State Variable Analysis

State equations Optimal Transfer State equations Damping ratio

of a control system function of the of the system of the system
mechanical based on system from from the by minimizing
system quadratic the state transfer the quadratic
performance equations function performance
index index

Discrete time Equivalent Controllability, State equations State diagram

state equation electrical observability, of the system of the system
from the network and and stability from the state from the state
continuous state equations from the state diagram equations
time state for a equations
equation mechanical
system

Analogue Damping ratio Transfer Solving the Design of a

simulation of of the system function of the state equation state variable
the system by minimizing system from feedback
from the the quadratic the state controller
transfer performance equations
function index

State variable Design of the State equations Transfer Design of a

model of a hoist observer for of a d.c shunt function of the state variable
given observer motor system from feedback
poles the state controller
diagram

Transfer The state Objective

function of the transition type
system from matrix
questions:
the state
diagram

WEBSITE takes you to start page after you have read this chapter.
Start page has links to other chapters.
Problem: 7.1 State equations of a mechanical system

(a) Fig.1 shows a mechanical system, where

F (t) =Force applied to the system

M= Mass
B= Coefficient of friction
K= Spring constant
y= Displacement of the mass.

Write the state equations of the system, choosing displacement and velocity of the mass as
state variables, x1 and x2, respectively.

Y (t)

K M
F (t)

Fig.1

(b) Write the difference equation obeyed by a sum of money in a bank account at a yearly
interest rate of R percent, compounded every T years. Denote the amount immediately
following the kth interest period as A (kT). Find the differential equation which results as
T 0.

Solution

(a) Force absorbed by mass, fM= Md2y/dt2

Force absorbed by friction, fB= Bdy/dt

Force absorbed by spring, fs= Ky

fM+ fB + fs = F
Let x1 = y, x2 = dy/dt

The state equations are:

dx1/dt = x2

dx2/dt = -Kx1/M –Bx2/M +F (t)/M

(b) The principal obeys the difference equation

A (kT+T) =A (kT) [1+RT/100].

Re-writing, [A (kT+T)-A (kT)]/T = RA (kT)/100.

Taking the limit as T tends to 0 and letting kT tend to t gives the differential

equation

dA/dt =RA (t)/100.

TOP
Problem: 7.2 Optimal control system based on quadratic performance index

(i) A control system is described by

dx/dt = Ax + Bu

where x = x1, A = 0 1 B= 0
x2 0 0 1

Assume the linear control law:

U =-Ktx = -k1x1-k2x2.

Note that Kt = -k1 -k2 .

Determine the constants k1and k2 so that the following performance index is minimized:

J = xtx dt
0
The initial condition is x(0) = 2
0

Choose the undamped natural frequency to be 2 rad/s.

(ii) Will the optimal control law of Part (i) for the system change if the initial conditions are
changed?

Solution

dx/dt =Ax-BKtx = 0 1 x1
-k1 -k2 x2 (1)

A-BKt = 0 1
-k1 -k2

Eliminating x2 from eqn (1) yields

d2x1/dt2 + k2dx1/dt + k1x1 = 0 . Since undamped natural frequency = 2, we obtain, k1=4.

Therefore A –BKt = 0 1
-4 -k2

A-BKt is a stable matrix if k2 >0.

J = xt (0) Px (0) is minimized where

P= (A-BKt )t P +P (A-BK) = -1

Or,
0 -4 p11 p12 + p11 p12 0 1 = -1 0
1 -k2 p12 p22 p12 p22 -4 - k2 0 -1

Therefore,

p11 p12 = (5/2k2) + (k2)/8 1/8

p12 p22 1/8 5/ (8k2)

J = xt (0) P x (0) = sqrt (2) 0 p11 p12 sqrt (2)

p12 p22 0

= ( 5/k2 ) + ( k2/4 )

To minimize J, differentiate J w.r.t k2 and set it = 0.

dJ/dk2 = (-5/k22) + (1/4)

Therefore , K2 = sqrt (20)

Jmin = sqrt (5).

The designed system has the control law

U = -4x1- sqrt (20). x2

TOP

Problem : 7.3 Transfer function of the system from the state equations

(a)

(i) Write the following differential equation

d3 y/dt3 +d2y/dt2 + 3 dy/dt + 2y = f(t)

in the state variable form by choosing the state variables:

x1 = y, x2 = dy/dt, and x3 = d2 y/dt2

and the input variable as f (t).

(ii) Is the obtained state variable form in Part (i) a unique description of the differential

equation?

(b) A control system is represented by the following state variable equations:

dx/dt = Ax + Bu
y = Cx

where x = x1 , A= 0 1 0
x2 0 0 1
x3 0 -4 -5

B= 1 , C= 1 0 0
-1
5

(i) Determine the transfer function of the system.

(ii) Determine the eigenvalues of the A matrix.

