2=−6 s, L3= ,
s
L4 =−12 and L5=−2 s.
4 4
There are no non-touching loops. So ∆=1+2+6 s+ +12+ 2 s=15+8 s + . When either
s s
of the forward paths is removed there are no loops left, so ∆ 1=∆ 2=1. The closed loop
transfer function is

T 1 ∆1 +T 2 ∆ 2 2+6 s 2 s(3 s+1)

T ( s )= = = 2
∆ 4 8 s +15 s+ 4
15+8 s+
s

Solving for the roots of the denominator one gets -0.3219 and -1.5531.
 Solutions to Case Studies Challenges 5-15

9.
K G (s) K
G (s)  T (s)   2
Since s ( s  20) , 1  G ( s ) s  20 s  K . Therefore, 2 = 20. Thus,  =
n
%OS
 ln( )
100  ln 0.15
   0.189
2 %OS  2
 ln 2 0.15
  ln (
2
)
10/n .For a 15% overshoot 100 . Hence, n =
52.95 = . Therefore K = 2803.3.

10.
K
T (s) 
s  s  K ;
2

%OS
 ln(
)
100  ln 0.1
   0.5912
2 %OS  2 2
  ln (
2
)  ln 0.1
100 ;

4
Ts  0.17
 n .

Therefore, n = 39.8; K = n2 = 1584;  = 2n = 47.06.

11.
4
We first find , ω n necessary for the specifications. We have T s= =3 and
ωn
π 3π
T p= =1.5. Eliminating ω n from both equations we get =1.5 . Cross-
ωn √ 1−❑ 2
4 √1−❑2
multiplying, squaring both sides and solving, we get ¿
√ 4
4+ π 2
=0.537 . ω n=2.4829. The closed

loop transfer function of the system is:

30 K 1
s (s +2) 30 K 1
T ( s )= = 2
30 K 1 30 K 2 s s + ( 30 K 2+2 ) s+30 K 1
1+ +
s(s+2) s(s+ 2)
5-16 Chapter 5: Reduction of Multiple Subsystems

From which we get that 30 K 1=ω n or K 1=0.2055 and 30 K 2 +2=2 ω n=2.667 or

2

K 2=0.0222.

12.
20
a. For the inner loop, Ge(s) = s (s +12) , and He(s) = 0.2s. Therefore, Te(s) = = . Combining with

the equivalent transfer function of the parallel pair, Gp(s) = 20, the system is reduced to an equivalent

unity feedback system with G(s) = Gp(s) Te(s) = . Hence, T(s) = = .

−zp / √1−z 2
b. n2 = 400; 2n = 16. Therefore, n = 20, and %OS=e x 100=25 . 38 ;

Ts = =0.5; Tp = =0.171. From Figure 4.16, nTr = 1.463. Hence, Tr = 0.0732.

d = Im = n= 18.33.

13.

The closed loop transfer function of the system is

8550
s (s +100) 8550
T ( s )= = 2
8550 s +100 s+8550
1+
s(s+100)

100 −π
From which we get ω n=92.47 and ¿ =0.54 . The %OS= √1−❑ =13.3%.
2

2ω n 100 e

π
T p= =0.0404 sec. T s= 4 =0.08 sec. From the figure we get that ω n T r ≈ 1.7
ωn √ 1−❑ 2
ωn
from which we get T r=0.0184 sec.

14.
For the generator, Eg(s) = Kf If (s). But, If (s)= . Therefore, = . For the motor,

consider Ra = 2 the sum of both resistors. Also, Je = Ja+JL()2 = 0.75+1x= 1; De

= DL()2 = 1. Therefore,
 Solutions to Case Studies Challenges 5-17

1
= = s (s +3 ) .

0 .5
But, = . Thus, = s (s +3 ) . Finally,

1
= = s(s+1)( s+3 ) .

15.
2
For the mechanical system, J () s22(s) = T() . For the potentiometer, Ei (s)= 10 , or
2(s) = Ei(s). For the network, Eo(s) = Ei(s) = Ei(s) , or Ei(s) = Eo(s) .
1
s+
p RC
q 2 ( s )= E (s )
Therefore, 5 o s . Substitute into mechanical equation and obtain,
5N 1
Eo ( s ) JpN 2
=
T (s)
(
s s+
1
RC . )
16.
a.
5-18 Chapter 5: Reduction of Multiple Subsystems

r 1 50
1 x4 x3 x2 s x1
s
2
s 1

2
2

s

1

b.

x3 xx xx
3 2
2 1
1

xx x
5
5 4
4
 Solutions to Case Studies Challenges 5-19

-G
1 - 1

r x G2 G
1 5 x 3 5 x2 G6 x
1
G3 1 1

x
4

G4
-G
7
c.

x
5
x4
x3 x2 x1

G - G6
8
r x x x x
1 5 G1 4 G3 3 1 2 G x
7 1

-G G4
2

- G
5
5-20 Chapter 5: Reduction of Multiple Subsystems

17.

a.

The transfer function is

C( s) 200
= 3 2
R( s) s +60 s + 1100s +6000

The corresponding differential equation is:

c⃛ +60 c̈ +1100 ċ +6000 c=200 r

The state variables are defined as: c=x 1; ċ=x 2 and c̈=x 3. From which we get

x˙1=x 2
x˙2=x 3

x˙1=−6000 x 1−1100 x 2−60 x 3 +200 r

y=x 1

The corresponding signal flow diagram:

b.

The transfer function can be expressed as G ( s )=200 ( s+10

1
)( s +20
1
)( s+30
1
) from which the

signal flow diagram is obtained.

The state space representation is

x˙1=−30 x 1+ x 2
 Solutions to Problems 5-21

x˙2=−20 x 2+ x3

x˙1=−10 x 3+ 200 r
y=x 1

18.
20
4 3 2
a. Since G(s) = s +15 s +66 s +80 s = ,

···· ··· ·· ·
c +15 c + 66 c + 80 c =20 r

Let,
c=x 1
·
c =x 2
··
c =x 3
···
c =x 4

Therefore,
·
x 1=x 2
·
x 2 =x3
·
x 3 =x 4
·
x 4 =- 80 x 2−66 x 3 −15 x 4 + 20 r
y=x 1

1 x 1 x3 1 x 1 x
r 20 s 4 s s 2 s 1 1 y

-15
-66

-80

20 1 1 1
b. G(s) = ( s )( s +2 )( s +5 )( s +8 ). Hence,
5-22 Chapter 5: Reduction of Multiple Subsystems

1 1 1
1 x x x x
r 20 4 1 s 3 1 s 2 1 s 1 1 y
s

-2 -5 -8

From which,
·
x 1 =- 8 x 1 + x 2
·
x 2 =- 5 x 2 + x 3
·
x 3 =- 2 x 3 + x 4
·
x 4 =20 r
y=x 1
19.

 = 1 + [G2G3G4 + G3G4 + G4 + 1] + [G3G4 + G4]; T1 = G1G2G3G4; 1 = 1. Therefore,

T(s) = =

20.

