Modeling in Frequency Domain

Control Theory

Author Al-Motasem Aldaoudeyeh, PhD

Institute Tafila Technical University

Date March 8, 2023

Al-Motasem Aldaoudeyeh, PhD Slide 1 of 242 BasmahWantsMeHappy

 Laplace Transform Review
▶ The most general form of Laplace transform of any function in the time domain f (t)
is the two-sided Laplace transform [HKD12; p. 541]
Z∞
L [f (t)] = F (s) = f (t) e−st dt (1)
−∞

where the symbol L is used to denote that we want to apply Laplace transform to its
argument f (t)

Al-Motasem Aldaoudeyeh, PhD Slide 2 of 242 BasmahWantsMeHappy

The One-Sided Laplace Transform [HKD12; p. 542]
▶ In control systems, forcing functions (e.g., current-sources or mechanical forces) do
not exist forever in time. Thus, we think of the response to have started at specific
instant t = 0. Thus, Eq. (1) becomes
Z∞
L [f (t)] u(t) = F (s) = f (t) e−st u(t)dt
−∞

Al-Motasem Aldaoudeyeh, PhD Slide 3 of 242 BasmahWantsMeHappy

The One-Sided Laplace Transform [HKD12; p. 542]
▶ Thus, the Laplace transform we study here is

Z∞
L [f (t)] u(t) = F (s) = f (t) e−st dt (2)
0−

where s is a complex variable given as

s = σ + jω

Al-Motasem Aldaoudeyeh, PhD Slide 4 of 242 BasmahWantsMeHappy

Invrse of Laplace Transforms [HKD12; p. 542]
▶ The function F (s) in Eq. (2) may also be converted back to its original time-domain
representation using the integral
σ0Z+j∞
−1 1
f (t) u(t) = L F (s)] =
[ est F (s) ds (3)
πj
2
σ0 −j∞

where u(t)is the unit-step function, which is given as

(
0 t<0
u(t) =
1 t>0

Al-Motasem Aldaoudeyeh, PhD Slide 5 of 242 BasmahWantsMeHappy

Common Laplace Transform Pairs [Nis11; p. 36]
▶ The following table shows the most common Laplace transform pairs

Item Item
f (t) F (s) f (t) F (s)
No No

1. δ (t) 1 1
5. e−at u(t)
1 s+a
2. u(t) ω
s 6. ω t) u(t)
sin (
1
s + ω2
2

3. t u(t) s
s2 7. cos ( ω t) u(t)
s2 + ω 2
n!
4. tn u(t)
sn+1

Al-Motasem Aldaoudeyeh, PhD Slide 6 of 242 BasmahWantsMeHappy

Common Laplace Transform Pairs
▶ The following table shows the most common Laplace transform pairs

Item
f (t) F (s)
No

1
1. t e−at u(t)
(s + a)2
s + a)
(
2. A e−at cos (ω t) u(t) A 2 2
(s + a) + ω

ω
3. B e−at sin (ω t) u(t) B 2 2
(s + a) + ω

Al-Motasem Aldaoudeyeh, PhD Slide 7 of 242 BasmahWantsMeHappy

Laplace Transform Theorems
Linearity Theorem [Nis11; p. 37]
▶ Let f (t), f1 (t), and f2 (t) be functions whose Laplace transforms exist, then
L [f1 (t) ± f2 (t)] = F1 (s) ± F2 (s) (4)
L [kf (t)] = kF (s) (5)

Al-Motasem Aldaoudeyeh, PhD Slide 8 of 242 BasmahWantsMeHappy

Laplace Transform Theorems
Linearity Theorem [Nis11; p. 37]
▶ Eq. (4) is called the additivity property.
d it means that the Laplace of two functions being added or subtracted from each other is
the result of taking the Laplace of each function alone then adding/subtracting the
individual Laplace transforms

▶ Eq. (5) is called the homogeneity property

d it means that the Laplace of any function f (t) scaled by a constant k is simply the result of
taking the Laplace of this function F (s) then multiplying it by k

