ME 413 Systems Dynamics & Control Chapter 9: Frequency Domain Analyis of Dynamic Systems Systems

Chapter 9: 
Frequency‐Domain Analysis of 
Dynamic Systems 
A. Bazoune 

9.1 INTRODUCTION
The term Frequency Response refers to the steady state response of a system to a sinusoidal
input.
An input f ( t ) is periodic with a period τ if f ( t + τ ) = f ( t ) for all values of time t ,
where τ is a constant called the period. Periodic inputs are commonly found in many
applications. The most common perhaps is ac voltage, which is sinusoidal. For the common
1
ac frequency of 60 Hz, the period is τ = s . Rotating unbalanced machinery produces
60
periodic forces on the supporting structures, internal combustion engines produce a periodic
torque, and reciprocating pumps produce hydraulic and pneumatic pressures that are
periodic.

Frequency response analysis focuses on sinusoidal inputs. A sine function has the
form A sin ωt , where A is the amplitude and ω is its frequency in radians/seconds. Notice
⎛ π⎞
that a cosine is simply a sine shifted by 90o or π 2 rad, as cos ωt = sin ⎜ ωt + ⎟
⎝ 2⎠

9.2 SINUSOIDAL TRANSFER FUNCTION (STF)

When a sinusoidal input is applied to a LTI system, the system will tend to vibrate at its own
natural frequency, as well as follow the frequency of the input. In the presence of damping, that
portion of motion sustained by the sinusoidal input will gradually die out. As a result, the
response at steady-state is sinusoidal at the same frequency as the input. The steady-state
output differs from the input only in the amplitude and the phase angle. See Figure 9-1 below.

ω3t
A sin4
142 X sin (ω t + φ )
14442444 3
Inpu t Output

Figure 9-1.

1/27
ME 413 Systems Dynamics & Control Chapter 9: Frequency Domain Analyis of Dynamic Systems Systems

Thus, the output-input amplitude ratio and the phase angle between the output and input
sinusoids are the only parameters needed to predict the steady state output of LTI systems
when the input is a sinusoid.

X
: output-input amplitude ratio
A
φ : phase angle between output and input

Forced Vibration without damping Figure 9-2 illustrates a spring-

mass system in which the mass is subjected to a
sinusoidal input force p (t ) = Po sin ωt . Let us find the
response of the system when it is initially at rest. The
equation of motion is k
p(t ) = Po sin ωt
x + k x = Po sin ωt
m &&
m
where x is the output, P is the amplitude of the
excitation and ω is the forcing (excitation) frequency. x
Figure 9-2 Spring-mass system.
The above equation can be written in the form

k P
x+
&& x = o sin ωt (9-1)
m m

where k m = ωn is known as the natural frequency of the system. The solution of Equation
(1) consists of the vibration at its natural frequency (the complementary solution) and that at
the forcing frequency (the particular solution) as shown in Figure 9-3. Thus,

x ( t ) = complementary solution + particular solution

Let us obtain the solution under the condition that the system is at rest. Take LT of both sides
of Equation (9-1) for zero initial conditions, i.e., x ( 0 ) = x& ( 0 ) = 0 .

⎛ 2 k⎞ Po ω
⎜ s + ⎟ X (s) =
⎝ m⎠ m s2 + ω 2

where X ( s ) = L [ x(t ) ] . Substituting k m = ωn2 and Solving for X (s) yields

Po ω 1
X (s) =
m s + ω s + ωn2
2 2 2

The above equation can be written in partial fraction as:

Po ω 1 A s + B1 A2 s + B2
X (s) = = 12 + 2
m s + ω s + ωn s + ω 2
2 2 2 2
s + ωn2

2/27
ME 413 Systems Dynamics & Control Chapter 9: Frequency Domain Analyis of Dynamic Systems Systems

where A1 , A2 , B1 and B2 are left as an exercise for the student.

