2019 Answers PDF
2019 Answers PDF
2019 Answers PDF
dx(t)
x(t) → y(t) = tx(t) + .
dt
For any scalar a, we have
d(ax(t))
ax(t) → t(ax(t)) + = ay(t).
dt
Thus, homogeneity is also satisfied and the system is linear.
1
(c) Consider arbitrary inputs x1 (t) and x2 (t).
But,
Then,
2. Consider the following systems whose input x(t) and output y(t) are related as follows. Which
of them are time-invariant? (Multiple answers possible)
(a) y(t) = (x(t))2 [correct]
dx(t)
(b) y(t) = tx(t) + dt
(c) y(t) = x(t) + 1 [correct]
(d) y(t) = et x(t)
Solution: A system with input x(t) and output y(t) is said to be time-invariant if an arbitrary
delay of t0 in x(t) results in the same amount of delay in y(t).
(a) Let y1 (t) be the output of the system to a delayed input x(t − t0 ). Thus,
d(x(t − t0 ))
x(t − t0 ) → y1 (t) = tx(t − t0 ) +
dt
d(x(t−t0 ))
However, a delay of t0 in the output means y(t − t0 ) = (t − t0 )x(t − t0 ) + dt 6= y1 (t).
Thus, the system is time-variant.
(c) Consider a delayed input x(t − t0 ).
2
(d) Consider a delayed input x(t − t0 ).
However, a delay of t0 in the output means y(t − t0 ) = e(t−t0 ) x(t − t0 ) 6= y1 (t). Thus, the
system is time-variant.
3. Consider a continous-time system with input x(t) and output y(t) related by
The system is
(a) Linear and causal [correct]
(b) Linear and non-causal
(c) Non-linear and causal
(d) Non-linear and non-causal
Solution: A system is said to be causal if its output depends only on the past or present values
of the input.
Since t − 1 < t, y(t) depends only on the past values of x(t). Thus, the system is causal.
Now, to verify linearity, suppose
Then,
Solution: A system is said to be static if its output depends only on the present value of its
input. Statement 1 is true, since all static systems are causal by definition.
A system is said to be dynamic if its output depends on the past values of its input. Statement
2 is false, since not all dynamic systems are non-causal. For example, the system y(t) = x(t − 1)
is dynamic, but causal.
3
5. Select the two equations which correctly model the electrical circuit shown in the figure below.
(multiple correct choices possible)
R i1 L i2
+
v C1 C2
−
Rt
(a) v = Ri1 + C11 −∞ (i2 − i1 ) dt
Rt Rt
(b) C11 −∞ (i1 − i2 ) dt = L di 1
dt + C2 −∞ i2 dt [correct]
2
Rt
(c) v = Ri1 + C11 −∞ (i1 − i2 ) dt [correct]
Rt Rt
(d) C11 −∞ (i2 − i1 ) dt = L di 1
dt + C2 −∞ i2 dt
2
Solution: Applying KVL to the two loops in the circuit give the equations mentioned in choices
(b) and (c).
6. Select the two equations which correctly model the mass-spring-damper system shown in the
figure below. (multiple correct choices possible)
x1 x2
B K
M1 M2
F1 F2
Solution: The system consists of two nodes, one at each mass M1 , M2 . The various forces at
these nodes are shown in the figure below.
B K
M1 M2
F1 F2
K(x2 − x1 ) K(x2 − x1 )
M1 x¨1 M2 x¨2
B x_1
Thus, choices (a) and (d) represent the equations of motion at node M1 and M2 respectively.
4
7. By the force-voltage analogy (F ≡ v), which electrical circuit is analogous to the mass-spring-
damper system shown in the figure below?
K x
B1 M
F
B2
i K B1 B2
(a) +
v 1=M
−
i 1=K B1 B2
(b) + [correct]
v M
−
i
B1
(c)
+ B2
v 1=M
−
1=K
i
B1
(d)
+ B2
v M
−
Solution: The system consists of one node, at mass M . The equation of motion at this node is
F = M ẍ + Kx + B1 ẋ + B2 ẋ. (1)
v = M q̈ + Kq + B1 q̇ + B2 q̇, (2)
where v is the source voltage and q is the charge. Since i = dq/dt = q̇, we can rewrite (2) as
Z
v = M i̇ + K i dt + B1 i + B2 i. (3)
5
Compare this with Z
1
v = Li̇ + i dt + R1 i + R2 i, (4)
C
where L is an inductor, C is a capacitor and R1 , R2 are resistors. We can see that the circuit
corresponding to equation (3) is choice (b).
8. Consider two gears G1 and G2 coupled with each other with the set of values as shown in the
table below.
G1 G2
Torque T1 T2
Angular displacement θ1 θ2
Radius r1 r2
Assume the coupling is lossless and the linear distance travelled by both gears is the same. Which
of the following relations is correct?
T1 θ1 r1
(a) T2 = θ2 = r2
T1 θ2 r1
(b) T2 = θ1 = r2 [correct]
T1 θ1 r2
(c) T2 = θ2 = r1
T1 θ2 r2
(d) T2 = θ1 = r1
Solution: Since the coupling is lossless, the power at G1 is equal to the power at G2 , i.e.,
T1 θ1 = T2 θ2 .
Since the linear distance travelled by G1 and G2 is the same, r1 θ1 = r2 θ2 .
Thus, (b) is the correct choice.
9. Consider the following predator-prey model, where x is the number of preys and y is the number
of predators.
dx
= 3x − 2xy,
dt (5)
dy
= −6y + 3xy.
dt
Which of these is not an equilibrium point of the model?
(a) x = 0, y = 0
(b) x = 2, y = 1.5
(c) x = 1.5, y = 2 [correct]
Solution: A point (x∗ , y ∗ ) is said to be an equilibrium point of a system if dx/dt, dy/dt are both
zero when evaluated at x = x∗ , y = y ∗ .
It is easy to verify that choices (a) and (b) satisfy the above definition, while (c) does not.
10. In the model described in Question 9, what happens to the predator population in the absence
of any preys?
(a) remains constant
(b) increases exponentially
(c) decreases exponentially [correct]
6
Solution: In the absence of any preys, i.e., x = 0, we get
dy
= −6y. (6)
dt
Integrating this equation, we get
Z Z
dy
= −6 dt + c
y
ln y = −6t + c
y = e−6t ec ,
7
Control Engineering
NPTEL Online Course (Jan-Apr 2019)
Indian Institute of Technology Madras
Solutions 2
Note: Unless mentioned otherwise, all signals are causal, i.e., their value is zero for all time t < 0.
1. Which of the following properties of a sinusoidal signal does not change when passed through
an LTI system?
(a) Amplitude
(b) Frequency [correct]
(c) Phase
Solution: Consider an LTI system with impulse response h(t). Suppose an input x(t) =
A sin(ωt + φ) is applied to it. We show that the output can have a different amplitude Ã
and a phase shift φ̃, but the sinusoidal nature of the signal and its frequency remain unchanged.