Al-Motasem Aldaoudeyeh, PhD Slide 9 of 242 BasmahWantsMeHappy

Laplace Transform Theorems
Differentiation Theorem [Nis11; p. 37]
▶ Let f (t) be a continuous on [0, ∞) where f ′ (t) be continuous on [0, ∞), then
 
df −
L = sF (s) − f (0 ) (6)
dt
 2 
d f 2 − ′ −
L = s F (s) − sf (0 ) − f (0 ) (7)
d t2

▶ In a more general (less useful for our analysis), the differentiation property is
 n 
d f n−1
L n
= s F (s) − s f (0− ) − sn−2 f ′ (0− ) − · · · − f (n−1) (0− ) (8)
d tn

Al-Motasem Aldaoudeyeh, PhD Slide 10 of 242 BasmahWantsMeHappy

Laplace Transform Theorems
Differentiation Theorem [Nis11; p. 37]
▶ Eqs. (6) to (8) show us that the differentiation theorem is probably the most useful
theorem for engineering analysis purposes
d the differentiation with respect to t in the time domain is replaced with multiplication by s in the
frequency domain, converting a differential equations into algebraic ones

Al-Motasem Aldaoudeyeh, PhD Slide 11 of 242 BasmahWantsMeHappy

Laplace Transform Theorems
Integration Theorem [Nis11; p. 37]
▶ Let f (t) be a continuous function on [0, ∞) where it is integrable, then
 
Zt
 1
L  f (τ ) dτ  = F (s) (9)

s
0−

▶ The integration theorem shows that integration in the time domain is replaced with
dividing over s. This also helps us convert equations containing integrals into algebraic
equations

Al-Motasem Aldaoudeyeh, PhD Slide 12 of 242 BasmahWantsMeHappy

Laplace Transform Theorems
Frequency Shift Theorem [Nis11; p. 37]
▶ This theorem states that for t ≥ 0
L f (t)e−at u(t) = F (s + a)
 
(10)

▶ This means that when f (t) is multiplied by e−at , then we simply calculate F (s) for
f (t) (usually using the tables), then replace every s with s + a

Al-Motasem Aldaoudeyeh, PhD Slide 13 of 242 BasmahWantsMeHappy

Laplace Transform Theorems
Time Shift Theorem [Nis11; p. 37]
▶ This theorem states that for t ≥ 0
L [f (t − a)u(t − a)] = e−as F (s) (11)

▶ This means that when every t is replaced with t − a, then we simply calculate F (s)
for f (t) (usually using the tables), then multiply the result by e−as

Al-Motasem Aldaoudeyeh, PhD Slide 14 of 242 BasmahWantsMeHappy

Laplace Transform Theorems
Initial Value Theorem This theorem states that when f (t) and its first derivative are
transformable then [Nis11; p. 37]
lim f (t) = lim sF (s) (12)
t→0+ s→∞

▶ Here, note that as t approaches 0+ in time domain, s approaches ∞ in frequency

domain

Al-Motasem Aldaoudeyeh, PhD Slide 15 of 242 BasmahWantsMeHappy

Laplace Transform Theorems
Final Value Theorem This theorem states that when f (t) and its first derivative are
transformable, then it is possible to calculate the final value of the time domain response
as follows [Nis11; p. 37]
lim f (t) = lim sF (s) (13)
t→∞ s→0

▶ Here, note that as t approaches ∞ in time domain, s approaches 0 in frequency

domain
▶ Thus, to calculate the final value of a response using a transfer function, multiply the
transfer function by s. Afterwards, substitute 0 for every s

Al-Motasem Aldaoudeyeh, PhD Slide 16 of 242 BasmahWantsMeHappy

Laplace Transform Theorems
Final Value Theorem Three important conditions must be satisfied for this theorem to
hold
▶ No poles contain pure imaginary parts (e.g., sp = ±jω)
▶ At most one pole at the origin (i.e., at most one pole whose value is sp )
= 0