Closed-loop gains: -s2; - ; - ; -s2

Forward-path gains: T1 = s; T2 =

Nontouching loops: None

 = 1 - (-s2 - - - s2)

1 = 2 = 1

G(s) = = =
 Solutions to Problems 5-23

21.
( s+1 )( s +2) 2 5 6
G ( s )= 2
= 2
− +
a. ( s +3 ) ( s + 4 ) ( s+3 ) s+3 s+ 4

Writing the state and output equations,

x1 = -3x1 + x2
x2 = -3x2 + r
x3 = -4x3 + r

y = 2x1 - 5x2 + 6x3

In vector-matrix form,

[ ] []
· −3 1 0 0
x = 0 −3 0 x+ 1 r
0 0 −4 1
y= [ 2 −5 6 ]
( s +2) 3 /4 1 5/ 4 1
G ( s )= =- + − −
b. G(s) =
2
( s +5 ) ( s +7 )2
( s+ 5) s +5 ( s +7 )
2 2 s +7

1 1
x x
s 2 1 s 1
- 3
-5 -5 4
r 1 y
1
1 1 x 1 x
s 4 1 3
s - 5
4
-7 -7

-1
5-24 Chapter 5: Reduction of Multiple Subsystems

Writing the state and output equations,

x1 = -5x1 + x2
x2 = -5x2 + r
x3 = -7x3 + x4
x4 = -7x4 + r
y = - x1 + x2 - x3 - x4
In vector matrix form,
-5 1 0 0 0
 0 -5 0 0 1
x= x+ r
0 0 -7 1 0
0 0 0 -7 1

3 5
y = - 1 - -1 x
4 4
c.

G( s )=
s +4
=
1 1
−
1 1
− ( ) ( ) ( ) (
1 1
+
1 1
(s +2)2 (s +5 )( s+ 6) 8 s +6 9 s+5 72 s+2 6 ( s+2 )2 )

Writing the state and output equations,

x1 = - 2x1 + x2
x2 = - 2x2 + r
x3 = - 5x3 + r
x4 = - 6x4 + r
1 1 1
y = x1 -
72 x2 - 9 x3 + 8 x4

In vector-matrix form,
 Solutions to Problems 5-25

y=
[ 1
6
−
1
72
−
1
9
1
8]x

22.
a.

Writing the state equations,

x1 = x2
x2 = - 7x1 - 2x2 + r

y = 3x1 + x2
In vector matrix form,

b.
5-26 Chapter 5: Reduction of Multiple Subsystems

Writing the state equations,

·
x 1=x 2
·
x 2 =x3
·
x 3 =- x 1 −2 x 2 −5 x3 +r
y=6 x 1 + 2 x 2 + x 3

In vector matrix form,

[ ] []
· 0 1 0 0
x= 0 0 1 X + 0 r
−1 −2 −5 1
y= [ 6 2 1 ] x
 Solutions to Problems 5-27

c.

x1 = x2

x2 = x3

x3 = x4

x4 = - 4x1 - 6x2 - 5x3 - 3x4 + r

y = x1 + 7x2 + 2x3 + x4

In vector matrix form,

23.
a. Controller canonical form:

From the phase-variable form in Problem 5.31(a), reverse the order of the state variables and obtain,
x2 = x1
x1 = - 7x2 - 2x1 + r

y = 3x2 + x1
5-28 Chapter 5: Reduction of Multiple Subsystems

Putting the equations in order,

x1 = - 2x1 - 7x2 + r
x2 = x1

y = x1 + 3x2

In vector-matrix form,

[ ] []
·
x = −2 −7 x + 1 r
1 0 0
y= [ 1 3 ] x
Observer canonical form:

G(s) = . Divide each term by and get

1 3
+
s 2
s C(s)
G(s) = =
2 7 R(s)
1+ +
s 2
s
Cross multiplying,

(+ ) R(s) = (1 + + ) C(s)
Thus,

+ = C(s)

Drawing the signal-flow graph,

R(s) 1

-2

-7
Writing the state and output equations,
x1 = - 2x1 + x2 + r
x2 = - 7x1 + 3r

y = x1
In vector matrix form,
 Solutions to Problems 5-29

[ ] []
·
−2 1 1
x= x+ r
−7 0 3
y= [ 1 0 ] x
b. Controller canonical form:

From the phase-variable form in Problem 5.31(b), reverse the order of the state variables and obtain,
·
x 3 =x 2
·
x 2 =x1
·
x 1 =- x 3 −2 x 2 −5 x1
y=6 x 3 +2 x 2 + x1
Putting the equations in order,
·
x 1 =- 5 x 1−2 x 2−x 3
·
x 2 =x1
·
x 3 =x 2
y=x 1 + 2 x 2 +6 x 3
In vector-matrix form,

[ ] []
· −5 −2 −1 1
x = 1 0 0 x+ 0 r
0 1 0 0
y= [ 1 2 6 ] x
Observer canonical form:

s 2 +2s +6 1
G ( s )= 3 2
s +5s +2s+1 . Divide each term by s 3 and get

1 2 6
+ +
s s2 s3 C (s)
G ( s )= =
5 2 1 R( s )
1+ + 2 + 3
s s s
Cross-multiplying,

( 1s + s2 + s6 ) R( s )=(1+ 5s + s2 + s1 ) C ( s )
2 3 2 3

Thus,
1 1 1
( R( s )−5c ( s ) )+ 2 ( 2R ( s )−2C( s ))+ 3 ( 6R( s )−C ( s ) )=C ( s )
s s s

Drawing the signal-flow graph,

5-30 Chapter 5: Reduction of Multiple Subsystems

1 1 1
6 s X3(s) X2(s) X1(s)
1 s 1 s 1
R(s) C(s)

-5

-2

-1
Writing the state and output equations,
·
x 1 =- 5 x 1 + x 2 +r
·
x 2 =- 2 x1 + x 3 +2 r
·
x 3 =- x 1 +6 r
y= [ 1 0 0 ] x
In vector-matrix form,

[ ] []
· −5 1 0 1
x = −2 0 1 x+ 2 r
−1 0 0 6
y= [ 1 0 0 ] x
c. Controller canonical form:

From the phase-variable form in Problem 5.31(c), reverse the order of the state variables and obtain,

x4 = x3

x3 = x2

x2 = x1

x1 = - 4x4 - 6x3 - 5x2 - 3x1 + r

y = x4 + 7x3 + 2x2 + x1
Putting the equations in order,

x1 = - 3x1 - 5x2 - 6x3 - 4x4 + r

x2 = x1
 Solutions to Problems 5-31

x3 = x2

x4 = x3

y = x1 + 2x2 +7x3 + x4
In vector-matrix form,

[ ] []
−3 −5 −6 −4 1
·
1 0 0 0 0
x= X+ r
0 1 0 0 0
0 0 1 0 0
y= [ 1 2 7 1 ] x
Observer canonical form:

G(s) = . Divide each term by and get

1 2 7 1
+ + +
s 2 3 4
s s s C(s)
G(s) = =
3 5 6 4 R(s)
1+ + + +
s 2 3 4
s s s

Cross multiplying,

(+ + + ) R(s) = (1 + + + + ) C(s)

Thus,

+ + + = C(s)

Drawing the signal-flow graph,

5-32 Chapter 5: Reduction of Multiple Subsystems

R(s) 2
7
1

-3

-5

-6

-4

Writing the state and output equations,

x1 = - 3x1 + x2 + r
x2 = - 5x1 + x3 + 2r
x3 = - 6x1 + x4 +7r
x4 = - 4x1 + r
y = x1
In vector matrix form,
 Solutions to Problems 5-33

24.
a.
1 1 x 1
1 s x3 1 s 2 1 s x1
50
r c= y

-9-2 -8-5 -2-7

-1

Writing the state equations,

In vector-matrix form,

b.