The output of the system is given by
Z ∞ Z ∞
y(t) = h(t) ? x(t) = h(u)x(t − u) du = h(u)A sin(ω(t − u) + φ) du
−∞ −∞
Z ∞
=A h(u) sin(ωt + φ − ωu) du
−∞
Z ∞
=A h(u) sin(ωt + φ) cos(ωu) − cos(ωt + φ) sin(ωu) du
−∞
Z ∞ Z ∞
=A h(u) cos(ωu) du sin(ωt + φ) − h(u) sin(ωu) du cos(ωt + φ)
−∞ −∞
R∞ R∞
Let Hc = −∞
h(u) cos(ωu) du and Hs = −∞
h(u) sin(ωu) du. Then,
y(t) = A Hc sin(ωt + φ) − Hs cos(ωt + φ)
p
2 2
Hc Hs
= A Hc + Hs p sin(ωt + φ) − p cos(ωt + φ)
Hc2 + Hs2 Hc2 + Hs2
2. What is the Region of Convergence (ROC) of the Laplace transform of the signal te−3t ? Note
that Re(s) denotes the real part of the complex number s.
1
Solution: The Laplace transform of the signal is given by
Z ∞ Z ∞
−3t −3t −st
X(s) = L{te } = te e dt = te−(s+3)t dt
0 0
−(s+3)t ∞ Z ∞ −(s+3)t
e
− e
= t dt
−(s + 3) 0
0 −(s + 3)
−(s+3)t ∞
1 e
=0+
s + 3 −(s + 3) 0
Now, limt→∞ e−(s+3)t exists and is equal to zero only if Re(s + 3) > 0, i.e., Re(s) > −3. Then,
1
X(s) =
(s + 3)2
2 2
Solution: We know L{sin(2t)} = s2 +4 . So, L{e4t sin(2t)} = (s−4)2 +4 .
s d s s2 −9
Also, we know L{cos(3t)} = s2 +9 . So, L{t cos(3t)} = − ds s2 +9 = (s2 +9)2 .
2 s2 −9
So, L{sin(2t) + t cos(3t)} = (s−4)2 +4 + (s2 +9)2 .
4. Using the initial value theorem of Laplace transform, what is the initial value of the signal x(t)
s+2
if its Laplace transform is X(s) = (s+1)2 +5 ?
(a) 4
(b) −2
(c) 1 [correct]
(d) does not exist
Solution: By initial value theorem, the initial value of x(t) is
s+2 s2 + 2s
lim x(t) = lim sX(s) = lim s = lim = 1.
t→0 s→∞ s→∞ (s + 1)2 + 5 s→∞ s2 + 2s + 6
5. Using the final value theorem of Laplace transform, what is the final value of the signal x(t) =
3 sin(2t) + 1?
(a) 4
(b) −2
(c) 1
(d) does not exist [correct]
Solution: Since a sinusoidal signal oscillates for all time t, the limit limt→∞ sin(2t) does not
exits. So, the final value of the given signal does not exist.
2
se−s
6. What is the inverse Laplace transform of (s+2)2 +9 ?
Now,
se−s
−1
2
L = e−2(t−1) cos(3(t − 1)) − sin(3(t − 1))
(s + 2)2 + 9 3
7. Using the convolution theorem of Laplace transform, what is the convolution of the signals et
and sin(2t)?
(a) 52 et + cos(2t) + 21 sin(2t)
1 2
Solution: We know that L{et } = s−1 and L{sin(2t)} = s2 +4 .
Suppose ? denotes convolution. By convolution theorem, we have
2
L{et ? sin(2t)} = L{et }L{sin(2t)} = .
(s − 1)(s2 + 4)
So,
2
et ? sin(2t) = L−1 .
(s − 1)(s2 + 4)
For partial fraction expansion, let
2 A Bs + C
= + 2 .
(s − 1)(s2 + 4) s−1 s +4
So, 2 = A(s2 + 4) + (Bs + C)(s − 1). Equating the coefficients of s on both sides and solving for
A, B, C, we get A = 2/5, B = −2/5 = C. Thus,
t −1 2 2 −1 1 s+1
e ? sin(2t) = L = L −
(s − 1)(s2 + 4) 5 s − 1 s2 + 4
2 1 s 1 2
= L−1 − 2 −
5 s − 1 s + 4 2 s2 + 4
2 1
= et − cos(2t) − sin(2t) .
5 2
8. Using Laplace transform, what is the solution of the ODE ẍ + 2ẋ + x = e−t with the initial
condition x(0) = 1, ẋ(0) = −1?
3
t2
(a) x(t) = e−t
2 + 1 [correct]
2
(b) x(t) = e−t t2 − 1
2
(c) x(t) = et t2 + 1
2
(d) x(t) = et t2 − 1
t2
−1 1 1 −t
x(t) = L + = e 1 + .
s + 1 (s + 1)3 2
(a) Only 1
(b) Only 2
(c) Both 1 and 2
(d) None of them [correct]
Solution: The transfer function of a system is a property of the system, hence it is independent
of the nature of the input applied. So, statement 1 is false.
It is possible that two systems with different physical natures, for example one electrical and
other mechanical, have the same transfer function. So, statement 2 is also false.
10. Consider the system shown in the figure below.
1 2H x
+ + 10Nm
v 1F vo 2kg
v
F = t
The input to the system is the supply voltage v(t). It generates a voltage vo (t) at the capacitor.
The force F experienced by the mass is equal to the rate of change of vo (t) with respect to time.
4
x(t) is the displacement of the mass from its rest position. The values of the parameters of the
system are as shown in the figure. What is the transfer function from the supply voltage to the
displacement, i.e., X(s)
V (s) ?
s
(a) (2s2 +1)(2s2 +s+10)
1
(b) (2s2 +1)(2s2 +s+10)
s
(c) (2s2 +10)(2s2 +s+1) [correct]
1
(d) (2s2 +10)(2s2 +s+1)
5
Control Engineering
NPTEL Online Course (Jan-Apr 2019)
Indian Institute of Technology Madras
Solutions 3
Note: Unless mentioned otherwise, all signals are causal, i.e., their value is zero for all time t < 0.
Y (s)
1. What is the transfer function R(s) in the block diagram shown below?
H3
R(s) + Y (s)
+ G1 + G2 + G3 G4
− −
H1 H2
G1 G2 G3 G4
(a) 1+G1 G2 H1 +G3 G4 H2 +G2 G3 H3 +G1 G2 G3 G4 H1 H2
G1 G2 G3 G4
(b) 1+G1 G2 H1 +G3 G4 H2 +G2 G3 H3 +G1 G2 G3 G4
G1 G2 G3 G4
(c) 1+G1 G2 H1 +G3 G4 H2 −G2 G3 H3 +G1 G2 G3 G4 H1 H2 [correct]
G1 G2 G3 G4
(d) 1+G1 G2 H1 +G3 G4 H2 −G2 G3 H3 +G1 G2 G3 G4
Solution: First, we shift G1 ahead of the second summing block and shift G4 before the second
take off point as below.
1 1
G1 H3 G4
R(s) + Y (s)
+ + G1 G2 + G3 G4
− −
H1 H2
Note that we can interchange the first and second summing blocks as follows.
1
1 1
G1 H3 G4
R(s) + Y (s)
+ + G1 G2 + G3 G4
− −
H1 H2
Now, we can simplify the inner two negative-feedback loops to get the following.
1 1
G1 H3 G4
R(s) + G1 G2 G3 G4
Y (s)
+ 1+G1 G2 H1 1+G3 G4 H2
Y (s) G1 G2 G3 G4
= .
