Control Engineering

NPTEL Online Course

Indian Institute of Technology Madras
Assignment 11
[Common data for Q1-Q2]
Consider the plant described by
   
0 1 1
ẋ = x+ u
7 −4 2
 
y= 1 3 x

1. Find the transfer function using matrix algebra

7s+27
(a) s2 +4s−7 . [Correct]
7s+27
(b) s2 +4s+7 .
7s+21
(c) s2 +4s−7 .
7s+21
(d) s2 +4s+7 .

2. Is system stable ?
(a) Yes
(b) No [Correct]
Solution: Using the formula G(s) = C(sI − A)−1 B,
we get G(s) = YU(s) 7s+27
(s = s2 +4s−7 .
One of the eigenvalues of A is in right half of ’s’ plane.
[Common data Q3-Q4]
Consider the circuit diagram shown below

Figure 1: Time Response

3. Find the output voltage for the circuit shown in Fig. 1, when the input is given at time t = 0.
Assume zero initial conditions.
(a) 5(1 − e3t/5 )
(b) 5(1 − e−3t/5 ) [correct]
(c) 5(1 + e−3t/5 )

1
 (d) 5(e−3t/5 )
4. Find the settling time (2%criterion) for the output voltage . [Answer: 6.67, Range:
5.85 - 7.00]
Solution:
Apply KVL around the circuit,

di
vin (t) = 5 + 3i
dt

In s-domain assuming zero initial conditions,

Vin (s) = (5s + 3)I(s)

Vo (s) = 3I(s)
Vin (s)
=3
(3 + 5s)
3×5
=
s(3 + 5s)
1 1
= 5( − )
s (s + 53 )

therefore,
3
vo (t) = 5(1 − e− 5 t )

This is of the form v0 (t) = A(1 − e−t/τ ) (or in RL circuit the time constant is L/R). Time
constant τ = 5/3 = 1.67
Thus, settling time T = 4τ = 6.67
5. Consider a system with state-space model

ẋ = Ax + Bu, y = Cx + Du

where    
3 2 1  
A= , B= , C= 1 0 , D = 0.
3 4 −1

Is it possible to obtain first-order state-space model of the system, i.e., a model of the form

ẋ = ax + bu, y = cx + du

where a, b, c, d are scalars.

(a) Yes [Correct]
(b) No

2
 Solution:
The input-output transfer function of the system is given by
 −1  
Y (s)  s−3 −2 1
= C(sI − A)−1 B + D = 1

0
U (s) −3 s−4 −1
  
1   s−4 2 1
= 1 0
(s − 1)(s − 6) 3 s − 3 −1
 
1   1
= s−4 2
(s − 1)(s − 6) −1
1
=
s−1
Now, from the above transfer function we get sY (s) − Y (s) = U (s). Taking inverse Laplace
transform (keeping in mind the zero initial conditions of a transfer function), we get ẏ − y = u.
Now, define the new state variable x = y. Then, we have ẋ = ẏ = y + u = x + u. Thus, a new
first-order state-space model for the same transfer function is

ẋ = x + u, y = x.

Common data for Q6-Q7  

2a b
The Jordon matrix of the transfer function 1/s is .
c d

6. The value of b= . [Answer = 1 Range: 0.99 to 1.01]

7. The value of c= . [Answer = 0 Range: -0.01 to 0.01]

Solution:
Not all the transfer functions can be expressed in diagonal form and this example is one of them.
(Why?)
First find the eigenvalues of the given matrix. If they are the same and if these eigenvalues do
not have independent eigenvectors, then use the Jordan normal form.
 
0 1
we get A = .
0 0
[Common data Q8-Q11]
For the nonlinear system, ẋ = f (x, u) and y = g(x, u) given by

x˙1 = −x1 x2 + 1
x˙2 = x1 x3 − x2 + u1 + u22
x˙3 = x23 (1 − x3 ) + 3u2
y = x21 + x2 + x3 + u1

8. Choose the correct equilibrium point (xeq , ueq ) of the system.

 
(a) ( 0 0 0 , 0)
 
(b) ( 1 1 1 , 0) [Correct]
 
(c) ( −1 −1 −1 , 0)

3
  
(d) ( 0 0 1 , 0)

9. Linearize the system about chosen equilibrium point, A matrix is

 
1 1 0
(a) 1 −1 −1
0 0 −1
 
0 0 0
(b) 0 −1 0
0 0 0
 
−1 −1 0
(c)  1 −1 1  [Correct]
0 0 −1
(d) None of the above
10. The B matrix is
 
0 0
(a) 1 2
0 3
 
0 0
(b) 1 0 [Correct]
0 3
 
0 0
(c) 1 0
0 1
(d) None of the above
11. Define the stability around the above chosen equilibrium point
(a) Stable [Correct]
(b) Unstable
(c) Can not define
(d) Marginal stable
Solution:
The equilibrium points of the system are obtained by setting all inputs to zero. Let (xeq , ueq ) be
the equilibrium point of the system. At equilibrium, f (xeq , ueq ) = 0. So

xeq eq
1 x2 = 1
xeq eq eq
1 x3 = x2
(xeq 2 eq
3 ) (1 − x3 ) = 0

Solving the above equations, we get 2 equilibrium points xeq = [1 1 1]T and [−1 − 1 1]T .