Solution

(a)

(i) d3y/dt3 +d2y/dt2 + 3 dy/dt + 2y = f (t)

dx1/dt = 0 1 0 x1 0 .f (t)
dx2/dt = 0 0 1 x2 + 0
dx3/dt = 1 3 2 x3 1

(ii) No

(b) (i)

G(s) = C (sI-A)-1 B

(sI –A) = s -1 0
0 s -1
0 4 s+5

G(s) = (s2+ 4s + 4) / (s3 + 5s2 + 4s)

(ii) eigenvalues are 0,-1 and -4.

TOP

Problem : 7.4 State equations of the system from the transfer function

(a) A system is described by the following equations:

dx/dt =Ax + Bu
y =Cx + Du
where

A = -1 1 B= 1 0 1
0 -2 0 1 1

C= 1 2
1 0
1 1

In addition, D is a null matrix of proper order. Find the transfer function matrix,
[Y(s)/U(s)].

(b) The overall transfer function of a single input single output control system is given by

Y(s)/U(s) = ( s2+4s+4)/(s3+5s2+4s)
i Find the A,B and C matrices in the state variable description of the system

ii Is the description of Part (i) above unique?

Solution

(a) (sI-A)-1 = s+1 -1 -1

= 1/(s+1) 1/(s+1) (s+2)
0 s+2 0 1/(s+2)

Transfer function, C (sI-A)-1B+D = 1/(s+1) (2s+3)/(s+1) (s+2) (3s+5)/(s+1) (s+2)

1/(s+1) 1/(s+1) (s+2) (s+3)/(s+1) (s+2)
1/(s+1) 1/(s+1) 2/(s+1)

(b)

Y(s)/U(s) = [(s+2)2]/[s(s+1) (s+4)] = (1/s) - [(1/3)/(s+1)] + [(1/3)/(s+4)]

This may be represented in the block diagram form as follows:

X1(s)
1/s 1

U(s) X2(s) Y(s)

1/s -1/3

-1

1/s X3(s)
1/3

-4

Choosing the state variables as shown above, we write

dx1/dt = u
dx2/dt = x2+u
dx3/dt = - 4x3 +u

and y = x1-x2/3+x3/3

Therefore, A= 0 0 0 B= 1 C = 1 -1/3 1/3

0 -1 0 1
0 0 -4 1
TOP

Problem : 7.5 Damping ratio of the system by minimizing the quadratic performance
index

The control system of Fig.2 is assumed to be at rest initially.

1
u + x c
s (s + )
-

Fig.2

Determine the damping ratio 0 so that, for unit-step input u (t), the following performance
index is minimized:

J = xT (t) Q x (t) dt
0
where

x1 x 1 0
x = = , Q = ( 0)
x2 dx /dt 0

What is the optimal value of the damping ratio if =1?

Solution

d2c/dt2 + dc/dt + c = u

Noting that

x= u –c, u = 1, the initial conditions are equal to zero, we have

d2x/dt2 + dx/dt +x =0 (t 0)

The state-space representation of this eqn. becomes

dx1 0 1 x1 0
dt
= , u

dx2 -1 - x2 1
dt

x1 (0) 1

= or, dx/dt =Ax

x2 (0) 0

where A = 0 1

, J = xt (0) Px (0), where P is determined from

-1 -

At P + PA = -Q

Or, 0 -1 p11 p12 p11 p12 0 1 -1 0

+ =
1 - p12 p22 p12 p22 -1 - 0 -

or,

-2p12 = -1

-p22 +p11- p12=0

2p12-2 p22=-

Therefore, P = ( /2) + 1+ 0.5

2

0.5 1+
2

J = xt (0) Px (0) = p11 = ( /2) + 1+

2
To minimize J, dJ/d = 0.5 – (1+ ) = 0 giving =sqrt (1+ )
2 2

If =1, the optimal value of = sqrt (2).

TOP

Problem : 7.6 Discrete time state equation from the continuous time state equation

(a) Obtain the solution of the state equation

dx1 1 1 x1 1
dt
= + u

dx2 0 1 x2 1
dt

where u (t) is a unit step. The initial conditions are xT (0) = [0 1].

( Note that ‘ T ‘ is the transpose of the vector x )

(c) The state equation for a continuous-time plant is

(d)

dx1 0 1 x1 0
dt
= + u

dx2 0 -1 x2 1
dt

Obtain the equivalent discrete–time equation, suitable for a time step, T , of 1 sec.

Solution

(a)

x1 et + tet -1

= 2 Answer

x2 et - 0.5
(b) The solution of the equation is

t
X (t) = exp (A (t-t0)) x (t0) + exp (A (t- )) Bu ( ) d
To

where A and B are known.