The expression for X ( s ) is therefore

Poω ⎛ 1 ⎞ ωn Po ω
X (s) = − ⎜ ⎟ 2 +
k − mω 2 ⎝ ωn ⎠ s + ωn k − mω s + ω
2 2 2 2

The inverse Laplace Transform of the above equation is given by

P ( ω ωn ) Po
[ x (t )] = L -1
⎡⎣ X ( s ) ⎤⎦ = − o sin ωnt + sin ωt
k −
14243 m ω 2
k −
1424m ω3
2

A B (9-2)
= A sin ωnt + B 1 in ω
4s24 3t
1424 3
Complementary Solution Particular Solution

where
Po (ω ωn ) Po (ω ωn ) Po P
A =− = and B= = o
k − mω 2 Den k − mω 2
Den

where Den = k − m ω .
2

⎛P ⎞
• As ω →0, lim A = 0 lim B = ⎜ o ⎟ ≡ static deflection
and
ω →0 ω →0 ⎝k ⎠
As ω increases from zero the denominator Den = k − m ω becomes small and the
2
•
amplitude increases, therefore, both A and B increase.
• The expression of the denominator Den can be written as
⎡k 2⎤
⎡ ω 2⎤
Den = k − m ω = m ⎢ − ω ⎥ = m ⎡⎣ωn − ω ⎤⎦ = m ωn ⎢1 − 2 ⎥
2 2 2 2

⎣m ⎦ ⎣ ωn ⎦
• It is clear that when ω = ωn the denominator becomes zero and the amplitude of
vibrations increases without bound, therefore resonance occurs.

Sinusoidal Transfer Function (STF) The sinusoidal Transfer Function

(STF) is defined as the transfer function G (s ) in which the variable (s ) is replaced by

( jω).

G (s ) s = jω G ( jω )

1442443 1442443
TF STF

Figure 9-3 Sinusoidal Transfer Function STF

3/27
ME 413 Systems Dynamics & Control Chapter 9: Frequency Domain Analyis of Dynamic Systems Systems

When only the steady-state solution (the particular solution) is wanted, the STF
G ( jω) can simplify the solution. In our discussion, we are concerned with the behavior of
stable, LTI system under steady state conditions, i.e., that is after the initial transients died
out. We shall see that sinusoidal inputs produce sinusoidal outputs in the steady state with
the amplitude and phase angle at each frequency ω determined by the magnitude and the
angle of G ( j ω ) , respectively.
Deriving Steady State Output caused by Sinusoidal Input
Figure 9-4 shows an LTI system for which the input P (s ) and the output is X (s ) .
p (t ) = Po sin ωt
G (s )
P (s ) X (s )
1 4 42 4 43
TF

Figure 9-4 Linear Time Invariant (LTI) System

The input p (t ) is sinusoidal and is given by

p (t ) = Po sin ωt

We shall show that the output x (t ) at steady state is given by

x (t ) = G ( j ω ) Po sin (ωt + φ )

where G ( jω) and φ are the magnitude and phase angle of G ( j ω ) , respectively.
Suppose that the transfer function G ( s ) can be written as a ratio of two polynomials
in s ; that is

K ( s + z 1 )( s + z 2 )L( s + z n )
G (s ) =
( s + s1 )( s + s 2 )L( s + s n )
The Laplace transform X (s ) is

X (s ) = G (s ) P (s ) (9-3)
where P ( s ) = L [ p(t )] .

Let us limit our discussion to stable systems. For such systems, the real parts of the
−s i are negative. The steady state response of a stable linear system to a sinusoidal input
does not depend on I. C’s, so they can be ignored.

4/27
ME 413 Systems Dynamics & Control Chapter 9: Frequency Domain Analyis of Dynamic Systems Systems

• If G ( s ) has only distinct poles, then the partial fraction of Equation (9-3) yields
Pω
X (s ) = G (s )
s 2 + ω2
(9-4)
a a b b b
= + + 1 + 2 +L+ n
s + j ω s − j ω s + s1 s + s 2 s + sn

where a and bi ( i = 1,2,L, n ) are constants and a is the complex conjugate of a . The
response x (t ) can be obtained by taking the inverse Laplace transform of Equation (9-4)

[ x (t )] = L -1
⎡⎣ X ( s ) ⎤⎦ = ae − j ωt + a e j ωt + b1 e −s1 t + b 2 e 2 + L + b n e n
144444244444
−s t
3
−s t

For a stable system these terms →0 as t →∞

since they have negative real part

• If G (s ) involves k multiple poles sj , then x (t ) will involve such terms as

−s j t
t he (where h = 0,1, 2,L, k − 1). Since the real part of the −s j is negative for a
−s j t
stable system, the terms t he →0 when t → ∞.
Regardless of whether the system involves multiple poles, the steady state response becomes

x (t ) = ae − j ωt + a e j ωt (9-5)

where the constants a and a can be evaluated from Equation (9-4):

Pω P
a = G (s ) 2 (
s + jω) = − G (− j ω )
s +ω
2
s =− j ω 2j
Pω P
a = G (s ) ( s − j ω ) = G ( jω)
s 2 + ω2 s = jω 2j

Notice that a is the complex conjugate of a . Referring to Figure 9-5, we can write

( j ω)

Gy G ( jω)

φ
σ
−φ Gx

G( −jω)

Figure 9-5 Complex function and its complex conjugate.