▶ All other poles must have negative real parts

Al-Motasem Aldaoudeyeh, PhD Slide 17 of 242 BasmahWantsMeHappy

Summary of Important Concepts
▶ Laplace transform helps us convert an algebraic equation into differential equation
▶ The conversion from time domain to frequency domain may be easily done by using
the Common Laplace Transform Pairs and Laplace Transform Theorems

Conversion from Time to Frequency Domain

▶ Avoid using the = sign when converting between time and frequency domains. For
example
1
e−at u(t) = incorrect
s+a
1
e−at u(t)⇔ correct
s+a

Al-Motasem Aldaoudeyeh, PhD Slide 18 of 242 BasmahWantsMeHappy

Laplace Transform Theorems [Nis11; p. 98]
Obtain the Laplace transform for the following time-domain function

f (t) = e−a t ω t) u(t)

cos (

Al-Motasem Aldaoudeyeh, PhD Slide 19 of 242 BasmahWantsMeHappy

Laplace Transform Theorems [Nis11; p. 98]
▶ From Laplace transform tables
s
cos ( ω t) u(t)⇔
s2 + ω 2
however, from frequency shift theorem
L f (t)e−at u(t) = F (s + a)
 

thus, we replace every s with s + a

s+a
e−a t cos ( ω t) u(t)⇔ 2 2
( s + a) + ω

Al-Motasem Aldaoudeyeh, PhD Slide 20 of 242 BasmahWantsMeHappy

Laplace Transform Theorems [Oga10; p. 863]
Obtain the inverse Laplace transform for the following frequency-domain function

1
F (s) =
( s + a)n

Al-Motasem Aldaoudeyeh, PhD Slide 21 of 242 BasmahWantsMeHappy

Laplace Transform Theorems [Oga10; p. 863]
▶ According to Laplace transform tables
n!
tn u(t)⇔
sn+1
however, from frequency shift theorem
L f (t)e−at u(t) = F (s + a)
 

thus, we multiply f (t) by e−a t and replace every s in F (s) with s + a

n!
e−a t tn u(t)⇔ n+1
(s + a)

Al-Motasem Aldaoudeyeh, PhD Slide 22 of 242 BasmahWantsMeHappy

Laplace Transform Theorems [Oga10; p. 863]
▶ Defining k = n + 1 and replacing every n with k − 1
k − 1)!
(
e−a t tk−1 u(t)⇔
s + a)k
(

dividing both sides by (k − 1)!

1 1
e−a t tk−1 u(t)⇔
( k − 1)! s + a)k
(

thus, we get
1 1
e−a t tn−1 u(t)⇔
( n − 1)! s + a)n
(

Al-Motasem Aldaoudeyeh, PhD Slide 23 of 242 BasmahWantsMeHappy

Laplace Transform Theorems [Oga10; p. 863]
Obtain the Laplace transform for the following time-domain function

ω t) u(t)
sinh (

Al-Motasem Aldaoudeyeh, PhD Slide 24 of 242 BasmahWantsMeHappy

Laplace Transform Theorems [Oga10; p. 863]
▶ From the definition of sinh (ω t)
eω t − e−ω t
sinh ( ω t) u(t) = u(t)s
2
1
eω t − e−ω t u(t)
 
sinh ( ω t) u(t) =
2
from linearity theorem
L [f1 (t) ± f2 (t)] = F1 (s) ± F2 (s) additivity property
L [kf (t)] = kF (s) homogeneity property

Al-Motasem Aldaoudeyeh, PhD Slide 25 of 242 BasmahWantsMeHappy

Laplace Transform Theorems [Oga10; p. 863]
▶ Thus
 
1  ωt −ω t
 1 1 1
e −e u(t)⇔ −
2 2 s−ω s+ω
unifying denominators and simplifying
1  ωt ω
e − e−ω t u(t)⇔