1 x3 1 x 1
2 x1
1 s 10 s s
r c= y

-6
-8
-24
-25

-1

Writing the state equations,

5-34 Chapter 5: Reduction of Multiple Subsystems

In vector-matrix form,

c.
1 x 1
1 100 1 s 2 1 s x1
r c= y
-1

-1 Tach feedback
before integrator

-1

x1 = x2

x2 = -x2 - x2 + 160(r-x1) = -160x1 -2x2 +160r

y = x1

In vector-matrix form,

y = 1 0 x
 Solutions to Problems 5-35

d. Since = , we draw the signal-flow as follows:

1

1 1
1 10 s x s x1 2
r 2
c= y
-2

-1

-1

Writing the state equations,

x1 = x2

x2 = -x1 - 2x2 + 16(r-c) = -x1 - 2x2 + 16(r - (2x1+x2) = -33x1 - 18x2 + 16r

y = 2x1 + x2
In vector-matrix form,

y= 2 1 x
25.

a. Phase Variable form

20
T ( s )= 3 2
s + 8 s +12 s+20

x˙1=x 2
x˙2=x 3

x˙3=−20 x 1−12 x2 −8 x3 +20 r

y=x 1

Or in matrix form
5-36 Chapter 5: Reduction of Multiple Subsystems

[ ][]
0 1 0 0
ẋ= 0 0 1 x+ 0 u
−20 −12 −8 20

y= [ 1 0 0 ] x

b. Parallel form
5 5 5
20 3 2 6
G ( s )= = − +
s (s+2)(s+6) s s +2 s+6

The state equations are:

5 −5 5 5 5
x˙1= ( u−x1 −x2 −x3 ) = x − x 2− x 3+ u
3 3 1 3 3 3
−5 5 1 5 5
x˙2=
2
( u−x 1−x 2−x 3) −2 x2 = 2 x + 2 x 2+ 2 x3 − 2 u
1
5 −5 5 41 5
x˙3= ( u−x 1−x 2−x 3 )−6 x 3= x − x2 − x 3+ u
6 6 1 6 6 6
y=x 1 + x 2+ x3

In matrix form:
 Solutions to Problems 5-37

[ ][]
−5 −5 −5 5
3 3 3 3
5 1 5 −5
ẋ= x+ u
2 2 2 2
−5 −5 −41 5
6 6 6 6

y= [ 1 1 1 ] x

26.
10 ( s 2 +5 s+6 )
T ( s )=
a. s 4 +16 s3 +99 s2 + 244 s +180

Drawing the signal-flow diagram,

5
1 1 1 1
10 s s s s 6
r y
x4 x x2 x1
3
16

99
244

180

Writing the state and output equations,

·
x 1=x 2
·
x 2 =x3
·
x 3 =x 4
·
x 1 =- 180 x 1−244 x 2 −99 x 3 −16 x 4 + 10 r
y=6 x 1 + 5 x 2 + x3

In vector-matrix form,
5-38 Chapter 5: Reduction of Multiple Subsystems

[ ][]
0 1 0 0 0
·
0 0 1 0 0
x= x+ r
0 0 0 1 0
−180 −244 −99 −16 10
y= [ 6 5 1 0 ] x

10( s+2 )( s +3) 1 /3 10/ 3 15 12

G( s )= = − + −
b. ( s +1)( s+ 4 )( s +5 )( s+ 6) s +1 s +4 s+5 s +6

Drawing the signal-flow diagram and including the unity-feedback path,

1
s

1 x1 1
-1
3
 10 1
1 3 s 1
r=u y
-4
x 1
2

15 1 1
s
x
-5 3

-12 1
1
s
x4
-6

-1

Writing the state and output equations,

·
1
x 1= ( u−x 1 −x 2−x 3 −x 4 )−x1
3
·
−10
x 2= ( u−x1 −x 2 −x3 −x 4 )−4 x 2
3
·
x 3 =15( u−x 1 −x 2−x 3 −x 4 )−5 x 3
·
x 4 =- 12( u−x 1 −x 2−x 3 −x 4 )−12 x 4
y=x 1 + x2 + x 3 + x 4
In vector-matrix form,
 Solutions to Problems 5-39

[ ][ ]
4 1 1 1 1
− − − −
3 3 3 3 3
·
10 2 10 10 10
x= − x+ − u
3 3 3 3 3
−15 −15 −20 −15 15
12 12 12 0 −12
y= [ 1 1 1 1 ] x
5-40 Chapter 5: Reduction of Multiple Subsystems

27.
Program:
'(a)'
'G(s)'
G=zpk([-2 -3],[-1 -4 -5 -6],10)
'T(s)'
T=feedback(G,1,-1)
[numt,dent]=tfdata(T,'v');
'Find controller canonical form'
[Acc,Bcc,Ccc,Dcc]=tf2ss(numt,dent)
A1=flipud(Acc);
'Transform to phase-variable form'
Apv=fliplr(A1)
Bpv=flipud(Bcc)
Cpv=fliplr(Ccc)
'(b)'
'G(s)'
G=zpk([-2 -3],[-1 -4 -5 -6],10)
'T(s)'
T=feedback(G,1,-1)
[numt,dent]=tfdata(T,'v');
'Find controller canonical form'
[Acc,Bcc,Ccc,Dcc]=tf2ss(numt,dent)
'Transform to modal form'
[A,B,C,D]=canon(Acc,Bcc,Ccc,Dcc,'modal')

Computer response:
ans =

(a)

ans =

G(s)

Zero/pole/gain:
10 (s+2) (s+3)
-----------------------
(s+1) (s+4) (s+5) (s+6)

ans =

T(s)

Zero/pole/gain:

10 (s+2) (s+3)
------------------------------------------
(s+1.264) (s+3.412) (s^2 + 11.32s + 41.73)

ans =

Find controller canonical form

Acc =

-16.0000 -99.0000 -244.0000 -180.0000

1.0000 0 0 0
0 1.0000 0 0
0 0 1.0000 0
Bcc =
 Solutions to Problems 5-41

1
0
0
0

Ccc =

0 10.0000 50.0000 60.0000

Dcc =

ans =

Transform to phase-variable form

Apv =

0 1.0000 0 0
0 0 1.0000 0
0 0 0 1.0000
-180.0000 -244.0000 -99.0000 -16.0000

Bpv =

0
0
0
1

Cpv =

60.0000 50.0000 10.0000 0

ans =

(b)

ans =

G(s)

Zero/pole/gain:
10 (s+2) (s+3)
-----------------------
(s+1) (s+4) (s+5) (s+6)

ans =

T(s)

Zero/pole/gain:
10 (s+2) (s+3)
------------------------------------------
(s+1.264) (s+3.412) (s^2 + 11.32s + 41.73)

ans =

Find controller canonical form

5-42 Chapter 5: Reduction of Multiple Subsystems

Acc =

-16.0000 -99.0000 -244.0000 -180.0000

1.0000 0 0 0
0 1.0000 0 0
0 0 1.0000 0

Bcc =

1
0
0
0

Ccc =

0 10.0000 50.0000 60.0000

Dcc =

ans =

Transform to modal form

A=

-5.6618 3.1109 0 0
-3.1109 -5.6618 0 0
0 0 -3.4124 0
0 0 0 -1.2639

B=

-4.1108
1.0468
1.3125
0.0487
C=

0.1827 0.6973 -0.1401 4.2067

D=

0
28.
 Solutions to Problems 5-43

x˙1=x 2

x˙2=−4 x 1+ x 3

x˙3=x 4
x˙4 =7 x1−x 2 +r

y=−7 x 1 + x 2

Or in Matrix form

29.
a.