R(s) 1 + G1 G2 H1 + G3 G4 H2 − G2 G3 H3 + G1 G2 G3 G4 H1 H2
2. Which signal flow graph correctly represents the block diagram shown in Question 1?
H3
R(s) 1 G1 G2 G3 G4 1 Y (s)
(a)
H1 H2
H3
R(s) 1 G1 G2 G3 G4 1 Y (s)
(b)
−H1 −H2
H3
H1 H2
2
H3
−H1 −H2
3. Which signal flow graph correctly represents the electrical system shown in the figure below?
i1 1Ω i2 2Ω
+
v 2H 1F
−
1
2s+1
(a)
(b)
(c) [correct]
(d)
3
I1 (s) I2 (s)
1 2
+
1
V (s) 2s s
−
So,
1 2s
I1 (s) = V (s) + I2 (s). (1)
2s + 1 2s + 1
Applying KVL in the right loop gives
1
2s(I1 (s) − I2 (s)) − 2I2 (s) − I2 (s) = 0.
2s
So,
2s2
I2 (s) = I1 (s). (2)
2s2 + 2s + 1
The signal flow graph in choice (c) is obtained from equations (1) and (2).
Y (s)
4. Using Mason’s gain formula, what is the transfer function R(s) in the signal flow graph shown
below?
G5
R(s) 1 G1 G2 G3 G4 1 Y (s)
−H1
−H2
G1 G2 G3 G4 +G4 G5 (1+G2 H1 )
(a) 1+G2 H1 +G4 G5 H2 +G1 G2 G3 G4 H2
G1 G2 G3 G4 +G4 G5 (1+G2 H1 )
(b) 1+G2 H1 +G4 G5 H2 +G1 G2 G3 G4 H2 +G2 G4 G5 H1 H2 [correct]
G1 G2 G3 G4 +G4 G5
(c) 1+G2 H1 +G4 G5 H2 +G1 G2 G3 G4 H2
G1 G2 G3 G4 +G4 G5
(d) 1+G2 H1 +G4 G5 H2 +G1 G2 G3 G4 H2 +G2 G4 G5 H1 H2
Solution: The graph has two forward paths, so N = 2. The gains of the two paths are
P1 = G1 G2 G3 G4 , P2 = G4 G5 .
4
The graph has three loops with gains
∆ = 1 − (L1 + L2 + L3 ) + L1 L3 = 1 + G2 H1 + G1 G2 G3 G4 H2 + G4 G5 H2 + G2 G4 G5 H1 H2 .
All three loops touch the first forward path P1 , so ∆1 = 1. The loop L1 does not touch the
second forward path P2 , so ∆2 = 1 − L1 = 1 + G2 H1 .
Finally, by Mason’s gain formula, we get
Y (s) P 1 ∆ 1 + P2 ∆ 2 G1 G2 G3 G4 + G4 G5 (1 + G2 H1 )
= = .
R(s) ∆ 1 + G2 H1 + G4 G5 H2 + G1 G2 G3 G4 H2 + G2 G4 G5 H1 H2
Y (s) 10
5. What is the response y(t) of the system R(s) = s(s+2) when the input r(t) is a unit ramp signal?
1 t t2 e−2t
(a) 5 2 + 4 + 4 + 2
1 t t2 e−2t
(b) 5 4 + 2 + 2 + 4
1 t t2 e−2t
(c) 5 2 − 4 + 4 − 2
1 t t2 e−2t
(d) 5 4 − 2 + 2 − 4 [correct]
1 10 10
Solution: r(t) = t. So, R(s) = s2 . Then, Y (s) = R(s) s(s+2) = s3 (s+2) . For partial fraction
expansion, let
10 A B C D
= + 2+ 3+ .
s3 (s+ 2) s s s s+2
Solving this, gives A = 5/4, B = −5/2, C = 5, D = −5/4. Thus,
1 1 1 1
Y (s) = 5 − + 3− .
4s 2s2 s 4(s + 2)
t2 e−2t
1 t
y(t) = 5 − + − .
4 2 2 4
6. Consider the two responses y1 (t), y2 (t) of a second-order system to a unit step input, correspond-
ing to damping ratios ζ1 , ζ2 respectively, as shown in the figure below. Assume that the natural
frequency ωn is same for both the responses.
5
1.5
y1 (t) corresponding to ζ1
y2 (t) corresponding to ζ2
0.5
0
0 2 4 6 8 10
t
Which of the following are possible values of ζ1 , ζ2 ? (multiple correct choices possible)
(a) ζ1 = 1, ζ2 > 1 [correct]
(b) ζ1 > 1, ζ2 = 1
(c) ζ1 > ζ2 > 1
(d) ζ2 > ζ1 > 1 [correct]
Solution: From the figure, we can infer that both responses are either of a critically damped
system or an overdamped system, i.e.,
ζ1 ≥ 1, ζ2 ≥ 1. (3)
Also, we can see from the figure that y1 (t) is faster than y2 (t), i.e.,
ζ2 > ζ1 . (4)
Choices (a) and (d) satisfy the constraints in equations (3) and (4), while choices (b) and (c) do
not.
7. What is the percentage peak overshoot Mp in the unit step response of the closed-loop system
Y (s) 100
R(s) = s2 +4s+36 ?
Accepted answer: within range 32-34.
Y (s)
Solution: Comparing the given transfer function with the one in the standard form R(s) =
2
Kωn
2 ,we
s2 +2ζωn +ωn get ωn = 6, ζ = 1/3.
Then, percentage peak overshoot is
√ 2
Mp = e−πζ/ 1−ζ × 100 = 32.9
8. In the system shown in the figure below, what should be the value of the gain K to obtain
response y(t) with a peak time tp = 1 sec, when r(t) is a unit step input?
6
R(s) Y (s)
1
+ K s(s+3)
−
Y (s) K
= 2 .
R(s) s + 3s + K
√ 3
Comparing this with the standard form gives ωn = K, ζ = √
2 K
. The formula for peak time of
the unit step response is
π π
tp = = p .
ωd ωn 1 − ζ 2
For the above values of ωn and ζ, we need tp = 1 sec, i.e.,
π
1= √ q .
9
K 1− 4K
K = 12.1
9. What is the steady-state error in the unit step response of the unit negative feedback system
1
with open-loop transfer function G(s) = s2 +5s+3 ?
(a) 0
(b) ∞
(c) 0.25
(d) 0.75 [correct]
Solution: For a unit step input to a unit negative feedback system, the steady-state error is
given by
1 1 1 1
ess = = = 1 = = 0.75
1 + Kp 1 + lims→0 G(s) 1 + lims→0 s2 +5s+3
1 + 1/3
10. In the system shown in the figure below, what should be the values of K and a to obtain a
steady-state error of 0.2 for a unit ramp input?
R(s) Y (s)
K 1
+ sa s+10
−
7
(c) K = 100, a = 1
(d) K = 100, a = 2
Since we want a non-zero, finite steady-state error for a ramp input, the system must be of type
1, i.e., a = 1.
Further, the exact value of the error is ess = 0.2. So,
1 1 1 10
0.2 = ess = = = K
= .
Kv lims→0 sG(s) lims→0 s s(s+10) K
Thus, K = 50.
8
NPTEL Online Course (January - April 2019)
Control Engineering
Assignment 4 Solution
Q.1
The forward-path transfer function is
K(1 + as)
G(s) = , a > 0, b > 0
s(1 + bs)
RH Tabulation is as follows
s2 b K
s1 1 + Ka 0
s0 K 0
For stability first row of RH must not change sign and since a and b arek positive, we have
i) 1 + Ka > 0 =⇒ K > − a1 and ii) K > 0.
Q.2
The given characteristic equation is
s4 1 13 K +5
s3 5 15 0
s2 10 K +5
25−K
s1 10 0
s0 K +5 0
For stability first row of RH must not change sign, so we have
i) 25−K
10 =⇒ K < 25 and ii) K + 5 > 0 =⇒ K > −5.