Setting t0 = kT and t = (k+1) T and assuming u (t) = U(kT) for kT t (k+1)T, we get
=

x (k+1) = x1 et +tet -1

=2 Answer

x2 et - 0.5

(T) x (k) + (t) u (k)

where (T) = exp (AT), (T) = exp (A ) Bd

0

For given A,

1 1- e-t
exp (At) = Setting T= 1 sec. we get
-t
0 e

(T) =
1 0.632
exp (At) = Setting T= 1 sec. we get

0 0.368

(T)= exp (A ) Bd
t
(1-exp (- ) d
0
t
= exp (- ) -d
0

Setting T= 1 sec. we get

0.368
=
0.632

Therefore,

1 0.632 x1 (k) 0.368

x (k+1) = + u (k) Answer

0 0.368 x2 (k) 0.632

TOP

Problem : 7.7 Equivalent electrical network and state equations for a mechanical
system

The suspension system for an automobile is designed to increase the comfort of the
passengers by absorbing the vibrations due to the terrain of the road. Fig.3 shows
the model of such a system. Here M represents the masses, B represents the
dashpot damping coefficients, and K represents the spring constants respectively.

M2

M1
u2 (t) = force due to winds

Passenger area
x2

B2 K2

Lower body
x1

B1 K1

Wheels

Road

u1 (t) = force due to road profile Fig.3

(i) Obtain an equivalent electrical network

(ii) Write a set of state equations for the system
(iii) Describe the procedure for determining the transfer function matrix relating the states
x1 and x2 to the inputs u1 and u2 respectively.

Solution

(a) dx1/dt B2 dx2/dt

K2

u1 B1 K1 M1 M2 u2

(b)

The equilibrium equations are:

M1 d2x1/dt2 + B1 dx1/dt +K1x1+B2 (dx1/dt-dx2/dt) +K2 (x1-x2)= u1 (t)

M2 d2x2/dt2 + B2 (dx2/dt –dx1/dt) +K2 (x2-x1)= u1 (t)

We need four state variables, x1, dx1/dt, x2, and dx2/dt. The state eqns. are:

dx/dt=Ax+Bu

where A= 0 1 0 0
-(K1+K2)/M1 -(B1+B2)/M K2/M1 B2/M1
0 0 0 1
K2/M2 B2/M2 -K2/M2 -B2/M2

B= 0 0 dx/dt = x1 u= u1
1/M1 0 dx1/dt u2
0 0 x2
0 1/M2 dx2/dt

(c) Transfer function Matrix:

(d)
-1
G(s) = 1 0 0 0 s -1 0 0 0 0

0 0 1 0 (K1+K2)/M1 s+(B1+B2)/M1 -K2/M1 -B2/M1 I/M1 0

0 0 0 -1 0 0
-K2/M2 -B2/M2 K2/M2 s+B2/M2 0
1/M2
TOP

Problem : 7.8 Controllability, observability, and stability from the state equations

The state equations of a system are as follows:

0 0 -8 2
dx/dt = 1 0 -14 x + 1 u
0 1 -7 0

c = (0 0 1) x

Determine if the system is

i. controllable
ii. observable, and
iii. stable

Solution

(i) Condition of controllability: Rank of U= (B, AB, A2B) = 3

Condition of controllability: Rank of U = (B, AB, A2B) = 2

U= 2 0 -8 Rank U = 2, Uncontrollable
1 2 -14
0 1 -5

(ii) Condition of observability: V is non-singular, and Rank of V=(C, CA, CA2)=3

V= 0 0 1 V is non –singular, Rank V =3, Observable

0 1 -7
1 -7 35

(ii) Characteristic eqn: s3 + 7s2 +14s +8

S3 1 14
S2 7 8
s 90/7
S0 8

Stable
TOP

Problem : 7.9 State equations of the system from the state diagram

(a) Fig.4 shows a feedback control system. The open –loop transfer function is

G(s) =
2(s+1)(s+3)
s(s+2)(s+4)
What is the closed-loop transfer function of the system?

G(s)
U(s) + C(s)

Fig.4

(b) Fig.5 shows the state diagram of the system of Part (a), The state variables are used to
provide the output signal.

b
U(s) 1 1/s 1/s 1/s C(s)

x3 x2 x1 d g
e
f

Fig.5
Write the state variable equations of the system in the form:

dx (t)/dt = A x (t) + Bu (t)

c (t) = D x (t)

What are the values of transmittances a, b, d, e, f and g shown in Fig.3?

Solution

Closed loop transfer function is

T(s) =(2s2+8s+6)/(s3+8s2+16s+6)

The vector differential equation for the state diagram is

Dx/dt = 0 1 0 0
0 0 1 + 0 u(t)
-6 -16 -8 1

and the output is

c (t) =(6, 8, 2) x1 a = 2, b = 8, d = 6, e = -8, f = -16, g = -6

x2
x3

TOP

Problem : 7.10 State diagram of the system from the state equations

(a) The state equation of a linear dynamic system is

dx1/dt = 0 1 x1 (t) 0
+ r (t)
dx2/dt = -2 -3 x2 (t) 1

where x1 (t), x2 (t) are the system states and r (t) is the reference input. Determine the system
states for t 0 when the input r (t) =1 for t 0; that is,
r (t) =u (t). The initial state of the system is

x1 (0) = 1

x2 (0) = 1

(b) Draw the state diagram for the system described in Part (a) by its state equation and initial
condition. Assign the state variables as the outputs of the integrators and the reference input and
the initial conditions as the system inputs.