Linearizing the system about (xeq , ueq ) = ([1 1 1]T , 0).

Let δx = x − xeq and δu = u − ueq , using Taylor’s series
˙ = ẋ = f (x, u) = f (xeq , ueq ) + df |(xeq ,ueq ) δx + df |(xeq ,ueq ) δu + ....
δx dx du
Since f (xeq , ueq ) = 0,
˙ = df |(xeq ,ueq ) δx + df |(xeq ,ueq ) δu + ....
δx dx du

4
 dg dg
y = g(x, u) = g(xeq , ueq ) + dx |(xeq ,ueq ) δx + du |(xeq ,ueq ) δu + ....
eq eq dg dg
δy = y − g(x , u ) = dx |(xeq ,ueq ) δx + du |(xeq ,ueq ) δu + ...
For the given system,   
−x2 −x1 0 −1 −1 0
df
dx |(x ,u ) =
eq eq  x3 −1 x1  =  1 −1 1 
0 0 2x3 − 3x23 (xeq ,0) 0 0 −1
   
0 0 0 0
df
| eq eq
du (x ,u ) =  1 2u 2
 =  1 0
0 3 (xeq ,0) 0 3
dg
   
dx |(x ,u ) =
eq eq 2x1 1 1 (xeq ,0) = 2 1 1
dg
   
du |(x ,u ) =
eq eq 1 0 (xeq ,0) = 1 0
Analyse the eigenvalues of A.
[Common data: Q12-Q13]
Given the system       
ẋ1 −4 1 x1 0
= + u(t)
ẋ2 −2 −1 x2 1

12. For u(t) being a step input, find the steady state value of x1 = . [Answer = 0.17
Range: 0.12 - 0.23]
13. For u(t) being a step input, find the steady state value of x2 = . [Answer = 0.67,
Range = 0.62 - 0.72]

Solution

The solution to state equation is given by

Z t
At
x(t) = e x(0) + eA(t−τ ) Bu(τ )dτ.
0

The first term of the right hand side of equation is response due to initial state (Called as zero
input response (ZIR) ?) Since given system is stable and we now that at steady-state the effect
due of ZIR is 0.
The state transition matrix is given by

eAt = L−1 (sI − A)−1 .

On simplification we get

−e−2t + 2e−3t e−2t − e−3t

 
At
e =
−2e−2t + 2e−3t 2e−2t − e−3t

Thus
" −2t 3t
#
1
− 13 − e 2 + e3
2
x(t) = 3t .
1 − 31 − e−2t + e3
 
0.167
At steady state(t → ∞) then, x(t) =
0.67
Alternate Method
At steady state, the values of x1 and x2 do not change i,e,. ẋ1 = ẋ2 = 0

5
 We get the following equations by substituting ẋ1 = ẋ2 = 0 in the model:

−4x1 + x2 = 0
−2x1 − x2 + u = 0

As u(t) is a step input, we can substitute u(t) = 1 and solve the above equations to get steady
state values of x1 & x2 . We get

x1 = 0.17
x2 = 0.67

[Common data Q14-Q15]

Consider the following system
      
ẋ1 1 0 x1 0
= + u
ẋ2 1 1 x2 1

where u(t) is a unit step occurring at t = 0, and x1 (0) = 0, x2 (0) = 1.

14. Find the state transition matrix (eAt ).
 −t 
e 0
(a)
−te−t e−t
 t 
e 0
(b) [Correct]
tet et
 −t 
e 0
(c)
te−t e−t
 −t
−te−t

e
(d)
0 e−t

15. The value of x(t) is

 
0
(a)
(1 − 2e−t )
 
0
(b)
(1 − 2et )
 
0
(c)
(2e−t − 1)
 
0
(d) [Correct]
(2et − 1)
Solution:

 −1
−1 s−1 0
(sI − A) =
−1s − 1
 1 
0
= s−11 1
(s−1)2 s−1
t
 
e 0
=
tet et

6
 et
 
0
eAt =
tet et
Using the solution for state equation
Z t
At
x(t) = e x(0) + eA(t−τ ) Bu(τ )dτ.
0
  Z t 
0 o
= t +
e 0 e(t−τ ) dτ
 
0
=
2et − 1