Q.3
The Routh table for the characteristic equation corresponding to the given transfer function is
s4 1 −6 25
s3 0 0
The entire row of s3 is zero. The auxiliary equation is given by
A(s) = s4 − 6s2 + 25 = 0
dA(s)
= 4s3 − 12s
ds
The 3rd row is replaced with the coefficients of the derivative of auxiliary equation A(s) = 0.
The Routh table becomes
s4 1 −6 25
s?3 4 −12
s2 −3 25
64
s1 3 0
s1 25
There are two sign change in the first column of the Routh table. So, there are 2 poles in the RHP.
Since Order of the characteristic equation, n = 4 = number of poles.
∴ Number of poles in the LHP = 4 − 2 = 2.
Answer: There are 2 poles in the RHP and 2 poles in the LHP
Q.4
Characteristic equation of the four systems are:
s + 2 =0
s1 1 0
s0 2 0
s2 + s + 1 = 0
s2 1 1
s1 1 0
s0 1 0
2s + 1 = 0
s1 2 0
s0 1 0
s3 + s2 + 1 = 0
s3 1 0
s2 1 1
s1 −1 0
s0 1 0
In all the characteristic equations except s3 + s2 + 1 = 0, the coefficients are positive (greater than 0), indicating
stability by Routh-Hurwitz criterion.
There are two sign changes in the first column of the Routh Table of s3 + s2 + 1 = 0. Therefore, two roots of the
closed loop system lie in the RHP. Hence, closed loop system is unstable.
Q.5
The open loop transfer function is given by
4
G(s) =
s(s2 + qs + 2p)
s3 + qs2 + 2ps + 4 = 0
Since the closed loop system is oscillating at a frequency of 4 rad/s then the coefficient of s1 must be 0 in general and to
be replaced by the coefficient of dA(s)
ds .
2pq−4
∴ q = 0 =⇒ pq = 2.
Now, the auxiliary equation A(s) is given by
A(s) = qs2 + 4 = 0
≡ q( jω)2 = −4
⇒ qω 2 = 4
⇒ q × 42 = 4
⇒ q = 0.25
2
∴ p= 0.25 = 8.
Option c is correct.
Q.6
The characteristic equation is
Q.7
The characteristic equation of the closed loop system is
Test:
a) K = 40, a = 0.5 −→ m = 10
b) K = 40, a = 1 −→ m = -26 =⇒ Case iii) violated.
c) K = 10, a = 5 −→ m = -20 =⇒ Case iii) violated.
d) K = 10, a = 2 −→ m = 1.6
e) K = 80, a = 2.5 −→ Case i) violated.
f) K = 80, a = 10 −→ Case i) violated.
Option a and d are correct.
Q.8
s6 +ve
s5 +ve
s4 +ve
s3 −ve
s2 −ve
s1 −ve
s0 −ve
The number of sign changes in first column of RH table is 1. So, number of roots in RH of s-plane = 1.
Hence system is unstable.
Option a and e are correct.
Q.9
From the Routh table of characteristic equation, we have
s5 1 18 81
s4 2 36 162
s3 0 0 0
s5 1 18 81
s4 2 36 162
s?3 8 72 0
s2 18 162 0
s1 0 0 0
Again, the entire row for s1 is zero.
The auxiliary equation is given by
Q.10
Case 1: For the poles lie to the left of s = -2
Substituting s = z − 2 into the characteristic equation of the given system, we have an equation in terms of z as
(z − 2)2 + 2ζ ωn (z − 2) + ωn2 = z2 + (2ζ ωn − 4)z + (ωn2 − 4ζ ωn + 4) = 0 (5)
Plugging ζ = 0.5, equation (5) reduces to
z2 + (ωn − 4)z + (ωn2 − 2ωn + 4) = 0 (6)
The Routh table of equation (6) is
z2 1 ωn2 − 2ωn + 4
z1 ωn − 4 0
z0 ωn2 − 2ωn + 4 0
Since the first element of first column is positive, the condition for poles to lie to the left are
ωn − 4 > 0 and ωn2 − 2ωn + 4 > 0. (7)
Since, ωn2 − 2ωn + 4 = (ωn − 1)2 + 3 > 0 ∀ ωn ∈ R.
Control Engineering
Assignment 5 Solution
Q.1
By Evan’s condition, the characteristic roots of the system must satisfy
Q.2
The following are true statements for root locus:
1. the root locus begins at the poles and ends at the zeros.
2. the value of K increases from zero to infinity.
3. the root locus on the real axis always lies in a section of the real axis to the left of an odd number of poles and zeros.
4. the root locus is symmetrical with respect to the horizontal real axis.
Option a and d are correct.
Q.6
The characteristic equation is given as s2 + αs + 1 = 0. The equation can be written as
s
s2 + αs + 1 = 0 =⇒ (s2 + 1) + αs = 0 =⇒ 1 + α 2 =0
s +1
s
The modified Ĝ(s) = s2 +1
. Hence, the poles are s = ± j and the zeros is s = 0.
Option b is correct.
Q.7
Given test points are s1 = −2 + j3 and s2 = −2 + j √12 .
For s1 , angle contribution of the system poles and zeros on the test point is
3 π
arctan( ) + arctan(3) − − π − arctan(3) ≈ −70.56o
2 2
For s2 , angle contribution of the system poles and zeros on the test point is
1 1 π 1
arctan( √ ) + arctan( √ ) − − π − arctan( √ ) = −180o
2 2 2 2 2
The angle contribution of all system poles and zeros on a point on the root locus is always an odd multiple of ±180o .
From the angle contributions computed above it can be concluded that s2 is a point on the root locus but not s1 .
Option c is correct.
Q.8
The open loop gain K at any point s0 on the root locus is determined as
∏ni=1 |s0 + pi |
K=
∏mi=1 |s0 + zi |
Here, s0 = −5 & poles are p1 = 0, p2 = −0.5, p3 = j2, p4 = − j2 and zero is z1 = −2.
5 × 4.5 × 29
∴ K= = 217.5
3
Option a is correct.
Q.9
The C.E. in the standard form of root locus can be expressed as
K
1+ =0
s4 + 12s3 + 64s2 + 128s
Determining the poles and zeros, we have
K
1+ =0
s(s + 4)(s + 4 + j4)(s + 4 − j4)
The center of the asymptotes is
∑ Re(pi ) − ∑ Re(zi ) (0 − 4 − 4 − 4) − (0)
σA = = = −3
#pi − #zi 4−0
Option c is correct.
Q.10
The angles of the asymptotes are
(2k + 1)
φA = 180◦ , k = 0, 1, 2, 3;
4
∴ φA = 45◦ , 135◦ , 225◦ , 315◦
Option d is correct.
Q.11
The break point is estimated by evaluating
between s = 4 and s = 0.
dK
= −(4s3 + 36s2 + 128s + 128)
ds
dK
= 0 =⇒ s = −1.557, −3.7 ± j2.55
ds
d2K
= −(12s2 + 72s + 128)|(s=−1.557) = −44.3 < 0
ds2
Since, s = −1.557 lies in the root locus and maximises the value of K, hence it is a break-away point.
Option a is correct.
Q.12
The angle of departure θd at the complex pole p1 = −4 + j4 can be estimated by utilizing the angle criterion as follows:
θd + ∠(p1 , −4) + ∠(p1 , −4 − j4) + ∠(p1 , 0) = 180◦ + k360◦ , k = ±1, ±2, ...