2 s − ω2
2

Al-Motasem Aldaoudeyeh, PhD Slide 26 of 242 BasmahWantsMeHappy

Try It in MATLAB
%% (nise2011, P2, 98)

syms a omega t s

=
f_t = exp( a*t)*cos(omega*t);
F_s = laplace(f_t, t, s);
pretty(F_s)

%% (ogata2010, 863)

syms a omega t s n

F_s = 1/(s + a)^n;

f_t = ilaplace(F_s, s, t);
pretty(f_t)

%% (ogata2010, 863)

Al-Motasem Aldaoudeyeh, PhD Slide 27 of 242 BasmahWantsMeHappy

syms a omega t s n

f_t = sinh(omega*t);
F_s = laplace(f_t, t, s);
pretty(F_s)

Al-Motasem Aldaoudeyeh, PhD Slide 28 of 242 BasmahWantsMeHappy

 Partial-Fraction Expansion
▶ Assuming we have complicated frequency domain function F1 (s) = N (s)/D(s). To
obtain f (t), we convert such function to simpler terms [Nis11; p. 37]
d the method which allows us to simplify F1 (s) is known as partial-fraction expansion

Al-Motasem Aldaoudeyeh, PhD Slide 29 of 242 BasmahWantsMeHappy

▶ For partial fraction expansion to be possible the order of N (s) (i.e., the numerator)
must be at most equal to the order of D(s) (i.e., the denominator). For example, the
order of N (s) in [Nis11; p. 37]
s3 + 2s2 + 6s + 7
F1 (s) =
s2 + s + 5
is 3, while the order of D(s) is 2. Thus, we simplify the expression as
2
= s+1+ 2 +
s s+5
the inverse Laplace is
 
dδ (t) −1 2
f1 (t) = + δ (t) + L
dt s2 + s + 5

Al-Motasem Aldaoudeyeh, PhD Slide 30 of 242 BasmahWantsMeHappy

Case 1. Roots of the Demoniator of F (s) Are Real and Distict [Nis11; p. 38]
▶ For example the denominator of
2
F (s) =
( s + 1)(s + 2)

has
d distinct roots (i.e., all factors are raised only to unity)
d real roots (i.e., solutions of (s + 1) = 0 is −1 and solution of (s + 2) = 0 is −2)

▶ We thus write the partial-fraction expansion as sum of terms with each factor
forming a denominator and constants at the numerator called residues
2 K1 K2
F (s) = = +
( s + 1)(s + 2) s + 1)
( s + 2)
(

Al-Motasem Aldaoudeyeh, PhD Slide 31 of 242 BasmahWantsMeHappy

Case 1. Roots of the Demoniator of F (s) Are Real and Distict [Nis11; p. 38]
▶ To calculate K1 multiply the previous equation by (s + 1)
2 K1 K2
F (s)(s + 1) = = + (s + 1)
s + 2)
( 1 s + 2)
(

by subtituting s = −1 we get
2 K1 K2
F (s) × 0 = = + ×0
−1 + 2)
( 1 −1 + 2)
(

K1 = 2

in a similar fashion, we obtain K2 = −2

Al-Motasem Aldaoudeyeh, PhD Slide 32 of 242 BasmahWantsMeHappy

Case 1. Roots of the Demoniator of F (s) Are Real and Distict [Nis11; p. 38]
▶ The expanded F (s) becomes
2 2
F (s) = −
( s + 1) ( s + 2)
from Laplace transform tables
1
e−at u(t)⇔
s+a
thus, converting to time domain
f (t) = (2e−t − 2e−2t ) u(t)

Al-Motasem Aldaoudeyeh, PhD Slide 33 of 242 BasmahWantsMeHappy

Case 1. Roots of the Demoniator of F (s) Are Real and Distict [Nis11; p. 38]
▶ We may generalize this as follows
N (s) N (s)
F (s) = =
D(s) (s + p1 )(s + p2 ) · · · (s + pm ) · · · (s + pn )
K1 K2 Km Kn
= + + ··· + + ··· +
(s + p1 ) (s + p2 ) (s + pm ) (s + pn )