Solution

(a)

A= 0 1 B= 0 x (0) = 1
-2 -3 1 1

sI-A = s -1
2 s+3

(sI –A)-1 = (1/(s2+3s +2) s+3 1

-2 s

(t) = L-1 [(sI-A)-1]= 2e-t-e-2t e-t-e-2t

-2e-t +2e-2t -e-t+2e-2t
t
x (t) = (t) x (0) + (t- ) Br ( ) d
0

x (t) =

2e-t-e-2t e-t-e-2t x (0) + 0.5 -e-t-0.5e-2t

-2e-t +2e-2t -t
-e +2e -2t
e-t -e-2t

= 0.5 +2e-t –2.5 e-2t

-2e-t +3 e-2t

(b) x 1 (0)/s x2(0)/s

r 1 dx2/dt dx1/dt = x2 x1
1/s 1/s

TOP

Problem : 7.11 Analogue simulation of the system from the transfer function

(a) A robot-arm drive system for one joint can be represented by the differential equation:

dv (t)/dt=-k1v (t)-k2y (t)+k3i (t)

where v (t) =velocity, y (t) =position, and i (t) is the control-motor current. Write the second-

order drive system in the state-variable form, when k1=k2=k3=1.

(b) Draw the block diagram for analogue computer simulation of a system having the following
transfer function:
Y(s)/U(s) = 1/(s3+7s2+14s+8).

Choose the state variables as follows:

x1 = y (t), x2= dx1/dt, x3= dx2/dt.

Solution
(a) Let x1 = y, x2= dy/dt=v, u= i. Then

dx1/dt = x2

dx2/dt = -k1x2 -k2x1 +k3 u = -x2 -x1 +u

(b) The transfer function corresponds to the differential equation:

d3 y/dt2 +7d2y/dt2+14dy/dt+8y = u which can be written as:

dx3/dt = -7x3-14x2-8x1+u and combining this with the state variable description, we get
dx1/dt = x2
dx2/dt = x3
y=x1

The state equations lead to the following simple analogue computer simulation.
+u integrator
x3 integrator
x2 integrator
x1= y

7 14 8

+ +

+ +

TOP

Problem : 7.12 Damping ratio of the system by minimizing the quadratic performance
index

Determine the value of the damping ratio 0 so that when the system shown in Fig.6 is
subjected to a unit step input, the following performance index is minimized:

J= xT (t) Q x (t) .dt

0

where

x = x1 = e Q= 1 0
x2 de/dt , 0 2
Assume the system to be at rest initially.

r(t) e (t) c (t)

K
+ - s(s + 3 )

Fig.6

Solution

Eqn. for the system:

d2c/dt2 +3 dc/dt+c=r

Since e= r-c, and initial conditions are zero, we have

d2x/dt2 +3 dx/dt + x=r

From the state –space representation of this eqn. we get

A= 0 1
-1 -3

Note that, since e= r-c, and applied input is a step function, the initial conditions for the state

variables are:

x1 (0) = 1
x2 (0) = 0

The performance index is:

J = xT (0) P x (0)

P is determined from

ATP+PA= -Q
(Elements of symmetric P matrix are p11, p12, and p22)
Solving the above eqn. we get

p11 = (1/2 ) +3 /2

p12 = ½

p22 = 1/2

Therefore J = (1/2 )+3 /2

To minimize J, obtain dJ/d = 0. This gives = 0.577

TOP

Problem : 7.13 Transfer function of the system from the state equations

(a) An integrator whose transfer function is 1/s has a negative feedback path with an adjustable
gain K.
(i) Demonstrate the manner in which the gain K affects the closed-loop transfer function and the
response of the system, when it is varied from zero to plus and minus infinity.
(ii) Draw the locus of the closed loop pole in the complex s- plane.
(b) Obtain the transfer function Y(s)/U(s)of the system shown in Fig.7 where x1 and x2 are the
state variables.
-5

+
+ x1
2 dt

+
u + y
3

+
-1

+
dt 2
5
+
+
-1 x2

Fig.7

Solution

(a)

The closed loop transfer function is T(s) =1/(s+K). Pole is at s = -K. Addition of the feedback
path has caused the transfer function pole to shift from the origin to the left in the s-plane as K
is increased from some positive value to . While increasing K, the system has been changed from
an integrator to a first-order low –pass filter whose half-power bandwidth is K rad/sec. For
negative K , system will be unstable. Locus of closed loop poles in the complex plane as K is
varied is sown below.
K= 40 K= 30 K= 20 K= 10 K= 0

-40 -30 -20 -10 0

(b)

Write dx1/dt = -5x1 — x2 + 2u

dx2/dt = 3x1-x2 + 5u

y = x1+ 2x2

Transfer function G(s) = C(sI-A)-1B

-1
= [1 2] s+5 1 2 = (12s+59)/(s+2) (s +4)
-3 s+1 5

TOP

Problem : 7.14 Solving the state equation

(a) Show how to reduce the following differential equation

.
d3x/dt3 + a1.d2x/dt2 + a2 .dx/dt + a3 .x = u

into the following set of first-order differential equations:

dx1/dt 0 1 0 x1 0
dx2/dt = 0 0 1 x2 + 0 u
dx3/dt -a3 -a2 -a1 x3 1

(b) A system is defined by the state equations:

dx1/dt =-2x1 +x2

dx2/dt = x1-2x2 +u

and the output equations:

y1 = 3x1+x2

y2 = x1+2x2

y3 = x2

The initial conditions are:

x1(0) =1, x2(0) = 2.