Q.13
The break point is estimated by evaluating
K(s − 2)(s − 4)
1+ =0
s2 + 6s + 25
s2 + 6s + 25
K=−
s2 − 6s + 8
dK 2(6s2 + 17s − 99)
=
ds (s2 − 6s + 8)2
dK
= 0 =⇒ s = 2.9, −5.7
ds
d2K 2(12s3 + 51s2 − 594s + 1052)
2
=− |(s=2.9) = 105 > 0
ds (s2 − 6s + 8)3
Since, s = 2.9 lies in the root locus and minimises the value of K, hence it is a break-in point. s = −5.7 does not lie on
root locus.
Option c is correct.
Root Locus
15
10
Imaginary Axis (seconds-1 )
-5
-10
-15
-15 -10 -5 0 5 10 15
-1
Real Axis (seconds )
Q.14
The characteristic equation is
(K + 1)s2 + (6 − 6K)s + 25 + 8K = 0
RH Tabulation is as follows
s2 K +1 25 + 8K
s1 6 − 6K 0
s0 25K + 8 0
The elements in the second row of RH table are zero for K = 1. Hence, the root locus crosses the imaginary-axis at
√
2s2 + 33 = 0 =⇒ s = ± j 16.5 = ± j4.06
Option b is correct.
Q.15
The elements in the second row of RH table are zero for K = 1. The value of the gain, K, at the jω-axis crossing is 1.
Option c is correct.
Q.16
Given ζ = 0.5.
The standard second order equation with ζ = 0.5 can be given as s2 + ωn s + ωn2 .
The characteristic equation of the given system is
(K + 1)s2 + (6 − 6K)s + 25 + 8K = 0
Comparing coefficients of both equation, we get
6 − 6K 25 + 8K
ωn = and ωn2 =
K +1 K +1
Solving for ωn , we get K = 0.108 and K = 3.64.
Option b is correct.
Root Locus
5
0.5
4
3
Imaginary Axis (seconds-1 )
-1
-2
-3
-4
0.5
-5
-4 -3 -2 -1 0 1 2 3 4
-1
Real Axis (seconds )
Control Engineering
Assignment 6 Solution
Q.1
The phase of the system at (-1,j0) is 180◦ .
Phase Margin=180◦ + ∠G( jω)H( jω)|ω=ωgc = 180◦ − 180◦ = 0
Q.2
The Bode plot of the transfer function G(s) = −1/s2 has -40 dB/dec slope and a phase shift of −π − π/2 − π/2 = −2π
radians.
N = #Z − #RHP = 1 − 0 = 1.
Hence, one encirclement in clockwise direction.
s2 + 3s + 12 (s + 1.5)2 + 9.75
1 + G2 (s) = =
(s + 1)(s + 2) (s + 1)(s + 2)
For 1 + G2 (s) we have #RHP = 0 and right half zeros #Z = 0.
N = #Z − #RHP = 0 − 0 = 0.
Hence, no encirclement.
Q.5
At gain crossover frequency (ωgc ), magnitude of |G( jωgc )| = 1.
Q.6
K K
G(s) = =⇒ |G( jω)| = 3
s3 ω
K
|G( jωgc )| = 1 =⇒ 3 = 1
ωgc
√3
∴ ωgc = K rad/s
Q.7
The open loop transfer function G(s)= 1 −2T s , T > 0.
s(3s+1) e
The phase equation is given as φ = −2ωT − 90◦ − tan−1 (3ω).
Let at ω = ω̂ the value of φ be 0.
Q.8
1
G(s) =
s(s + 1)(s + 2)
Put s = jω
At ω −→ 0+ , x −→ −3/4 and y −→ −∞
Q.9
Ke−s Ke− jω
G(s) = =
s jω
At the gain crossover frequency, ωg , the gain is 1. Then,
K
|G( jωg )| = = 1 ⇒ ωg = K rad/s
ωg
The phase at the gain crossover frequency is
π π
∠G( jωg ) = −ωg − = −K −
2 2
π π π
P.M. = π − K − = =⇒ K = = 1.047
2 6 3
1
|G( jωg )| = q q =1
ωg (ωg2 + a2 ) (ωg2 + b2 )
⇒ ωg2 (ωg2 + a2 )(ωg2 + b2 ) = 1
⇒ 0.2(0.2 + a2 )(0.2 + b2 ) = 1
⇒ (a2 + b2 ) + 5a2 b2 + 0.2 = 25
⇒ (a2 + b2 ) + 5 × 4 + 0.2 = 25
⇒ (a2 + b2 ) = 4.8
Solving a2 + b2 = 4.8 and ab = 2, we get b = 1.93 and a = 1.04 or a = 1.93 and b = 1.04.
Since it is given that b > a, hence b = 1.93 and a = 1.04.
Q.12
K + j0.366ω
G( jω) = , H( jω) = 1
( jω)(1 + jω)
0.366ω
gc
P.M. = 180◦ + ∠G( jωgc )H( jωgc ) = 60◦ =⇒ 180◦ + tan−1 − 90◦ − tan−1 (ωgc ) = 60◦
K
0.366ω
gc
=⇒ tan−1 (ωgc ) − tan−1 = 30◦
K
0.366ωgc
ωgc − K 1
∴ 0.366ωgc
=√ (1)
1 + ωgc ( K ) 3
Substituting ωgc = 1 rad/sec, we get K = 1.366
Q.13
630
G(s)H(s) =
(s + 3)(s + 4)(s + 6)
At the gain crossover frequency, ωg , the gain is 1. Then,
630
|G( jωg )| = q q q =1
(ωg2 + 9) (ωg2 + 16) (ωg2 + 36)
⇒ (ωg2 + 9)(ωg2 + 16)(ωg2 + 36) = 6302
Taking ωg2 = t
⇒ (t + 9)(t + 16)(t + 36) = 396900
⇒ t 3 + 61t 2 + 1044t − 391716 = 0
⇒ t = 54
√ √
⇒ ωg = 54 = 3 6 = 7.35
Q. 14
At the phase crossover frequency, ω p , ∠GH = −π. Then,
ωp ωp ωp
− arctan − arctan − arctan = −π
3 4 6
ωp ωp ωp
⇒ arctan + arctan + arctan =π
3 4 6
ωp ωp
ωp 3 + 6
⇒ arctan + arctan ω ω =π
4 1 − 3p 6p
ωp
ωp 2
⇒ arctan + arctan =π
4 ω2
1 − 18p
ωp
ωp 2
4 + ω2
1− 18p
⇒ arctan ωp =π
ωp
1− 4
2
ω2
1− 18p
p ω
ωp
⇒ + 2 2 =0
4 ω
1 − 18p
ωp ω p2 ωp
⇒ × (1 − )+ =0
4 18 2
ωp ω p2
⇒ (1 − + 2) = 0
4 √ 18 √
⇒ ω p = 54 = 3 6 = 7.35
Q.15
1
G.M.(in dB) = 20 log = 20 log(1) = 0 dB
GH(ω p )
Q.16
Q.17
Since ωg = ω p also G.M= 0 and P.M. =0, it is clear that system is marginally stable.