q1 + 5q1 + 61 - 3q2 - 42 = 0

-3q1 - 41 + q2 + 5q2 + 52 = T

or

q1 = - 5q1 - 61 + 3q2 + 42

q2 = 3q1 + 41 - 5q2 - 52 + T

Letting, 1 = x1 ; q1 = x2 ; 2 = x3 ; q2 = x4 ,
5-44 Chapter 5: Reduction of Multiple Subsystems

1 1
s x2 s
x1

4 3 -5
-6

4
3

1 1
s x4 s x3
T
1
-5
-5

where x = .

b. Using the signal-flow diagram,

x1 = x2

x2 = -6x1 - 5x2 + 4x3 + 3x4

x3 = x4

x4 = 4x1 + 3x2 - 5x3 - 5x4 + T

y = x3
In vector-matrix form,
0 1 0 0 0
 -6 -5 4 3 0
x= x+ T
0 0 0 1 0
4 3 -5 -5 1

y= 0 0 1 0 x
 Solutions to Problems 5-45

30.

x1 = A1x1 + B1r (1)

y1 = C1x1 (2)

x2 = A2x2 + B2y1 (3)

y2 = C2x2 (4)

Substituting Eq. (2) into Eq. (3),

x1 = A1x1 + B1r

x2 = B2C1x1 + A2x2

y2 = C2x2
In vector-matrix notation,

x1 A 1 O x1
= - + B1 r
x2 B2 C 1 A 2 x 2 O

y 2 = O C 2 x1
x2

31.

x1 = A1x1 + B1e (1)

y = C1x1 (2)

x2 = A2x2 + B2y (3)

p = C2x2 (4)
Substituting e = r - p into Eq. (1) and substituting Eq. (2) into (3), we obtain,

x1 = A1x1 + B1(r - p) (5)

y = C1x1 (6)
x2 = A2x2 + B2C1x1 (7)
p = C2x2 (8)

Substituting Eq. (8) into Eq. (5),

x1 = A1x1 - B1C2x2 + B1r

x2 = B2C1x1 + A2x2

y = C1x1
In vector-matrix form,
5-46 Chapter 5: Reduction of Multiple Subsystems

x1 = A 1 - B 1 C 2 x1 + B 1 r
x2 B 2 C1 A2 x2 0
x1
y = C1 0
x2
32.
·
z =P−1 APz+P−1 Bu
y=CPz

33.

Eigenvalues are 1, -2, and 3 since,

|I - A | = ( - 3) ( + 2) ( - 1)

Solving for the eigenvectors, Ax = x

or,

For  = 1, x1 = x2 = . For  = -2, x1 = 2x3, x2 = -3x3. For  = 3, x1 = x3 , x2 = -2x3 . Thus,

z= P-1APz + P-1Bu ; y = CPz, where

 Solutions to Problems 5-47

34.
Program:
A=[-10 -3 7;18.25 6.25 -11.75;-7.25 -2.25 5.75];
B=[1;3;2];
C=[1 -2 4];
[P,d]=eig(A);
Ad=inv(P)*A*P
Bd=inv(P)*B
Cd=C*P

Computer response:
Ad =

-2.0000 0.0000 0.0000

-0.0000 3.0000 -0.0000
0.0000 0.0000 1.0000

Bd =

1.8708
-3.6742
3.6742

Cd =

3.2071 3.6742 2.8577

35.
Push Pitch Gain to the right past the pickoff point.

Collapse the summing junctions and add the feedback transfer functions.
5-48 Chapter 5: Reduction of Multiple Subsystems

Apply the feedback formula and obtain,

G( s ) 0 . 25( s +0 . 435 )
T ( s )= = 4
1+G( s ) H ( s ) s +3 . 4586 s +3 . 4569 s 2 +0 . 9693 s +0 . 15032
3

36.
 Solutions to Problems 5-49

37.
a. The first equation follows from the schematic. The second equation is obtained by applying the
voltage divider rule at the op-amp’s inverting terminal, noting that since the op-amp considered is
ideal, there is no current demand there.
Ri Ri
L=− A L=− A
T =A;
b. 1
Ri + R f ;
Ri + R f ;  1 1
vo T 1 Δ1 A
= =
vi Δ Ri
1+ A
Ri + R f
vo A 1 Rf
=Lim = =1+
vi A →∞ Ri Ri Ri
1+ A
c.
Ri + R f Ri + Rf
 5-50 Chapter 5: Reduction of Multiple Subsystems

38.
a. The three equations follow by direct observation from the small signal circuit.
b.The block diagram is given by

vi Vi Ri + vgs Vo
Ri+ Rs + gm ( R s||r o ) vo
-
-

a. From the block diagram we get

vo gm ( R s||r o ) Ri
=
v i 1+ g m ( R s||r o ) Ri + Rs

39.

a. Using Mason’s rule

Kt 1 ω 20 ω 20 ε εmr
T 1= 2 2 M
= 2 2
L 1 =+ 2 2 s +ε
mr L =−
s + ω0 US s +ω 0 ; Loops s +ω0 and
2
s+ ε , no

non-touching loops.
Δ 1=1
ω 20 ω20
X 3 T 1 Δ1 s2 +ω20 s 2 +ω 20
= = =
R Δ ε ω 20 ε ε mr s 2
1+m r − 2 m 1+
s+ ε s +ω 0 s+ ε r
2
( s+ ε )( s2 +ω20 )
a. From part (a)

ω20 ε
X1 X3 ε 2
s + ω20 s +ε
= =
R R s +ε ε m r s2
1+
( s +ε )( s 2 + ω20 )
 Solutions to Problems 5-51

40.

a.

>> A=[-100.2 -20.7 -30.7 200.3; 40 -20.22 49.95 526.1;...

0 10.22 -59.95 -526.1; 0 0 0 0];

>> B=[208; -208; -108.8; -1];

>> C = [0 1570 1570 59400];

>> D = -6240;

>> [n,d]=ss2tf(A,B,C,D)

n=

1.0e+009 *

Columns 1 through 3

-0.00000624000000 -0.00168228480000 -0.14206098728000

Columns 4 through 5

-3.91955218234127 -9.08349454230472

d=

1.0e+005 *

Columns 1 through 3
5-52 Chapter 5: Reduction of Multiple Subsystems

0.00001000000000 0.00180370000000 0.09562734000000

Columns 4 through 5

1.32499100000000 0

>> roots(n)

ans =

1.0e+002 *

-1.34317654991673

-0.78476212102923

-0.54257777928519

-0.02545278053809

>> roots(d)

ans =

-92.38329312886714

-66.38046756013043

-21.60623931100260
 Solutions to Problems 5-53

Y
s (0 )=68555 . 14
Note that U , follows that

Y 6348 . 17( s+2 . 5)( s+54 . 3 )(s+78. 5 )(s+134 .3 )

( s)=−
U s(s+21. 6 )(s+66 . 4 )( s+92 . 4 )

b.

>> [r,p,k]=residue(n,d)

r=

1.0e+005 *

-0.73309459854184

-0.51344619392820

-3.63566779304453

-0.68555141448543

p=

-92.38329312886714

-66.38046756013043

-21.60623931100260

k=
5-54 Chapter 5: Reduction of Multiple Subsystems

-6240

Y 73309 . 46 51344 . 6 363566. 8 68555 .14

( s)=−6240− − − −
or U s s+21. 6 s+66 . 4 s+92. 4

c.

d. The corresponding state space representation is:

[ ][ ][ ] [ ]
ẋ 1 0 0 0 0 x1 −73309. 5
ẋ 2 −21 x2
= 0 .6 0 0 + −51344 . 6 u(t )
ẋ 3 0 0 −66 . 4 0 x 3 −363566 . 8
ẋ 4 0 0 0 −9 . 4 x 4 −68555 . 14
 Solutions to Problems 5-55

[]
x1
x
y= [ 1 1 1 1 ] 2 −6240 u(t )
x3
x4

41.

a.