5/27
ME 413 Systems Dynamics & Control Chapter 9: Frequency Domain Analyis of Dynamic Systems Systems

G ( j ω ) = Gx +G y
= G ( j ω ) cos φ + j G ( j ω ) sin φ
= G ( j ω ) ( cos φ + j sin φ )
= G ( j ω ) e jφ

Notice that ∠G ( j ω ) = ∠e j φ = φ . Similarly,

G ( − j ω ) = G ( − j ω ) e jφ = G ( j ω ) e jφ

Substitute the expressions of G ( jω) and G (− j ω ) into the expressions of a and a , one
can get
P
a=− G ( j ω ) e − jφ
2j
P
a= G ( − j ω ) e jφ
2j
Then Equation (9-5) can be written as

j (ωt +φ )
−e ( )
− j ωt +φ
e
x (t ) = G ( j ω ) Po
2j
1442443
sin (ωt +φ )

= G ( j ω ) Po sin (ωt + φ ) (9-6)

= X sin (ωt + φ )

where X = G ( j ω ) Po and φ = ∠G ( j ω )

Same frequency

p ( t ) = Po sin ωt x(t ) = X sin (ω t + φ )

Input G ( j ω) Output

X = P G ( jω )
o

φ = ∠ G ( jω )
Output amplitude

Phase of the output

Figure 9-6 Input output relationships for sinusoidal inputs.

6/27
ME 413 Systems Dynamics & Control Chapter 9: Frequency Domain Analyis of Dynamic Systems Systems

Therefore for sinusoidal inputs,

X ( j ω ) amplitude ratio of the output

G ( jω) = = (9-7)
P ( j ω ) sinusoid to the input sinusoid

X ( jω) ⎡ imaginary part of G ( j ω ) ⎤

∠G ( j ω ) = ∠ = tan −1 ⎢ ⎥ (9-8)
P ( jω) ⎣ real part of G ( j ω ) ⎦
phase shift of the output sinusoid
=
with respect to the input sinusoid

█ Example 9-1 (Textbook Page 437)

Consider the TF
X (s) 1
= G(s) =
P(s) Ts + 1

For the sinusoidal input p(t ) = Po sin ωt , what is the steady-state output x ( t ) .

█ Solution

Substituting jω for s in G (s) yields

1
G ( jω ) =
Tjω + 1
The output-input amplitude ratio is

1
G ( jω ) =
T ω2 +1
2

and the phase angle φ is

φ = ∠G ( jω ) = − tan −1 T ω

So, for the input p (t ) = Po sin ωt , the steady-state output x ( t ) can be found as

sin (ωt − tan −1 T ω )

Po
x(t ) = (9-9)
T ω +1
2 2

█ Example 9-2

Find the steady state response of the following system:

7/27
ME 413 Systems Dynamics & Control Chapter 9: Frequency Domain Analyis of Dynamic Systems Systems

y& + 5 y = 4 p& + 12 p

if the input is p (t ) = 20sin 4t

█ Solution

First obtain the TF

Y ( s ) 4s + 12 s +3
G (s ) = = =4
P (s ) s +5 s +5

From the input p (t ) = 20sin 4t , it is clear that ω = 4 rad/s. Therefore, the sinusoidal
Transfer function is

Y ( jω) jω + 3 j4+3 3+ j 4
G ( jω) = =4 =4 =4
G ( jω) jω + 5 j4+5 5+ j4
Then

32 + ( 4 )
2
3+ j 4 25
G ( jω) = 4 =4 =4 = 3.123
5+ j4 52 + ( 4 )
2
41
and
⎛ 3 + jω ⎞
φ = ∠G ( j ω ) = ∠ ⎜ 4 ⎟ = ∠4 + ∠ ( 3 + j ω ) − ∠ ( 5 + j ω )
⎝ 5 + jω ⎠
ω ω 4 4
= 0o + tan −1 − tan −1 = 0o + tan −1 − tan −1 = 0.253 rad
3 5 3 5
The steady state response is

y (t ) = G ( j ω ) Po sin (ωt + φ )
= 3.123 × 20sin ( 4t + 0.253) = 62.46sin ( 4t + 0.253)

█ Example 9-3 (Example 9-2 in the Textbook

Page 437-438)

Suppose that a sinusoidal force p (t ) = Po sin ωt

is applied to the mechanical system shown in Figure
9-7. Assuming that the displacement x is measured
from the equilibrium position, find the steady-state
output.