▶ Km may this be calculated as


N (s) 
Km =  (14)
D̂(s) s=−pm

where D̂(s) is the denominator D(s) with the complex factor (s + pm ) removed

Al-Motasem Aldaoudeyeh, PhD Slide 34 of 242 BasmahWantsMeHappy

Partial Fraction Expansion with Distinct Poles [IN15; p. 565]
Calculate the inverse Laplace for the following function

s + 18
F (s) =
s (s + 2)(s + 3)

Al-Motasem Aldaoudeyeh, PhD Slide 35 of 242 BasmahWantsMeHappy

Inverse Laplace Transform with Distinct Poles [IN15; p. 565]
▶ Expanding F (s)
s + 18 K1 K2 K3
F (s) = = + +
s (s + 2)(s + 3) s s+2 s+3
K1 and K2 , and K3 are calculated using Eq. (14)

N (s) 
Km = 
D̂(s) s=−p
m

s + 18 
K1 =  = 3
(s + 2)(s + 3)

s =0

s + 18 
K2 = = −8
s (s + 3) s=−2

s + 18 
K3 = = 5
s (s + 2) s=−3

Al-Motasem Aldaoudeyeh, PhD Slide 36 of 242 BasmahWantsMeHappy

Inverse Laplace Transform with Distinct Poles [IN15; p. 565]
▶ Substituting numbers
3 −8 5
F (s) = + +
s s+2 s+3
from Laplace transform tables
1
e−at u(t)⇔
s+a
thus f (t) is
− 8e−2t + 5e−3t u(t)


f (t) = 3

Al-Motasem Aldaoudeyeh, PhD Slide 37 of 242 BasmahWantsMeHappy

Try It in MATLAB
syms s t

F_s = (s + 18)/(s*(s + 2)*(s + 3));

f_t = ilaplace(F_s, s, t);
pretty(f_t)

Al-Motasem Aldaoudeyeh, PhD Slide 38 of 242 BasmahWantsMeHappy

Case 2. Roots of the Demoniator of F (s) Are Real and Repeated [Nis11; pp. 40–41]
▶ For example. The denominator of
2
F (s) =
s + 1)(s + 2)2
(

has repeated roots since the factor (s + 2)2 is raised to an integer greater than 1
▶ We thus write the expansion as normal except that each set of repeated roots generate
additional terms of reduced multiplicity
2 K1 K2 K3
F (s) = 2
= +
2
+
s + 1)(s
( + 2) (s + 1) s
( + 2) (s + 2)

Al-Motasem Aldaoudeyeh, PhD Slide 39 of 242 BasmahWantsMeHappy

Case 2. Roots of the Demoniator of F (s) Are Real and Repeated [Nis11; pp. 40–41]
▶ K1 = 2 (calculated as described in Case 1)
▶ K2 may be calculated as by mutliplying both sides by (s + 2)2
2 K1 K2 K2
2
( s + 2)2 =
2
(s + 2) +
2
2
(s + 2) + (s + 2)
2
(s + 1)(s + 2) (s + 1) (s + 2) (s + 2)

2 K1 2
= (s + 2) + K2 + K3 (s + 2)
(s + 1) (s + 1)

substituting s = −2 yields the value of K2


2 
K2 =  = −2
s + 1 s =− 2

Al-Motasem Aldaoudeyeh, PhD Slide 40 of 242 BasmahWantsMeHappy

Case 2. Roots of the Demoniator of F (s) Are Real and Repeated [Nis11; pp. 40–41]