Q.18
100
G(s)H(s) =
s2 + 2bs + 100
100 100(100 − ω 2 − j2bω)
G( jω)H( jω) = 2
=
100 − ω + j2bω (100 − ω 2 )2 + (2bω)2
The Nyquist plot intersect imaginary axis at s = 0 − j1.18. Comparing coefficients, we get
100 5
100(100 − ω 2 ) = 0 =⇒ ω = 10 and = 1.18 =⇒ b = = 4.24
2bω 1.18
Q.19
Comparing the given system with a standard second order system, we get
Q.20
Ks
G(s)H(s) =
(s − 1)(s2 + s + 1)
( jω) −(ω 4 ) − jω
G( jω)H( jω) = = 2
( jω − 1)(1 − ω + jω) (ω + 1)(1 − ω 2 + ω 4 )
2
Q.21
s2 + s − 2
G(s)H(s) =
s2 + 16
( jω)2 + jω − 2 −(ω 2 + 2) + jω
G( jω)H( jω) = =
( jω)2 + 16 16 − ω 2
p
(ω 2 + 2)2 + ω 2
|G( jω)H( jω)| =
|16 − ω 2 |
(
180◦ − tan−1 ω 2ω+2 ,
ω ∈ (−4, 4)
∠G( jω)H( jω) = −1 ω
− tan ω 2 +2
, otherwise.
K(s + z1 )(s + z2 ) . . . (s + zm )
G(s) = .
(s + p1 )(s + p2 ) . . . (s + pn )
With a proportianl controller Kc in cascade, the position error constant of the system will be
1
Figure 1: Mechanical system with a controller
10s3 + 5s2 + KP s + 1 = 0.
s3 10 KP
s2 5 1
5KP −10
s1 5
s0 1
5KP −10
Thus, for stability, we need KP > 0 and 5 > 0, which implies KP > 2.
5. If Gc (s) = KP is only a proportional controller, then what should be the value of KP to get a
velocity error constant of 10? [Answer: 50]
Solution: With Gc (s) = KP , the open-loop transfer function will be
KP
.
s(Js + b)
2
So,
KP KP
Kv = lim s = .
s→0 s(Js + b) 5
Thus, KP = 50 will give Kv = 10.
6. With the controller obtained in Question 5, the closed-loop system is .
(a) stable [correct]
(b) unstable
(c) marginally stable
(d) cannot determine stability
KP 50
Solution: The characteristic equation of the system is 1 + s(Js+b) = 0, i.e, 1 + s(10s+5) = 0, i.e,
s2 + 0.5s + 5 = 0.
The roots of this equation are −0.25 ± j2.22, which are in the left-half of the s-plane, hence the
system is stable.
7. If Gc (s) is a PD controller with the system having friction coefficient b = 0, then what should
be the gain of the derivative mode to obtain a response with a settling time of 4 sec (with the 2
percent criterion)? [Answer: 20]
Solution: In this case, the open-loop transfer function of the system is
KP + KD s
.
10s2
KP +KD s
So, the characteristic equation is 1 + 10s2 = 0, i.e.,
KD KP
s2 + s+ = 0.
10 10
Comparing this with the standard characteristic equation, we get 2ζωn = KD /10.
4
Settling time is given by ts = ζωn = 4 sec. So, ζωn = 1, which implies KD = 20.
Td (s)
+
R(s) + Controller Automobile system Y (s)
1
Gc (s) + (s+2)(s+8)
−
3
8. It is desired to have a zero steady state error to a step input, and a steady-state error due to a
ramp input of less than 25 % of input magnitude. What type of controller could help achieve
this?
(a) A simple proportional controller
(b) A proportional plus derivative controller
(c) A proportional plus integral controller [correct]
(d) A PID Control [correct]
Solution: The open-loop system is of type 0. Since it is required to have zero steady-state error
to a step input and a bounded constant error to a ramp input, the controller must be of type 1.
PI and PID controllers satisfy this criterion.
KI
9. Suppose we employ a PI controller of the form Gc (s) = KP + s . What are the requirements
on KP , KI for the system to be stable?
(a) KI > 0, KP > 0
(b) KI > 0, KP > KI
KI
(c) KI > 0, KP > 10 − 16 [correct]
(d) KP > 0, KI > 1.6KP
Solution: With Gc (s) = KP + KsI , the characteristic polynomial of the closed-loop system is
s3 + 10s2 + (16 + KP )s + KI . To check for stability of the closed-loop system, we construct the
Routh-Hurwitz array.
s3 1 16 + KP
s2 10 KI
10(16+KP )−KI
s1 10
s0 KI
KI
For the system to be stable, we need KI > 0, 10(16+KP )−KI > 0 =⇒ KI > 0, KP > 10 − 16.
10. Give any one value of KI that will meet the steady state requirements in part (a).
Solution: The open-loop transfer function of the system is KP (s+Ks
I /KP ) 1
(s+2)(s+8) . The steady
state error to a ramp input is given by ess = 1/Kv , where Kv is the velocity error constant.
KP (s + KI /KP ) 1 KI
Kv = lim s =
s→0 s (s + 2)(s + 8) 16
For the error to be less than 25% of the magnitude, we need 1/Kv < 0.25 =⇒ Kv > 4 =⇒
KI > 64.
Hence, KI = 128 and KI = 256 are correct choices.
11. It is desired to have the overshoot of the system to be less than 5% and a settling time less than
1.5 seconds to a unit step input. The desired region for the poles of the closed-loop transfer
function is shown by
Solution:
P.O. ≤ 5% implies ζ ≥ 0.69
4
Ts ≈ ≤ 1.5 implies ζωn ≥ 2.6
ζωn
∴ −ζωn ≤ −2.6
4
(a) Correct
Figure 3: Desired region (shaded region) in the complex plane for locating the dominant system
12. Based on the location of dominant poles from the previous problem (Q 11), the values of KI and
KP must satisfy (Hint: Use root locus approach)
(a) KI /KP < 4.7 [Correct]
(b) KI /KP < 9.4
(c) KI /KP > 9.4
(d) 1.5 < KI /KP < 9.4
Solution: We want to have the dominant poles to the left of the s = −2.6 line. We know from
our experience of sketching the root locus that since we have three poles (at s = 0, −2, and − 8)
and one zero (at s = −KI /KP ), we expect two branches of the loci to go to infinity along two
asymptotes at φ = −90◦ and +90◦ centered at
P P
(−pi ) − (−zi )
σA =
np − nz
where np = 3 and nz = 1. In our case
KI
−2 − 8 − (− K ) 1 KI
σA = P
= −5 +
2 2 KP
We want to have α < −2.6 so that the two branches will bend into the desired regions.
Therefore,
1 KI KI
−5 + < −2.6 =⇒ < 4.7
2 KP KP
5
13. Let us choose the zero of the controller to be s = −2.5. Then the value of KP must satisfy (Hint:
Use root locus approach)
(a) KP < 55
(b) KP < 30 [Correct]
(c) KP < 20
(d) 20 < KP < 55
Solution:
s + K /K
I P
Gc (s) = KP
s
Given KI /KP = 2.5. Then the open-loop transfer function is
KP (s + 2.5)
G(s)Gc (s) =
s(s + 2)(s + 8)
The root locus is shown in Figure 4. To meet the ζ ≥ 0.69. We need to select KP < 30.
Root Locus
15
System: untitled1
Gain: 30.6
10 0.69 Pole: -3.51 + 3.66i
)
Damping: 0.692
-1
-5
-10 0.69
-15
-10 -8 -6 -4 -2 0 2
-1
Real Axis (seconds )
6
Question 14-15
Selecting KP = 26, the percent overshoot (in %)and settling time are and ,
respectively. [Answer: 8 ; 1.45 Range: 7-9 ; 1.3- 1.6 ] (Hint: Use MATLAB)
Solution: Given KP = 26, the closed-loop transfer function becomes
26(s + 2.5)
T (s) =
s3 + 10s2 + 42s + 65
System: untitled1
Step Response
1.2 Peak amplitude: 1.08
Overshoot (%): 8.04
At time (seconds): 0.899 System: untitled1
Settling time (seconds): 1.44
1
0.8
Amplitude
0.6
0.4
0.2
0
0 0.5 1 1.5 2 2.5
Time (seconds)
14.