>> A = [0 1 0; 0 -68.3 -7.2; 0 3.2 -0.7]

A=

0 1.0000 0

0 -68.3000 -7.2000

0 3.2000 -0.7000

>> [V,D]=eig(A)

V=

1.0000 0.0147 -0.1016

0 -0.9988 0.1059

0 0.0475 -0.9892

D=

0 0 0

0 -67.9574 0
5-56 Chapter 5: Reduction of Multiple Subsystems

0 0 -1.0426

Matrix V is the sought similarity transformation.

b.

>> Ad = inv(V)*A*V

Ad =

0 -0.0000 -0.0000

0 -67.9574 0.0000

0 -0.0000 -1.0426

>> B = [0;425.4;0]

B=

425.4000

>> Bd = inv(V)*B

Bd =

4.2030

-428.1077
 Solutions to Problems 5-57

-20.5661

The diagonalized system is:

[][ ][ ] [ ]
ż1 0 0 0 z1 4.2
ż 2 = 0 −67 . 96 0 z2 + −428. 11 e m
ż 3 0 0 −1 . 0426 z −20. 57
3

42.

a.

b.There is only one forward path

T 1= H e H h

There are three loops:

L1 =−H e H h ; L2 =−H i H h and L3 =−H i H r H act
L1 and L3 are non-touching loops so
Δ=1−L1 −L2 −L3 +L1 L3 =1+H e H h +H i H h +H i H r H act + H e H h H i H r H act

When
T 1 is eliminated only L3 is left so Δ 1=1−L3 =1+ H r H i H act
Fh T Δ H e H h ( 1+ H r H i H act )
( s )= 1 1 =
Finally
D Δ 1+ H e H h + H i H h + H i H r H act + H e H h H i H r H act

43.
a. There are two forwards paths:

1 1
M 1= K h K hs C s 2 and
M 2= −K h T es K es C s
s s

The loops are:

1
L1=−K h K hs C s T
2 hs
s
1
L2=+ K h T es K es Cs T hs
s
5-58 Chapter 5: Reduction of Multiple Subsystems

L3=−C j T ej K ej

L4 =−K h C j T hj

There are no non-touching loops. Therefore

1 1
∆=1+ K h K hs C s T −K h T es K es C s T hs + C j T ej K ej + K h C j T hj
2 hs
s s

Also ∆ 1=∆ 2=1

1 1
−K h T es K es C s
K h K hs C s
Y s M 1 ∆1 + M 2 ∆2 s
2
s
= =
Uh ∆ 1 1
1+ K h K hs C s 2 T hs−K h T es K es C s T hs +C j T ej K ej + K h C j T hj
s s

b.

There is only one forward path M 1=K h C j

The loops and are the same as in part a. Also ∆ 1=1. It follows that
Y j M 1 ∆1 KhC j
= =
Uh ∆ 1 1
1+ K h K hs C s T −K h T es K es C s T hs +C j T ej K ej + K h C j T hj
2 hs
s s

44.
−1 −1
a. Assuming Z h=0 there are two forward paths, M 1=Z m and M 2=C 6 Z m
The loops are
L1=−G s C s
−1
L2=−Z m C m
−1
L3=−Z m C 1 G s C 4
−1
L4 =−Z m C s G s Z e C 2

There are two non-touching loops L1 and L2.

−1 −1 −1 −1
∆=1+Gs C s +Z m C m+ Z m C1 Gs C 4 +Z m C s G s Z e C 2−Gs C s Z m C m
−1 −1 −1
¿ 1+Gs C s +Z m C m+ Z m C s Gs Z e C 2=1+Gs C s +Z m (C ¿ ¿ m+C s G s Z e C 2)¿
We also have that by eliminating M 1 or M 2
∆ 1=∆ 2=1+Gs C s
−1 −1 −1
X h M 1 ∆1+ M 2 ∆ 2 (Z m + Z m C6 )(1+Gs C s) Z m C2 (1+G s C s )
Y ( s )= = = −1
= −1
Fh ∆ 1+Gs C s +Z m (C m +C 2 Z e Gs C s) 1+Gs C s+ Z m (C m +C 2 Z e Gs C s)

b. The system can be described by means of the following diagram:

 Solutions to Problems 5-59

It follows that
Xh Y (s )
=
F h 1+ Y (s )Z h

45.
a. Substituting the values given above into the block diagram, when vp = 0, we have:

GV(s) GX(s) Gm(s)

GC(s)

R(s) ΔHm(s
E(s) YC(s) X(s)
+ 0.63 )
s  0.926
_
β

The Mould Level Block Diagram for vp = 0

Thus, the closed-loop transfer function is:

ΔH m ( s ) G C ( s )⋅G V ( s )⋅G X ( s )⋅G m ( s )

T ( s )= =
R (s) 1+ β⋅G C ( s )⋅G V ( s )⋅G X ( s )⋅G m ( s ) =
5-60 Chapter 5: Reduction of Multiple Subsystems

1 . 6 ( s2 +1. 25 s +0 . 25 ) 100
s ( )( s+00..63926 ) 0 .51 s
s 2 +10 s +100
1 . 6 ( s +1 . 25 s +0 .25 ) 100
( s +10 s +100 ) s+0. 926 0 . 5 s
( )
2
0 .63 1
1+0 .5× 2
s
= =

201 .6 ( s 2 +1 .25 s+0 .25 )

= s 2 ( s2 +10 s+100 ) ( s+0 . 926 ) +0 . 5×201 . 6 ( s 2 +1 . 25 s+0 .25 ) =

201 .6 ( s 2 +1 .25 s+0 .25 )

5 4 3 2
= s +10 . 93 s +109 . 26 s +193 . 4 s +126 s+25 . 2

b. Simulink was used to simulate the system. The model of that system is shown in Figure P5.x-4*. The
parameters of the PID controller were set to: Kp = 2, Kd = 1.6, and KI = 0.4. The reference step, r(t)
= 5 u(t), and the casting speed step, vp (t) = 0.97 u(t) were set to start at t = 0. An adder was used to
add the initial value,
Hm (0—) = – 75 mm, at the output, to the change in mould level, Hm.

The time and mould level (in array format) were output to “workspace ” sinks, each of which was
given the respective variable name. After the simulation ended, Matlab plot commands were used
to obtain and edit the graph of hm(t) from 0 to t = 80 seconds.

Casting
Speed
Casting Speed Coefficient 75 time

Initial M ould Clock Time

6
Level

100 0.63 1
PID(s)
s2+10s+100 s+0.926 0.5s Level XY Graph
Flow-in
Reference PID Controller Linearized M ould Adder
Valve minus
Step Flow-in Flow-out
Level Sensor Hm
Sensitivity M ould Level

0.5

Simulink Model of the Mould Level Control System

 Solutions to Problems 5-61

-64

-66
Mould Level, hm, in Response to a Reference Signal r(t) = 5 u(t)
and a Casting Speed vp (t) = 0.97 u(t).
At t = 0, hm (0) = - 75.0 mm
-68
Steady state (final) value (at t = 80 sec), hm = - 65.0 mm
Mould Level, hm, mm

-70

-72

-74

-76

-78
0 10 20 30 40 50 60 70 80
Time, sec

Response of the Mould Level to Simultaneous Step Changes in Reference Input, r(t) = 5 u(t), and Casting

Speed, vp (t) = 0.97 u(t) at an Initial Level, Hm (0—) = – 75

46.