2 K1
F (s)(s + 2)2 = =
2
(s + 2) + K2 + K3 (s + 2)
( s + 1) s + 1)
(

▶ Differentiating F (s)(s + 2)2 yields

s + 2)s
2 (
− 22
K1 + K3
=
(s (s + 1)
+ 1)

substituting s = −2 yields the value of K3


2 (s + 2)s 
K3 = − 2
− 2
K1  = −2
(s + 1) (s + 1)
s=−2

Al-Motasem Aldaoudeyeh, PhD Slide 41 of 242 BasmahWantsMeHappy

Case 2. Roots of the Demoniator of F (s) Are Real and Repeated [Nis11; pp. 40–41]
▶ The expanded F (s) becomes
2
F (s) =
(s + 1)(s + 2)2
2 −2 −2
= + +
(s + 1) (s + 2)
2 s + 2)
(

from Laplace transform tables

1
e−at u(t)⇔
s+a
1
t e−at u(t)⇔
s + a)2
(

Al-Motasem Aldaoudeyeh, PhD Slide 42 of 242 BasmahWantsMeHappy

Case 2. Roots of the Demoniator of F (s) Are Real and Repeated [Nis11; pp. 40–41]
▶ The solution thus becomes
e−t − 2t e−2t − 2e−2t u(t)


f (t) = 2

Al-Motasem Aldaoudeyeh, PhD Slide 43 of 242 BasmahWantsMeHappy

Case 2. Roots of the Demoniator of F (s) Are Real and Repeated [Nis11; pp. 40–41]
▶ We may generalize this as follows
N (s) N (s)
F (s) = =
D(s) s + p1 )r (s + p2 ) · · · (s + pn )
(

K1 K2 Kr
= + + ··· +
(s + p1 )
r (s + p1 )
r− 1 (s + p1 )

Kr+1 Kn
+ + ··· +
(s + p2 ) (s + pn )

Al-Motasem Aldaoudeyeh, PhD Slide 44 of 242 BasmahWantsMeHappy

Case 2. Roots of the Demoniator of F (s) Are Real and Repeated [Nis11; pp. 40–41]
▶ Multiplying F (s) by (s + p1 )r
F̂ (s) = F (s)(s + p1 )r
K1 r K2 r Kr r
= (s + p1 ) + (s + p1 ) + · · · + (s + p1 )
(s + p1 )
r (s + p1 )
r−1 (s + p1 )

Kr+1 r Kn r
+ (s + p1 ) + · · · + (s + p1 )
(s + p2 ) (s + pn )

Al-Motasem Aldaoudeyeh, PhD Slide 45 of 242 BasmahWantsMeHappy

Case 2. Roots of the Demoniator of F (s) Are Real and Repeated [Nis11; pp. 40–41]
▶ Multiplying F (s) by (s + p1 )r
F̂ (s) = F (s)(s + p1 )r
= K1 + (s + p1 )K2 + (s + p1 )2 K3 + · · · + (s + p1 )r−1 Kr
Kr+1 r Kn r
+ (s + p1 ) + · · · + (s + p1 )
(s + p2 ) (s + pn )

▶ We calculate the Ki as follows


1 di−1 F̂ (s) 
Ki = ; i = 1, 2, · · · , r; 0! = 1 (15)
(i − 1)! d si−1 

s=−p1

Al-Motasem Aldaoudeyeh, PhD Slide 46 of 242 BasmahWantsMeHappy

Partial Fraction Expansion with Repeated Poles [IN15; p. 566]
Calculate the inverse Laplace for the following function

s+8
F (s) = 2(
s s + 2)

Al-Motasem Aldaoudeyeh, PhD Slide 47 of 242 BasmahWantsMeHappy

Partial Fraction Expansion with Repeated Poles [IN15; p. 566]
▶ Expanding F (s)
s+8 K1 K2 K3
F (s) = =
2(
+ +
s s + 2) s2 s (s + 2)

K3 is calculated using Eq. (14)