Find the time solution of the output variables when the input u =1 for t 0 and 0 for t 0.

Solution

(a)

d3x/dt3+ a1.d2x/dt2 + a2 .dx/dt +a3 .x=u

Choose x1= x, x2 = dx/dt, x3 = d2x/dt2 so that dx3 /dt = d3x/dt3

Then the differential equation may be written as

dx1/dt = x2
dx2/dt = x3
dx3/dt =- a3x1- a2x2- a1x3 + u.

These first order equations are then written in the matrix-vector form.

(b)

If a = e-t and b = e-3t, then the solution for the states is

X1 = (1/3) +a – (b/3)

X2=(2/3) +a + (b/3) and the outputs are:

Y1=(5/3) +4a – (2b/3)

Y2=(5/3) +3a + (b/3)

Y3=(2/3) +a + (b/3)

TOP

Problem : 7.15 Design of a state variable feedback controller

(a) Describe a system involving feedback that arises in the following items:

i. Automobiles
ii. Economics

Attempt to identify the plant or process, the input (both reference and disturbance), and the
feedback path(s).

(b) A multi-input /multi-output plant with a state-variable feedback controller is shown in Fig.8.

W U X Y
B 1/s C

Controller

Fig.8
0 1 0 0
A= 0 0 1 B= 0
0 -30 -11 1

C= 1 0 0

Find the feedback matrix K such that the poles are at s= -2,-5, -6.

Solution

(a)

Category Feedback Plant Reference Disturbance Feedback

system input input path
Economics Supply of Production Desired Cost of raw Prices
goods & profit material
distribution
Automobile Automatic Carburetor Fuel/air ratio Engine Heat sensor
choke versus temperature on choke
engine
temperature

TOP

Problem : 7.16 State variable model of a hoist

A hoist may be thought of as a motor lifting a mass in the presence of gravity, as shown in Fig.9
Kb
R L Jdrum
KT
I JM D
V

Load
M

g
Fig.9

Write the appropriate equations and develop a state variable model. Kb is the back emf constant
and KT is the torque constant. JM and Jdrum are the moments of inertia of the motor and drum
respectively.

Solution

V= IR +LdI/dt +Kb dc/dt

KT I =(JM + J drum) d2c/dt2 +MgD/2

Let J= JM +Jdrum

Choose x1=c,x2= dc/dt, x3= I.

Substitute & manipulate:

dx/dt= 0 1 0 x1 - 0 MgD/2J + 0 V
0 0 KT/J x2 1 0
0 -Kb/L -R/L x3 0 1/L

TOP
Problem : 7.17 Design of the observer for given observer poles

( a) State the conditions for complete observability in terms of transfer functions or transfer
matrices.

(b)For a unity feedback closed loop system having an open loop transfer function
2
G(s) = 100/(s(s +0.1s +10)),

write the state equations and design an observer for observer poles at - 50, -10-j10, -10+ j10.

Solution

dx1/dt = x2

dx2/dt = 1x2+x3

dx3/dt = 10x2 +100 (R-x1)

K0 K1 K2

1/s 1/s 1/s

.1

100
s (s2 +.1s +10)
PLANT

Observer characteristic equation:

s3 + ( .1 +10K2)s2 +(10 +100K1 +K2)s+100K0 +100K2=0

For observer poles s=-50,-10+ j10,-10-j10,

s3 +70s2+1200s +10000=0

Equating coefficients, K2=6.99, K1=19.83, K0=80.17

TOP

Problem : 7.18 State equations of a d.c shunt motor

A d.c shunt motor, Fig.10, drives an inertia load. The shaft connecting motor to loads is not
rigid and must be considered as spring. The applied voltage is V.
R=armature resistance
L=armature inductance
Kb=back emf constant
KT= Torque constant
Ks=spring constant
JM=rotor moment of inertia
JL=load moment of inertia
I =armature current
M=rotor angle
L=load angle, Write the state equations of the control system using M, L, I, d M/dt, d L/dt as
the state variables.