P.O. = 8.04%
15.
Ts = 1.44sec
7
16. We can observe that zero of the controller does not allow the response to exactly meet the
specifications. So let us choose the zero of the controller to be s = −2 such that pole-zero
cancellation take place. Then for which of the following value(s) of KP the specifications are
met. (Hint: Use MATLAB)
(a) KP = 20 [Correct]
(b) KP = 30 [Correct]
(c) KP = 60
(d) KP = 120
8
Control Engineering
NPTEL Online Course(Jan-Apr 2019)
Indian Institute of Technology Madras
Assignment 8 Solutions
The open loop transfer function of a unity feedback system is given by
10
G(s) =
s(s + 3)
1. It is desired to design a lead compensator such that the closed loop system has a peak overshoot
of 9.478%, with a settling time of 1 sec (2% tolerance band). The angle deficiency (in degrees) is:
[Ans = 47.5 Range: 46.5-48.5]
Solution: Given ts = 1 and P.O = 9.478%. Therefore:
− √ πζ
e 1−ζ 2 = 0.09478
=⇒ ζ = 0.6
4
ts =
ζωn
4
1=
0.6 ∗ ωn
4
ωn =
0.6
ωn = 6.66
p
The closed loop poles are given by: sd = −ζωn ± jωn 1 − ζ 2 = −4 ± j5.33. The angle deficiency
can be calculated using the angle criterion of the root locus.
2. What should be the location of the lead compensator zero? [Ans: -4.2638, Accepted range:
-4.5 to -4.0]
Solution: First we need to calculate θ and γ. Taking ζ from the characteristic equation of dominant
poles, we get
θ = cos−1 (ζ) = cos−1 (0.6) = 53.12o
1
Therefore,
1
γ= (180 − θ − φ)
2
1
γ = (180 − 53.12 − 47.5)
2
γ = 39.7o
Solution: The open loop transfer function of compensated system is given by:
(s + 4.2638) 10
Gc (s)G(s) = Kc
(s + 10.4246) s(s + 3)
The compensator gain Kc should be such that the closed loop poles of the system are at −4 + j5.33.
Using the magnitude condition of the root locus:
(sd + 4.2638) 10
Kc =1
(sd + 10.4246) sd (sd + 2)
(0.2638 + j5.33) 10
Kc
=1
(6.4246 + j5.33) (−4 + j5.33)(−1 + j5.33)
=⇒ Kc = 5.62
This can also be verified through the root locus plot in MATLAB, see Figure 1.
MATLAB Code:
2
Root Locus
25
20 System: untitled1
Gain: 5.63
15 Pole: -3.99 + 5.31i
)
Damping: 0.6
-1
-5
-10
-15
-20
-25
-12 -10 -8 -6 -4 -2 0 2
-1
Real Axis (seconds )
Figure 1: Q4
5. What is the velocity error constant of the compensated system? [Ans: 7.66, Accepted range:
7 to 8]
(s + 4.2638) 10
Kv = lim sKc
s→0 (s + 10.4246) s(s + 3)
5.62 × 4.2638 × 10
Kv =
10.4246 × 3
Kv = 7.66
6. What should be the lead compensator gain (with poles and zeros as above) to achieve a peak over-
shoot of 16.3%? [Answer = 8.36, range 8.1 - 8.5]
Solution: Given P.O. = 16.3% implies ζ = 0.5, from root locus plot of G(s)Gc (s) we get Kp = 8.36
3
Root Locus 25
25
0.44
0.32 0.23 0.16 0.1 0.05
20
20 System: sys1
Gain: 8.36
0.6 Pole: -4.28 + 7.41i 15
15
Damping: 0.5
Overshoot (%): 16.3 10
)
-1
5
5
-5
5
0.84
-10
10
-15 0.6 15
-20
20
0.32 0.23 0.16 0.1 0.05
0.44
-25
-12 -10 -8 -6 -4 -2 250 2
-1
Real Axis (seconds )
Figure 2: Q6-Q7
7. With this gain adjustment what will be the new settling time? [Answer = 0.9356, range 0.9 -
0.97]
8. What is the β of the lag compensator that would achieve this [Answer = 10, range 9 - 11]
Solution: The lead compensated system as above has a Kv of 7.66 and desired value is K̂v = 76.6.
Therefore,
K̂v
β=
Kv
76.6
β=
7.66
β = 10
4
9. Choose the lag compensator zero at −0.1 and place the corresponding pole of the lead compensator.
The overall root locus
(a) Shifts slightly to the right [Correct answer]
(b) Shifts slightly to the left
Solution: Given zc = −0.1, we know that:
zc
β=
pc
−0.1
10 =
pc
pc = −0.01
MATLAB Code:
>> Gc2=tf([1 0.1],[1 0.01]); % Lag compensator
>> sys1=G*Gc; % Lead compensated system
>> sys2=G*Gc*Gc2; % Lead-Lag compensated system
>> rlocus(sys1);
>> hold on;
>> rlocus(sys2);
Root Locus
25
sys1
20 sys2
15
)
-1
10
Imaginary Axis (seconds
-5
-10
-15
-20
-25
-12 -10 -8 -6 -4 -2 0 2
-1
Real Axis (seconds )
Figure 3: Q9
5
Control Engineering
NPTEL Online Course(Jan-Apr 2019)
Indian Institute of Technology Madras
Assignment 9 Solutions
We are required to design a lead compensator, such that the position error constant Kp = 10
and the phase margin is atleast 25 degrees.
(a) What is the value of K that meets the steady state requirements. [Answer K = 10].
(b) What is the phase margin of the gain compensated system Answer = 17.2 degrees,
Range = 16 - 18
(c) Allowing for an extra margin of 10 degrees, what is the corresponding α of the lead
compensator Answer = 0.5318, Range = 0.48 to 0.58
(d) What is the magnitude of the compensator (in dB) given the frequency at which the
lead compensator provides a maximum phase is 1.83 rad/s? Answer = 2.74, Range
= 2.4 to 3.1
1+sT
(e) For a lead compensator of the form 1+sαT , what is the value of T . answer = 0.7494,
range = 0.7 to 0.8.
(f) What is the gain margin of the compensated system (in dB) Answer = 6.87, Range
= 6.5 to 7.2
Solution:
(a)
Kp = lim KG(s) = K = 10
s→0
(b) First, we find the gain crossover frequency ωgc . We know, at ω = ωgc , |G(jω)| = 1. So,
30
q =1
2
(ωgc + 2
9)(ωgc 2 + 1)
+ 1)(9ωgc
6 4 2
9ωgc + 91ωgc + 91ωgc − 891 = 0
2
ωgc = 2.435
ωgc = 1.560
PM = 180 + ∠G(jωgc )
= 180 + (− tan−1 (ωgc /3) − tan−1 (ωgc ) − tan−1 (3ωgc ))
= 180 − 162.766
= 17.234
1
(c) Required phase margin, with the extra margin is 25 + 10 = 35 degrees. Phase mar-
gin of the gain compensated system as determined above is 17.2 degrees. So, phase
contribution from the compensator is 35 - 17.2 = 17.8 degrees. So,
1 − sin(17.8)
α= = 0.5318
1 + sin(17.8)
(d) The magnitude of the compensator at maximum phase lead is
√
20 log(1/ α) = 2.74
(e)
1 1
T = √ = √ = 0.749
ωm α 1.83 0.5318
.