█ Solution

The equation of motion for the system is

x + bx& + k x = p(t )
m &&
Figure 9-7 Mechanical system

8/27
ME 413 Systems Dynamics & Control Chapter 9: Frequency Domain Analyis of Dynamic Systems Systems

The Laplace Transform of this equation, assuming zero I.C’s, is

(m s 2
+ bs + k ) X ( s ) = P ( s )

where X ( s ) = L [ x(t ) ] and P ( s ) = L [ p(t )] . (Notice that the I.C’s do not affect the
steady state output and so can be taken to be zero). The TF is

X (s) 1
= G (s) =
P( s) ( m s 2 + bs + k )
Since the input is a sinusoidal function p (t ) = Po sin ωt , we can use the STF to obtain the
steady-state solution. The STF is

X ( jω ) 1 1
= G ( jω ) = =
P ( jω ) − m ω + bjω + k ( k − m ω 2 ) + jbω
2

From Equation (9-6), the steady-state output x ( t ) can be written as

x(t ) = G ( jω ) Po sin (ωt + φ )

where
1
G ( jω ) =
(k − m ω ) 2 2
+ b 2ω 2
and
1 bω
φ = ∠G ( jω ) = ∠ = − tan −1
( k − m ω ) + jbω2
k − m ω2

therefore

Po ⎛ bω ⎞
sin ⎜ ωt − tan
−1
x (t ) = ⎟
(k − m ω ) ⎝ k −m ω ⎠
2 2
+b ω
2 2 2

Since ωn2 = k m and b / k = 2ζ / ωn , the equation for x ( t ) can be written as

⎛ b
ω
⎞
Po / k ⎜ k ⎟
sin ⎜ ωt − tan
−1
x (t ) =
k m 2⎟
⎛ k − m ω2 ⎞ + b ω2 ⎜
2
− ω ⎟
2

⎜ ⎟ ⎝ k k ⎠
⎝k k ⎠ k 2

or
Po / k ⎛ 2ζ ( ω ωn ) ⎞
sin ⎜ ωt − tan
−1
x (t ) = ⎟
1 − ( ω ωn ) ⎠
2

⎡⎣1 − ( ω ωn )2 ⎤⎦ + ( 2ζ ( ω ωn ) )2 ⎝
2

9/27
ME 413 Systems Dynamics & Control Chapter 9: Frequency Domain Analyis of Dynamic Systems Systems

Let β = ω ωn ≡ frequency ratio , the above equation can be written as

x st ⎛ 2ζβ ⎞
sin ⎜ ωt − tan
−1
x (t ) = ⎟ (9-10)
⎝ 1− β ⎠
2

⎡⎣1 − β ⎤⎦ + ( 2ζβ )
2 2 2

where xst = Po k is the static deflection. Writing the amplitude of x ( t ) as X , we find that
the amplitude ratio X / xst and the phase shift φ are

X 1 2ζβ
= and φ = − tan −1
1− β
2
x st ⎡⎣1 − β 2 ⎤⎦ + ( 2ζβ )
2 2

The variations of the amplitude ratio X / xst and the phase shift φ are shown in figures 9-8
and 9-9 as a function of β for different values of ζ .

Frequency Response Magnitude Ratio

15

ζ = 0.0

0.05
10

0.1
X/xst

0.25
0.50

0
5.0 1.00
2.0

-5
0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2
β = ω /ωn

Figure 9-8 Variation of the amplitude ratio X / xst with the frequency ratio β.

10/27
ME 413 Systems Dynamics & Control Chapter 9: Frequency Domain Analyis of Dynamic Systems Systems

Frequency Response Phase Angle

π

05
0.
1
ζ = 0.0 0.
0.25

0.50 1.00

2.0

5.0

φ(ω) π/2

ζ = 0.0

0
0 0.5 1 1.5 2 2.5 3
β = ω / ωn

Figure 9-9 Variation of the phase φ with the frequency ratio β.

11/27