I
V R L JM Ks JL

M L
FIG. 10

Solution

V= IR +L dI/dt + Kb d M/dt

KTI = JM d M/dt + Ks ( M- L)

0= JL d L/dt - Ks ( M- L)
x1 = L

x2 = d L/dt

x3 = M

x4 = d M/dt

x5 =I

0 1 0 0 0
0 0 - Ks/ JL Ks/ JL 0
A= 0 0 0 1 0
Ks/ JM 0 Ks/ JM 0 KT/ JM
0 0 0 Kb/L R/L

BT= ( 0 0 0 0 Kb/L )

TOP

Problem : 7.19 Transfer function of the system from the state diagram

(a) Write down the state and output equations of the system shown in Fig.11.
(b) Obtain its transfer function, C(s)/R(s)

- 10

+ dt
4 +

+
r (t) 6 + c (t)

-2
+
+ +
10 4
+
dt
-2

Fig.11
Solution
:
dX1/dt=-10x1-2x2+44

dX2/dt=-6x1-2x2+104

y=x1+4 x2

A= -10, -2
6, -2 B= 4
10 C= 1,4

Transfer function = C (sI-A) –1B = 11(s+11)/[(s+1)(s+2)]

TOP

Problem : 7.20 Design of a state variable feedback controller

(a) Discuss the realization of a feedback controller for pole allocation.

(b) Fig.12 shows a single-input single-output second-order system with state-variable
feedback controller. The system equations are:
(c)
dx/dt = A x + b u
y = Ct x
where

A = 1 1 , b= 0 ,
2 0 1

Ct = [1 0]
The plant input, u and the system input, w are related by

U = w + Kx .

.
w u + dx/dt x y
b 1/s
Ct
+ +

Fig.12

Find the feedback matrix K such that the poles are at

s = -1,-2.

Solution

A= 1 1 B= 0
2 0 1

C= 1 0

T.F = Y(s)/U(s) = C(sI-A) -1 B= 1/(s-1)(s-2) with poles at 1, 2.

Desired characteristic equation (s+1) (s+2)= s2+3s+2=0

Feedback matrix K= (K1,K2)

Anew = A + BK = 1 1
2+K1 K2

det (sI – Anew) = 0 . This gives s2-(K2+1) s-2K1=0. Compare with desired equation and obtain
K1=0, K2=-4

TOP

Problem : 7.21 Transfer function of the system from the state diagram

(a) Write the state and output equations:

dx/dt = Ax+Bu
y = Cx+Du

of the system shown in Fig.13.

1/8

+
+ x1 +
y1 (t)
u1(t) + +
-3/8 1/8 2 4

+ +
x2
+ + y2 (t)
u2(t) +

5/8

Fig.13 - Symbol for Integrator

(b) -If the state space description of the system of Fig.1 is given by

A= 1/8 -3/8 B= 1 0
1/8 5/8 0 1

C= 1 4
2 1

D = a null matrix of proper order , find the transfer function matrix relating the input vector
with the output vector.

Solution

(a) dx1/dt = x1/8-3x2/8 +u1

dx2/dt = x1/8+5x2/8 +u2
y1= x1+4x2
y2= 2x1+x2

(b) B=Unit matrix, D= null matrix

G(s) =C(sI-A) –1B+D

-1
= 1 4 s 0 1/8 -3/8
2 1 0 s - 1/8 5/8

= a/e, b/e
c/e, d/e

where a= 8s-1,b=32s-1,c=16s-9,d=8s-7,e=8s2-6s+1

TOP

Problem : 7.22 The state transition matrix

(a) Find the inverse of matrix A.

5 0 2
A = 0 2 -3
2 1 2
(b) Find the characteristic polynomial and the eigenvalues of matrix B.

0 1 0
B= 0 0 1
-4 -6 -4

(c) Find the state-transition matrix of the system described by the following state
equations:

dx1/dt = x2
dx2/dt = -2x1-3x2

Solution

(a)
Inverse =(1/27) 7 2 -4
-6 6 15
-4 -5 10

(b) Eigenvalues are : 2, -1+j, -1-j

(c) State Transition matrix: 2et-e-2t et-e2t

-2et+2e2t -e-t+2e2t

TOP

Objective type questions:

(i) Which one of the following statements is true?

(a) State space analysis is not applicable to linear time invariant systems
(b) Conventional root-locus and frequency response methods are applicable to the design
of optimal and adaptive control systems that are mostly time varying and/or nonlinear.
( c) State space analysis is a powerful method for time-varying systems, nonlinear
systems, and multiple- input-multiple output systems
(d)Frequency response methods are a powerful method for time-varying systems,
nonlinear systems, and multiple- input-multiple output systems
Ans: ( c)

(ii) Fig. 6 shows a control system defined by the following equations:

R(s) + E(s) Process C(s)

-

Fig.6

C(s)/R(s) =1/[s2+2ξ s +1] (ξ 1)

E(s) =Ł [e(t)] = R(s) – C(s)

where ξ is the damping ratio. The system is initially at rest and is subjected to
a unit –step input. The performance index of the system

t │e(t)│dt
0
is

(a) 2 ξ
(b) 4 ξ 2 -1 +2ξ
(c) 4 ξ 2 -1
(d) ξ

Ans: ( c )

(iii) The continuous time system described by

dx1/dt =x1
dx2/dt =-x1 +u
y = x1

(a) is completely state controllable but not completely observable

(b) is completely state controllable and completely observable
(c ) is completely observable but not completely state controllable
(d)is not completely state controllable and not observable.