R(s) C(s)
+
+

R(s) - H (s) G( s ) C(s)

- -

It can be easily verified that the closed loop transfer function for this system is identical to the

original.
5-62 Chapter 5: Reduction of Multiple Subsystems

47.
∆Vt G(s)
The closed loop transfer function is = where
∆ V ref 1+G ( s ) H (s)
2 2
0.3 s +1.6 s +0.4 10 1 1 75 s + 400 s+ 100
G ( s )= = and
s 0.1 s+1 0.4 s+ 1 s+ 1 s(s+ 1)( s+2.5)(s+10)
1 100
H ( s )= =
0.01 s +1 s+100

Substituting
2
75 s + 400 s +100
∆Vt s( s+ 1)(s+ 2.5)(s+ 10) 75 s 3 +7900 s 2 +40100 s+10000
= 2
= 5 4 3 2
∆ V ref 7500 s +40000 s+10000 s + 113.5s +1387.5 s +11275 s + 42500 s +10000
1+
s(s+ 1)(s+2.5)( s+ 10)(s+ 100)

48.
The MATLAB M-file is:

num1=25;
num2=[1 1.2 12500];
den1=[1 0];
den2=[1 5.6 62000];
num=conv(num1, num2);
den=conv(den1, den2);
G=tf(num, den);
D=feedback(G,0.1);
[numd,dend]=tfdata(D,'v');
numcm=[40 5];
dencm=[1 0];
numOL=conv(numd, numcm);
denOL=conv(dend, dencm);
Omega_OL=tf(numOL, denOL);
Omega_CL=260*feedback(Omega_OL,1);
step(Omega_CL, 0:0.002:0.2);
grid;

After the above file is run, MATLAB’s command window may be used to obtain the

requested minor-loop transfer function:

 Solutions to Problems 5-63

D=

25 s^2 + 30 s + 312500
----------------------------------------
s^3 + 8.1 s^2 + 62003 s + 31250

Continuous-time transfer function.

The MATLAB figure, shown below, illustrates the step-response obtained with all of the
requested important characteristics
System:noted on it.
Omega_CL
Peak amplitude: 300
Overshoot (%): 15.3
At time (seconds): 0.042 Step Response
300
System: Omega_CL
Final value: 260

250 System: Omega_CL

Settling time (seconds): 0.107
Shaft Angular Speed, rad/sec

System: Omega_CL
Rise time (seconds): 0.0262
200

150

100

50

0
0 0.02 0.04 0.06 0.08 0.1 0.12 0.14 0.16 0.18 0.2
Time (seconds)

49.

a.

Note that due to the topology, the loop on the top should have no influence whatsoever on

output x ss. Applying Mason’s: There are two forward paths T 1=s m ss H ss and T 2=C ss H ss .
2

There are two loops L1=−C ss H ss and L2=−C ls H ls; both loops are non-touching. Thus

∆=1+C ss H ss +C ls H ls +C ss H ss Cls H ls . Eliminating forward path 1: ∆ 1=1+Cls H ls.

Eliminating forward path 2: ∆ 2=1+C ls H ls. The closed loop transfer function is:
2 2 2
x ss (s) T 1 ∆ 1+T 2 ∆2 (s mss H ss +C ss H ss )(1+Cls H ls ) s mss H ss +C ss H ss s mss H ss +C
= = = =
r (s ) ∆ 1+C ss H ss +Cls H ls +C ss H ss C ls H ls C H +C H C H 1+C ss H
1+ ss ss ss ss ls ls
1+ Cls H ls
5-64 Chapter 5: Reduction of Multiple Subsystems

b.
2 2
In this case there are three forward paths T 1=s ¿¿ ; T 2=−s m ss H ss Cls H ls;
T 3=−C ss H ss C ls H ls. The loops and ∆ are just as in part a. Eliminating forward path
1 ∆ 1=1+C ss H ss . Eliminating either paths 2 or 3 ∆ 2=∆ 3=1. The closed loop
transfer function is:

xls (s) T 1 ∆1 +T 2 ∆ 2+ T 3 ∆3 2
= =s ¿ ¿ ¿
r ( s) ∆

50.
There are three forward paths T 1=k f F m T pi T ma; T 2=T pa T ma and T 3=T fa. There are
two loops: L1=−F m T pi R i H c and L2=−F m T pi T ma k r . There are no non-touching
loops. So ∆=1+ F m T pi R i H c + F m T pi T ma k r. Eliminating forward paths 1 or 2 leaves no
loops so ∆ 1=∆ 2=1. Eliminating forward path 3 leaves one loop so ∆ 3=1+ F m T pi Ri H c .
The closed loop transfer function is:

v^0 ( s) T 1 ∆1 +T 2 ∆2 +T 3 ∆ 3 k f Fm T pi T ma+T pa T ma +T fa (1+ Fm T pi Ri H c )

= =
v^i (s) ∆ 1+ F m T pi Ri H c + F m T pi T ma k r

51.
a.
Showing the equivalent circuit
here for reference we have:

1
Cs [ acF
V c (s )= I (s )−I acR ( s ) ]

(1)

Substituting the equation given in

the problem into (1), re-arranging,

and simplifying gives:

( )
1
LC
V c ( s )=[ V a ( s )−Ls I acR ( s ) ]
2 R 1
s + s+
L LC (2)
 Solutions to Problems 5-65

b.
Substituting circuit parameters into the equation given in the problem statement & (2), and
assuming zero initial conditions, we have:

103 (s +50000) 103 s

I acF ( s )= 2 I (s )+ V (s )
(s +10 3 s +5×10 7 ) acR s 2 +103 s+5×107 a
(3)

5×10 7 5×10 4 s
V c (s )=V c ( s )−V c (s )= V a ( s )− I acR ( s )
1 2 s 2 +103 s+5×107 s 2 +10 3 s +5×10 7
(4)

The Simulink model is shown below. The current source, IacR, (a step input) was set to

model iacR (t) = 10 u(t), amps, and the voltage ramp, va (t) = 20 t u(t), was created using a

step of 20 volts and an integrator with the upper saturation limit set to 20 volts and the lower

to 0. The result of the simulation was captured on a “scope” that was set to display the two

input variables (using a mux) from t = 0 to 15 ms, with the y-axis range set from – 60 to +60.

Simulink model of the system

5-66 Chapter 5: Reduction of Multiple Subsystems

Inductor Current, iacF(t) in amps, and Capacitor Voltage, vC(t) in volts, as Displayed on the scope at the end

of simulation (set to 15 ms).

 Solutions to Design Problems 5-67

SOLUTIONS TO DESIGN PROBLEMS

52.

Je = Ja+JL()2 = 2+2 = 4; De = Da+DL()2 = 2+DL()2. Therefore, the forward-path transfer function is,

G(s) = (1000) . Thus, T(s) = = .

Hence,  = = 0.456; n = ; 2n = . Therefore De = 10.9; from which DL = 3560.

53.

The equivalent forward path transfer function is Ge(s) = . Thus, T(s) = = . Prior to tachometer

compensation (K2 = 0), T(s) = . Therefore K = n2 = 100. Thus, after tachometer compensation, T(s)

= . Hence, n = 10; 2n = 1+K2. Therefore, K2 = 2n - 1 = 2(0.69)(10) - 1 = 12.8.