N (s) 
K3 = 
D̂(s) s=−p
m

s + 8 
= = 1.5
s2 s=−2

Al-Motasem Aldaoudeyeh, PhD Slide 48 of 242 BasmahWantsMeHappy

Partial Fraction Expansion with Repeated Poles [IN15; p. 566]
▶ Multiplying F (s) by s2
s+8 2 K1 2 K2 2 K3 2
F̂ (s) = s = s + s + s
s2 (s + 2) s2 s (s + 2)

simplifying
s+8 K3 2
= = K1 + K2 s + s
s + 2)
( s + 2)
(

Al-Motasem Aldaoudeyeh, PhD Slide 49 of 242 BasmahWantsMeHappy

Partial Fraction Expansion with Repeated Poles [IN15; p. 566]

s+8 K3 2
F̂ (s) = = K1 + K2 s + s
(s + 2) (s + 2)

▶ K1 and K2 are calculated from Eq. (15)


1 di−1 F̂ (s) 
Ki = ; i = 1, 2, · · · , r; 0! = 1
(i − 1)! d si−1 

s=−p1

(s + 8) 
K1 =  = 4
(s + 2)

s=0
d (ss+8
 
1 6

+2)  
K2 = = −  = −1.5
(2 − 1)! ds  2 
(s + 2 )

s=0 s=0

Al-Motasem Aldaoudeyeh, PhD Slide 50 of 242 BasmahWantsMeHappy

Partial Fraction Expansion with Repeated Poles [IN15; p. 566]
▶ Substituting numbers
4 −1.5 1.5
F (s) = 2
+ +
s s ( s + 2)
from Laplace transform tables
1
u(t)⇔
s
−at 1
e u(t)⇔
s+a
1
t u(t)⇔
s2
thus f (t) is
t − 1.5 + 1.5e−2t u(t)


f (t) = 4

Al-Motasem Aldaoudeyeh, PhD Slide 51 of 242 BasmahWantsMeHappy

Try It in MATLAB
syms s t

F_s = (s + 8)/(s^2*(s + 2));

f_t = ilaplace(F_s, s, t);
pretty(f_t)

Al-Motasem Aldaoudeyeh, PhD Slide 52 of 242 BasmahWantsMeHappy

Case 3. Roots of the Demoniator of F (s) Are Complex or Imaginary [Nis11; pp. 41–44]
▶ For example
3
F (s) = 2 +2
s s
( s + 5)
can be expanded as follows

K1 s + K2 K3
= +
2
s + 2s + 5 s
K3 = 0.6, while K1 and K2 are calculated by multiplying both sides by the
dominator of F (s) (i.e., s (s2 + 2s + 5))
3 K1 s + K2 K3
s (s2 + 2s + 5) = 2 s (s2 + 2s + 5) + s (s2 + 2s + 5)
s (s2 + 2s + 5) s + 2s + 5 s

Al-Motasem Aldaoudeyeh, PhD Slide 53 of 242 BasmahWantsMeHappy

Case 3. Roots of the Demoniator of F (s) Are Complex or Imaginary [Nis11; pp. 41–44]
▶ Substituting K3 = 0.6 and simplifying
3 = ( K1 + 0.6) s2 + (K2 + 1.2) s + 3
the previous equation is satisfied only when K1 = −0.6 and K2 = −1.2
0.6s + 1.2 0.6
F (s) = 2 +
s + 2s + 5 s
2
the solution of s + 2s + 5 is s1, 2 = −1 ± j 2. We thus rewrite the previous equation as
(s + 2) 0.6
= −0.6 +
2
(s + 1) + 2
2 s
(s + 1) + 1 0.6
= −0.6 +
2
(s + 1) + 2
2 s

Al-Motasem Aldaoudeyeh, PhD Slide 54 of 242 BasmahWantsMeHappy

Rewriting Complex Factor with Complex Roots
▶ For a complex factor s2 + α s + β whose roots are s1, 2 = −a ± jω, it is possible to
rewrite it as
s2 + α s + β s + a)2 + ω 2
= ( (16)