Ans: ( b )

(iv ) For the third order plant which of the following may be a set of state variables?

(a) x(t), dx(t)/dt, d2x(t)/dt2

(b) x(t), dx(t)/dt, x(t) + d2x(t)/dt2
(c) x(t), d2x(t)/dt2, x(t) + d2x(t)/dt2
(d) any of the above

Ans: ( a )

(v) Which one of the following minimum information is necessary to formulate the
problem of control systems optimization?

(a) System state equation, output equation and the control vector only
(b) System state equation, output equation, control vector, and the performance index only
(c) System state equation, output equation, control vector, Constraints of the problem, the
performance index, and system parameters

Ans: ( c )

(vi) Which one of the following characteristics are the attributes of an ideal relay?
(a)
OUTPUT

+
INPUT
-

(b)
OUTPUT

+
INPUT
-

(c)
OUTPUT

+
INPUT

Ans: (a)

(vii) In a series RLC circuit containing a voltage source v (t), the current i and the voltage
vc across the capacitor are chosen as the state variables x1 and x2 respectively. Show that
the state equations are:

(a) dx1 R 1 x1 1
dt = L L + L v (t)

dx2 1 0 x2 0
dt C
(b) dx1 -R -1 x1 1
dt = L L + L v (t)

dx2 1 0 x2 0
dt C

(c) dx1 -R -1 x1 1
dt = L L + C v (t)

dx2 1 0 x2 0
dt L

Ans: ( b )

(viii) In the RLC circuit of the above question, part (v), show that the characteristic
equation of the circuit is

(a) RCs2+LCs +1 =0

(b) s2+RCs +(1/LC) =0

(c) LCs2+RCs +1 =0

Ans: ( c )

(ix )

(a) The following system is not completely state controllable:

dx1/dt -1 0 x1 2 u
dx2/dt = 0 -2 x2 + 5

(b) The following system is not completely state controllable:

dx1/dt -1 0 x1 2 u
dx2/dt = 0 -2 x2 + 0
(c) The following system is not completely observable:

dx1/dt 1 1 x1 0 u
dx2/dt = -2 -1 x2 + 1

y = (1 0) x1
x2

Ans: ( b )

( x ) With the knowledge of the state variable description of a plant, the transfer function of the plant

(a) Cannot be determined

(b) Can be determined completely
(c) There is no relationship between the state variables and the transfer function

Ans: (b )

( xi ) Which one of the following statements is true?

( a ) A scalar function V (x) is said to be positive definite in a region (which
includes the origin of the state space) if V (x) 0 for all nonzero states x in the region
and V (0) = 0
( b ) A scalar function V (x) is said to be negative definite if minus of V (x) is negative
definite
( c )A scalar function V (x) is said to be positive semi definite if it is positive
at all states in the region except at the origin and at certain other states,
where it is zero
Ans: ( c )

( xii) Which of the following is the A matrix of the system?

(a) 0 1
-5 -4

(b) 0 1
5 4

(c) -5 -4
0 1

Ans: ( a)
( xiii ) In a control system

(a) The state variable can always be measured

(b) The output variable cannot be defined as a function of the state variables
(c) The output is a variable that can be measured

Ans: (c )

( xiv ) The state transition matrix (t) possesses the following properties:

(a) [ (t)] k = (kt) for k = integer

(b) -1(t) is not equal to (-t)
(c) (t2-t1) (t1-t0) is not equal to (t2-t0) for any t0,t1,t2

Ans: ( a )

( xv) Which one of the following statements is true?

(a) The frequency domain techniques are applicable to non-linear and time-varying systems.
(b) The frequency domain techniques are particularly useful for multivariable control systems
(c) Time- domain methods cannot be utilized to design an optimum control system
(d) Time- domain techniques can be utilized for non-linear, time-varying, and multivariable
systems.

Ans:( d )

(xvi) The state-variable equations of a single-input single-output system are:

dx/dt=Ax + Bu
y=Ctx

where A is of order (nxn), and B and C are each of order (nx1).

Which one of the following describes the transfer function, G(s)= y(s)/u(s), for this system?

(a) (s I-A) -1 BCt

(b) Ct B (sI-A) –1
(c) C (sI-A) –1 B

where I is an (nxn) identity matrix.

Ans: (c)

(xvii) The transfer function

(a) can be uniquely determined from the state variable equations and the state
variable equations can also be uniquely determined from the transfer function.

(b) for a given set of state variable equations is not unique, but the state variable
equations can be uniquely determined from the transfer function.

(c) can be uniquely determined from the state variable equations, but the state
variable equations for a given transfer function are not unique.

Ans: (a)

(xviii) Which one of the following statements is not true?

(a) A SISO (single-input single-output) system with the state equations

dx1/dt =-3x1-x2+u
dx2/dt =2x1
y = x1-x2

is controllable

(b) A SISO system with the state equations

dx1/dt = -3x1-x2+u
dx2/dt = 2x1
y = x1-x2

is observable.