(f) The compensated system is
(10/α)(s + 1/T ) 10(s + 1.33)
Gc (s)G(s) = =
(s + 1/αT )(s/3 + 1)(s + 1)(3s + 1) (s + 2.51)(s/3 + 1)(s + 1)(3s + 1)
The phase crossover frequency is given by solving
tan−1 (ω/1.33) − tan−1 (ω/2.51) − tan−1 (ω/3) − tan−1 (ω) − tan−1 (3ω) = −180,
which gives ωpc = 2.79 rad/sec.
Now, the gain margin of the system is
q
(10/0.5318) 2 + 1.333
ωpc
G.M. = 20 log q = 6.87 dB.
2 + 2.512 )(ω 2 /9 + 1)(ω 2 + 1)(9ω 2 + 1)
(ωpc pc pc pc
(b) What is the value of the compensator gain needed to achieve the requirements on the
steady state? [Ans: 10]
(c) Considering β = 9.12 and T = 10 what is the phase margin of the lag compensated
system? [Ans: 43.8], [accepted range: 40 - 47]
(d) What is the frequency (in rad/s) at which the maximum phase lag occurs in the com-
pensator? (upto three decimal places) [Ans:0.033], [accepted range: (0.025,0.040)]
Solution:
(a) The gain crossover frequency of the uncompensated system is given by solution of
|G(jωgc )| = 1
1
| q |=1
ωgc ωgc2 +1
4 2
ωgc + ωgc −1=0
ωgc = 0.786 rad/sec
Phase margin of the system is given by
180 + ∠G(jωgc ) = 180 + (−90 − tan−1 (ωgc )) = 51.828.
2
(b) First step in compensator design is to adjust gain which meets steady state require-
ments.(Refer week-9,lecture-3 slides).
The standard lag compensator form is given by,
(1+jωT )
Gc = βk (1+jβωT )
(1+jωT )
=⇒ Gc = K 0 (1+jβωT )
Kv = lim sK 0 Gc (s)G(s)
s→0
= K 0 lim sG(s)
s→0
we have, lim sG(s) = 1
s→0
10(s + 1/10) 1
Gc (jω)G(jω) = .
(s + 1/91.2) s(s + 1)
The gain crossover frequency of the lag compensated system is given by solving
|Gc (jω)G(jω)|ω=ωgc = 1
q
10 ωgc2 + 1/100
q =1
2 + 1/91.22 )(ω 2 + 1)
ωgc (ωgc gc
ωgc = 0.8438.
180 + ∠Gc (jωgc )G(jωgc ) = 180 + (tan−1 (10ωgc ) − 90 − tan−1 (ωgc ) − tan−1 (91.2ωgc ))
= 180 − 136.17 = 43.83
3
Control Engineering
NPTEL Online Course (Jan-Apr 2019)
Indian Institute of Technology Madras
Assignment 10
2
1. Consider the transfer function G (s) = (s+1)(s+2)
, The phase margin of G (s) in degrees is (Accepted
answer : 180)
Solution:
Calculate ω gc
2
=1
( jω + 1)( jω + 2)
4
√ √ =1
ω + 1 ω2 + 4
2
On solving, we get √
ω gc = 0 rad/sec and ω gc = ± j 5 (invalid ?)
Thus, PM = 180◦ − φ
where φ = − tan−1 ω gc − tan−1 ω gc = 0◦
PM = 180◦
2. The Bode magnitude plot of the transfer function (For Q2- Q4)
K (1 + as)(1 + 0.25s)
G (s) =
s(1 + 8s )(1 + bs)(1 + 36
s
)
What is the value of a/b ? [Ans = 12] (Allow answers between 11 and 13)
Solution:
Given Bode plot as in Figure the corner frequencies are ω1 = 2, ω2 = 4, ω3 = 8, ω4 = 24, ω5 =
36 Clearly, zeros are at ω1 = 2 and ω2 = 4. Poles are at ω3 = 8, ω4 = 24 and ω5 = 36.
1
1 1 1
So, we have a = ω1 = 0.5 and b = ω4 = 24 = 0.04.
1/2
∴ a/b = = 12
1/24
3. What is the value of K rounded to 2 decimal points? [Ans = 8] (Allow answers between 7 and
9)
Solution:
The given system is type-1 system. The intersection of the initial -20dB/dec segment(or its
extension) with the 0dB line has frequency numerically equal to Kv . (i.e. Kv /jω1 = 1)
so, here we have Kv = ω3 = 8 = K.
80
M agnitude
dB)
60
40
20
0
P hase
Deg)
180
135
90
45
0
102 101 10 101 102 103 104
F requency
radsec)
( s −1)2
(a) 104 (s+100)2
( s +1)2
(b) 104 (s+100)2 [correct]
( s −1)
(c) 104 (s+100)2
( s −1)2
(d) 104 (s−100)2
Solution:
For detailed solutions use the steps provided for Q-9
Analyse the phase diagram of the asymptotic bode plot using the poles and zeros of all the sys-
tems given in the options.
2
The magnitude condition for all the option remain same (?)
* Effect of each negative zero gives positive phase of 90◦ (increasing from 0◦ to 90◦ )
* Effect of each negative pole give negative phase of 90◦ (decreasing from 0◦ to -90◦ )
You can verify the above two statements by taking some arbitrary poles and zeros in RHP and
LHP and analyse their Phase diagram (Use MATLAB to simplify your work)
So based on above two statements we can see that phase diagram can be of 2 negative zeros and
2 negative poles given in option(b)
Here Negative zero implies zeros in left hand side of s-plane and similarly negative poles im-
plies poles lying in left hand side of s-plane
6. Which of the following transfer function is/are minimum phase transfer function(s)
1
(a) ( s −1)
( s −1)
(b) (s+3)(s+4)
( s +2)
(c) (s+3)(s−4)
(d) None [correct]
Solution:
All transfer function have atleast one zero or pole is in right half s-plane, hence all are non
minimum phase system.
(1− s )
7. The transfer function of a system is (1+ s )
. The system is then which one of the following
Gain dB)
20dB=dec
26dB
20dB=decade
20dB
log
0dB 10 2
Figure 3: C
3
Solution:
26−20
Slope = 20 = log ω1 −log 10 =ω = 19.9526
26−0
−20 = log 20−log ω2 =ω2 = 398.1071
Thus, ω2 − ω1 = 378
9. The asymptotic approximation of the log-magnitude versus frequency plot of a system is shown
in Figure 4.
Solution:
Step1:
The initial slope of the Bode plot is -40dB/dec which implies that system is of Type 2.
Step2:
The corner frequencies are : 2, 5 and 25.
The slope of Bode plot just after ω = 2 rad/s is -60 dB/dec. Hence, ω = 2 rad/s is the pole of the
system.
The slope of Bode plot just after ω = 5 rad/s is -40 dB/dec. Hence, ω = 5 rad/s is the zero of the
system.
The slope of Bode plot just after ω = 25 rad/s is -60 dB/dec. Hence, ω = 25 rad/s is the pole of
the system.
Step3:
4
Gain of the system
5K
54 = 20 log( ) =⇒ 10K = 102.7 =⇒ K = 101.7 = 50
50ω 2
Step4:
50(s + 5)
T.F. =
s2 (s + 2)(s + 25)
10. If an asymptotically stable transfer function possesses at least one positive zero, then the step
response of the system undergoes at least one zero crossing.