5-68 Chapter 5: Reduction of Multiple Subsystems

54.
At the N2 shaft, with rotation,q L ( s )

(J eq s2 +Deq s)q L (s)+F (s)r=T eq ( s)

F ( s)=( Ms 2 +f v s) X (s)
Thus,

(J eq s2 +Deq s )q L (s )+( Ms 2 +f v s ) X (s )r=T eq (s )

But, X ( s )=rq L ( s ) . Hence,

[ (J eq + Mr 2 )s 2+( Deq + f v r 2) s ] q L ( s )=T eq ( s )

where

J eq =J a ( 2)2 + J =5
D eq =D a (2)2 + D=4+ D
r =2
Thus, the total load inertia and load damping is

J L=J eq +Mr 2 =5+4 M

D L=Deq +f v r 2 =4 +D+(1 )(2)2 =8+D
Reflecting JL and DL to the motor yields,
( 5+ 4 M ) ( 8+ D )
J m= ; D m=
4 4
Thus, the motor transfer function is
Kt 1
qm( s ) R a Jm Jm
=
Ea ( s ) 1 Kt Ka 1
s( s+ ( D m+ )) s( s+ ( D m +1) )
Jm Ra Jm

The gears are (10/20)(1) = 1/2. Thus, the forward-path transfer function is

( )
1
Jm 1
G e ( s )=(500 )
1 2
s ( s+ ( D m+1 ))
Jm

Finding the closed-loop transfer function yields,

Ge ( s ) 250/ J m
T ( s )= =
1+G e ( s ) D +1 250
s2+ m s+
Jm Jm
Dm+ 1
=4
For Ts = 2, Jm . For 20% overshoot,  = 0.456. Thus,
 Solutions to Design Problems 5-69

D m +1
2 zwn =2( 0 . 456 ) w n= =4
Jm

Or,
w n =4 . 386=
√ 250
J m ; from which J m=13 and hence,
D m=51 . But,

( 5+ 4 M ) ( 8+ D )
J m= ; D m=
4 4 . Thus, M = 11.75 and D = 196.

55.

a. The leftmost op-amp equation can be obtained by superposition. Let

v o =0 , then the circuit is
10 k
v 1=− v =−v in v =0 , the circuit is a non-inverting
an inverting amplifier thus 10 k in . Now let in

amplifier with an equal resistor voltage divider at its input, thus

10 k 10 k
v 1= (1+ )v =v
10 k +10 k 10 k o o . Adding both input components v 1=v o −v in

b. The two equations representing the system are:

v 1=v o −v in and
1 1
vo 0 .1 μs 0. 1 μR
=− =−
v in 1 1
R+ s+
0 .1 μs 0. 1 μR
The block diagram is:

-1 1
+ Vi + V
- 0.1 R o

1
s
0.1 R

c. From the figure

5-70 Chapter 5: Reduction of Multiple Subsystems

1
0 .1 μR
1 1
vo s+
0 .1 μR 0 . 1 μR
= =
v in 1 2
s+
0 . 1 μR 0 . 1 μR
1+
1
s+
0. 1 μR
4
T s= =0 . 2 μR=1m sec
2
d. The system is first order so 0 . 1 μR from which
1m
R= =5 k Ω
0.2μ
vo 2000
=
e.
vi s+ 4000 For a unit step input the output will look as follows

Step Response
0.5

0.45

0.4

0.35

0.3
Amplitude

0.25

0.2

0.15

0.1

0.05

0
0 0.5 1 1.5
Time (sec) -3
x 10
 Solutions to Design Problems 5-71

56.

a. The transfer function obtained in Problem 3.32 is

Y −520 s−10. 3844

= 3
U 1 s +2 . 6817 s 2 +0 . 11s +0 . 0126 by inspection we write the phase-variable form

[][ ][ ] [ ]
ẋ1 0 1 0 x1 0
ẋ 2 = 0 0 1 x 2 + 0 u1
ẋ 3 −0. 0126 −0 .11 −2. 6817 x 1
3

[]
x1
y= [−10 . 3844 −520 0 ] x2
x3
b.We renumber the phase-variable form state variables in reverse order

[ ][ ][ ] [ ]
ẋ 3 0 1 0 x3 0
ẋ 2 = 0 0 1 x2 + 0 u1
ẋ1 −0 .0126 −0 .11 −2. 6817 x 1
1

[]
x3
y= [−10 . 3844 −520 0 ] x2
x1

And we rearrange in ascending numerical order to obtain the controller canonical form:

[][ ][ ] [ ]
ẋ1 −2 . 6817 −0 . 11 −0 . 0126 x 1 1
ẋ 2 = 1 0 0 x 2 + 0 u1
ẋ 3 0 0 1 x3 0

[]
x1
y= [ 0 −520 −10 .3844 ] x 2
x3
c. To obtain the observer canonical form we rewrite the system’s transfer function as:

520 10 . 3844
− −
Y s2 s3
=
U1 2 . 6817 0 . 11 0 . 0126
1+ + 2 +
s s s3

We cross-multiply to obtain
5-72 Chapter 5: Reduction of Multiple Subsystems

[ −
520 10 . 3844
s2
−
s 3 ] [
U 1=Y 1+
2 .6817 0 .11 0 . 0126
s
+ 2 +
s s3 ]
Combining terms with like powers of integration:

1 1 1
Y = [ −2 . 6817 Y ] + 2 [ −520 R−0 . 11Y ] + 3 [ −10 .3844 R−0 . 0126 Y ]
s s s
¿
1
s [ 1
s ( 1
−2 .6817 Y + [ −520 R−0 . 11Y ] + [−10 . 3844 R−0. 0126 Y ]
s )]
We draw the signal flow graph:

The following equations follow:

ẋ 1=−2. 6817 x 1 +x 2
x 2  0.11x1  x3  520r
ẋ 3 =−0 . 0126 x 1 −10 .38 r
y=x 1

Which lead to observer canonical form:

 Solutions to Design Problems 5-73

[][ ][ ] [ ] []
ẋ1 −2. 6817 1 0 x1 1 x1
ẋ 2 = −0 . 11 0 1 x 2 + −520 u1 y= [ 1 0 0 ] x2
ẋ 3 −0. 0126 0 0 x
3
−10 . 38 x3
;

d.

>> A=[-0.04167 0 -0.0058; 0.0217 -0.24 0.0058; 0 100 -2.4];

>> B=[5.2;-5.2;0];

>> C=[0 0 1];

>> [V,D]=eig(A);

>> Bd=inv(V)*B

Bd =

1.0e+002 *

-0.9936 + 0.0371i

-0.9936 - 0.0371i

1.9797

>> Cd = C*V

Cd =

0.9963 0.9963 1.0000

>> D

D=
5-74 Chapter 5: Reduction of Multiple Subsystems

-0.0192 + 0.0658i 0 0

0 -0.0192 - 0.0658i 0

0 0 -2.6433

So a diagonalized version of the system is

[][ ][ ] [ ]
ẋ1 −0. 0192+ j0 . 0658 0 0 x1 −99 . 36+ j3 .71
ẋ 2 = 0 −0 . 0192− j0 . 0658 0 x2 + −99 . 36− j3 . 71 u1
ẋ 3 0 0 −2 . 6433 x 197 . 97
3

[]
x1
y= [ 0 . 9963 0 . 9963 1 ] x2
x3
 Solutions to Design Problems 5-75

57.
a. Substituting all values and transfer functions into the respective blocks of the system
(Figure 4), we get:

0.6154

Torque
Controller
Ref. & Power
Signal Amplifier
Rv(s)