(c) A SISO system with the state equations

dx1/dt=-2x1+3x2+u
dx2/dt=x1
y=x1+u

is unobservable.

Ans: (c)
(xix) For the state-transition equation,
t
x (t) = (t-t0) x (t0)+ (t- ) Br ( ) d ,
t0
the Laplace transformed equation is

(a) X(s) = (sI –A) x (t0)+ (sI –A) –1BR(s)

(b) X(s) = (sI –A) –1x (t0)+ (sI –A) –1BR(s)
(c) X(s) = (sI –A) –1x (t0)+ (sI –A) BR(s)

Ans: (b)

( xx ) For a third-order plant, which of the following may be a set of state variables?

(a) y, dy/dt, d2y/dt 2

(b) y, dy/dt, y + d2 y/dt2
(c) y, d2y/dt2, y + d2y/dt2

Ans: (a)

(xxi ) Which one of the following systems is completely state controllable?

(a) dx1/dt= x2
dx2/dt =-x1+u
(b) dx1/dt=-2x1+x2+u
dx2/dt =-x2
(c) dx1/dt=-0.5x1
dx2/dt=-2x2+u
(d) dx1/dt=-x1+2u
dx2/dt=-2x2

Ans: (a)

( xxii) Which one of the following systems is completely observable?

(a) dx1/dt = -x1

dx2/dt = -2x2
y = x2
(b) dx1/dt = x1+x2
dx2/dt = -2x1-x2+u
y = x1
(c) dx1/dt = -2x1+3u
dx2/dt = -x2+u
y = x1
(d) dx1/dt = x2
dx2/dt = -x1-2x2
y = x1

Ans: (b)

(xxiii ) The transfer function

(a) can be uniquely determined from the state variable equations and the state variable
equations can also be uniquely determined from the transfer function.
(b) for a given set of state variable equations is not unique, but the state variable equations
can be uniquely determined from the transfer function.
(c) can be uniquely determined from the state variable equations, but the state variable
equations for a given transfer function are not unique.
(d) cannot be uniquely determined from the state variable equations, and the state
variable equations cannot be uniquely determined from the transfer function.

Ans:( c)

(xxiv ) The state transition matrix of the system defined by

dx1/dt =x1
dx2/dt = -2x1-3x2

is

(a) 2a -b
2b -a

(b) 2a-b a-b

-2a+2b -a + 2b

(c) a-b a-b

-a+b -a + b

(d) 2a-b a-b where a = e-t, b = e-2t

-2a-2b -a - 2b
Ans: b

( xxv ) The number of rabbits is x1 , and if left alone would grow indefinitely until the food
supply was exhausted.

dx1/dt = Kx1

However, with foxes present, we have

dx1/dt = Kx1- a .x2

where x2 is the number of foxes. Now, if the foxes must have rabbits to exist, we have

dx2/dt = -hx2 +b .x1

This system is stable and thus decays to the condition

x1 (t) = x2(t) = 0 at t = when

(a) h K, ab Kh
(b) h K, ab Kh
(c) h K, ab Kh

Ans:a

(xxvi ) In the problem of rabbits and foxes given in Part (xv), if K=1, h=3, and a=b=2, then

the transition matrix of the system is

(a) (t) = 2e-t -2 e-t

2 e-t -2t e-t

b) (t) = (2t +1)e-t -2t e-t

2t e-t (-2t +1) e-t

c) None of the above.

Ans :b
(xxvii ) In a series RLC circuit, if the current in the circuit and the voltage across the capacitor
are selected as the state variables, x1 and x2, respectively, then the state space description of the
circuit is given by

(a)

A= -R/L -1/L B = 1/L

0 1/C 0

(b)

A= -R/L -1/L B = 1/C

1/C 0 0

(c)

A = -R/L -1/L B = 1/L

1/C 0 0

Ans: c

(xxviii) Which one of the following statements is not true?

(a) A SISO(single-input single-output) system with the state equations

dx1/dt =-3x1-x2+u
dx2/dt =2x1
y=x1-x2

is controllable.

(b) A SISO system with the state equations

dx1/dt = -2x1+3x2+u
dx2/dt = x1+ u
y = x1

is uncontrollable.

(c ) A SISO system with the state equations

dx1/dt = -3x1-x2+u
dx2/dt = 2x1
y = x1-x2

is observable.

(d) A SISO system with the state equations

dx1/dt =-2x1+3x2+u
dx2/dt =x1
y = x1+u

is unobservable.

Ans: (d)

(xxix) Which of the systems with the following state space descriptions is completely
controllable?

(a)
A= -1 0 B= 2
0 -2 5

(b)
A= -1 0 B= 2
0 -2 0

( c)

A= 0 1 B= 1
2.5 -1.5 1
Ans:a

( xxx) Which of the following systems is not completely observable?

Note that Ct is the transpose of C.

(a)

A= -1 0 Ct = 1
0 -2 3
(b)

A= -1 0 Ct = 0
0 -2 1

(c)

A= 1 1 Ct = 1
-2 -1 0

B= 0
1

Ans: b

TOP