+ Vehicle
Speed
V(s)

r 0. 3
= =0 .06154
i
Moving the last pick-off point to the left past the tot
4 .875 block and changing the position of the

back-emf feedback pick-off point, so that it becomes an outer loop, we obtain the block-diagram shown below. In
1
that diagram the 7.226⋅s block (representing the total inertia) has two parallel feedback blocks. Reducing these two

blocks into one, we have the following equivalent feedback transfer function:
1
Ω (s ) 7. 226 s 0 . 1384
Geq (s )= = =
T ( s) 0 . 13787 s+0 . 01908
1+
7 . 226 s
5-76 Chapter 5: Reduction of Multiple Subsystems

2
0.03787
Eb(s)  (s)
TL (s)
UC(s) Ia(s) T(s) _
Rv(s) Ev(s) _ V(s)
100 s  40 10 s  6 UA 1
+ 1
+ 1.8 0.06154
+ s s + _ 7.226 s
_ _
KCS Ia(s) Tf (s)
KSS (s)
0.1
0.5

0.0443

Replacing that feedback loop with its equivalent transfer function, Geq(s), we have:

2
Eb(s)
Geq(s)
UC(s) T(s)  (s)
Rv(s) Ev(s) _ Ia(s) V(s)
100 s  40 10 s  6 + 1 1.8
0.1384
0.06154
s + s  0.01908
+ _ s UA (s)
_
KCS Ia(s)
KSS(s)
0.5

0.0443

T (s )
I ( s ) and G (s) blocks, the above block-diagram
Moving the armature current pick-off point to the right past the a eq

becomes as shown below.

 Solutions to Design Problems 5-77

2
Eb(s)
UC(s) T(s)  (s)
Rv(s) Ev(s) + _ Ia(s) V(s)
100 s  40 10 s  6 0.1384
+ 1 1.8 0.06154
+ s s s  0.01908
_ _ UA (s)
KCS Ia(s)
KSS(s)
0.5 s  0.01908
0.2491

0.0443

The latter, in turn, can be reduced to that shown next as the cascaded blocks in the feedback to the torque controller
K CS I a ( s ) s +0 .01908
=
are replaced by the single block: Ω( s ) 0 . 4982 and the inner feedback loop is replaced by its

equivalent transfer function:

0.2491
Ω ( s) s+0.01908 0.2491
= =
U A ( s) 0.2491 s+0.5173
1+ ×2
s+0.01908

Rv(s) Ev(s) UA (s)  (s) V(s)

100 s  40 UC(s) 10 s  6 0.2491
0.06154
+ s + _ s s  0.5173
_
KCS Ia(s)
KSS (s) s  0.01908
0.4982

0.0443
5-78 Chapter 5: Reduction of Multiple Subsystems

Ω (s)
=
( s )( s+0.5173 )
10 s+6 0.2491
=
0.2491 (10 s+6)

1+ ( )( s+0.5173 )( 0.4982 )
U (s)
C 10 s+6 0.2491 s+0.01908 s ( s+0.5173) + 0.5 (10 s+6) ( s+0.01908 )
Thus: s

Ω (s )
=
( 100 s +40
s ) ( s (s +0 .5173 ) + 0 . 5 (10 s+6) ( s+0. 01908 ) )
0 . 2491 (10 s+6)

) ( s ( s+0. 5173 ) + 0 .5 (10 s+6 ) ( s+0 . 01908 ) )

R (s )
1+ 0 . 0443 (
v 100 s+40 0 .2491 (10 s+6 )
Finally
s or

Ω (s) 249.1 ( s+0.4 ) ( s+0.6 )

=
R v ( s) s ( 6 s 2 +3.613 s +0.0572 ) +11.035 ( s 2 +s+0.24 )
249.1 ( s+0.4 ) ( s+0.6 )
= 3 2
6 s +14 .644 s +11.09 s+2.65

V (s ) Ω (s ) 15 . 33 ( s+0 . 4 ) ( s+0 . 6 )
=0 . 06154 =
Hence:
R v (s ) R v ( s ) 6 s +14 . 644 s 2 +11. 09 s +2 .65
3

b. Simulink was used to model the HEV cascade control system . That model is shown below. The reference
signal, rv (t), was set as a step input with a zero initial value, a step time = 0 seconds, and a final value
equal to 4 volts [corresponding to the desired final car speed, v ( ) = 60 km/h, e.g. a desired final value of
the change in car speed, v ( ) = 5.55 m/s]. The variables of interest [time, change in car speed,
acceleration, and motor armature current] were output (in array format) to four “workspace” sinks, each of
which was assigned the respective variable name. After the simulation ended, Matlab plot commands were
utilized to obtain and edit the required three graphs. These graphs are shown below.

The simulations show that in response to such a speed reference command, car acceleration would go
initially to a maximum value of 10.22 m/s 2 and the motor armature current would reach a maximum value
of 666.7 A. That would require an electric motor drive rated around 80 kW or using both the electric motor
and gas or diesel engine, when fast acceleration is required. Most practical HEV control systems, however,
use current-limiting and acceleration-limiting devices or software programs.
 Solutions to Design Problems 5-79

time
Clock
To Workspace1

0.61543

Speed vs Time
Armature Aerodynamic Drag
Res., Ra
1
PI(s) PI(s) 1 1.8 0.06154 du/dt
7.226s
Referece Derivative Acceleration
Speed Torque Torque Const. Motive HEV Inertia
Signal Linear vs Time
Controller Controller x Efficiency minus Referred to
Speed /
and Power Resistive Motor Shaft
Angular
Amplifier Torques Speed
0.1

Armature
Back EMF Friction
Current vs Time
Const, Kb1 Coeff., D

2 acceleration

To Workspace2

0.5
Speed

Current Sensor To Workspace

Sensitivity, Kcs
0.0443 Current

To Workspace3
Speed Sensor
Sensitivity, Kss

Model of the HEV Cascade Control System

Change in car speed in response to a speed reference signal step of 4 volts

5-80 Chapter 5: Reduction of Multiple Subsystems

Car acceleration reponse to a speed reference signal step of 4 volts

Motor armature current reponse to a speed reference signal step of 4 volts

 Solutions to Design Problems 5-81

58.

a. There is only 1 forward path. The gain of the forward path is M 1=GP . There are
~
two loops L1=−GP and L2=+G P , and no non-touching loop; so
~
∆=1+GP−G P . After the forward loop is eliminated ∆ 1=1. The closed-loop
transfer function from command input to output is:
Y M ∆ GP
( s )= 1 1 =
R ∆ 1+G (P−~
P)
b. There is only 1 forward path. The gain of the forward path is M 1=1 There are two
~ ~
loops L1=−GP and L2=+G P , and no non-touching loop; so ∆=1+GP−G P .
~
After the forward loop is eliminated one loop remains so ∆ 1=1−G P . The
closed-loop transfer function from command input to output is:
M ∆ ~
Y 1−G P
( s )= 1 1 =
D ∆ 1+G( P−~P)
c. The total output is
~
GP 1−G P
Y= ~ R+ ~ D
1+G(P− P) 1+G(P− P)
G(s)
d. In Figure P5.54(b) let G C ( s )= ~ . The closed-loop transfer function
1−G ( s ) P (s )
from command input to system output is:
G (s )
~ P(s)
Y 1−G ( s ) P ( s ) G( s) P(s)
= = ~
R G ( s) 1+G(s)( P(s)− P (s ))
1+ ~ P( s)
1−G ( s ) P ( s )
Similarly the transfer function from disturbance to system outputs is:

~
Y 1 1−G( s) P (s)
= = ~
D G (s ) 1+G(s)(P(s)− P (s))
1+ ~ P (s )
1−G ( s ) P ( s )