Al-Motasem Aldaoudeyeh, PhD Slide 55 of 242 BasmahWantsMeHappy

Case 3. Roots of the Demoniator of F (s) Are Complex or Imaginary [Nis11; pp. 41–44]
▶ We thus get the following
(s + 1) 1 0.6
F (s) = − 0.6 − 0.6 +
( s + 1)2 + 22 2
(s + 1) + 2
2 s
(s + 1) 0.6 2 0.6
= − 0.6 2 2
− 2 2
+
(s + 1) + 2 2 (s + 1) + 2 s
from Laplace transform tables
s + a)
(
A e−at cos (ω t) u(t)⇔A 2 2
(s + a) + ω
ω
B e−at sin (ω t) u(t)⇔B 2 2
(s + a) + ω

Al-Motasem Aldaoudeyeh, PhD Slide 56 of 242 BasmahWantsMeHappy

Case 3. Roots of the Demoniator of F (s) Are Complex or Imaginary [Nis11; pp. 41–44]
▶ The solution thus becomes
0.6
f (t) = −0.6e−t cos (2t) u(t) − e−t sin (2t) u(t) + 0.6u(t)
2
  
−t 1
= 0.6 − 0.6e cos (2t) + sin (2t) u(t)
2

Al-Motasem Aldaoudeyeh, PhD Slide 57 of 242 BasmahWantsMeHappy

Partial Fraction Expansion with Complex Poles [IN15; p. 565]
Calculate the inverse Laplace for the following function

s + 3)(s2 + 10s + 15)

(
F (s) = 2
(s + 10)(s + 15)(s + 22s + 137)

Al-Motasem Aldaoudeyeh, PhD Slide 58 of 242 BasmahWantsMeHappy

Partial Fraction Expansion with Complex Poles [IN15; p. 565]
▶ Solving (s2 + 22s + 137) yields s1, 2 = −11 ± j 4. Thus, using Eq. (16) we rewrite the
transfer function as
s + 3)(s2 + 10s + 15)
(
F (s) = 2 2
(s + 10)(s + 15) [(s + 11) + 4 ]

expanding F (s)
K1 s + K2 K3 K4
= + +
( s 2
+ 11) + 4
2 s + 10 s + 15

Al-Motasem Aldaoudeyeh, PhD Slide 59 of 242 BasmahWantsMeHappy

Partial Fraction Expansion with Complex Poles [IN15; p. 565]

s + 3)(s2 + 10s + 15)

( K1 s + K2 K3 K4
F (s) = 2 2
=
2 2
+ +
(s + 10)(s + 15) [(s + 11) + 4 ] s
( + 11) + 4 s + 10 s + 15

▶ K3 and K4 are calculated using Eq. (14)


N (s) 
Km = 
D̂(s) s=−p
m
2

(s + 3)(s + 10s + 15) 
K3 =
2
= −1.235
(s + 15)(s + 22s + 137)

s=−10
2

(s + 3)(s + 10s + 15) 
K4 =
2
 = 6.75
(s + 10)(s + 22s + 137)

s=−15

Al-Motasem Aldaoudeyeh, PhD Slide 60 of 242 BasmahWantsMeHappy

Partial Fraction Expansion with Complex Poles [IN15; p. 565]

s + 3)(s2 + 10s + 15)

( K1 s + K2 K3 K4
F (s) = 2 2
=
2 2
+ +
(s + 10)(s + 15) [(s + 11) + 4 ] s
( + 11) + 4 s + 10 s + 15

▶ To calculate K1 and K2 multiply both sides by (s + 10)(s + 15) (s + 11)2 + 42

 

(s + 3)(s2 + 10s + 15) = (K1 s + K2 )(s + 10)(s + 15)

2 2 2 2
   
+K3 (s + 15) (s + 11) + 4 + K4 (s + 10) (s + 11) + 4

to calculate K2 , sbustitute s = 0
2 2
 
3 × 15 = K2 × 10 × 15 + (−1.235) × 15 11 + 4
 2 2

+6.75 × 10 11 + 4

K2 = −44.43

Al-Motasem Aldaoudeyeh, PhD Slide 61 of 242 BasmahWantsMeHappy

Partial Fraction Expansion with Complex Poles [IN15; p. 565]

( s + 3)(s2 + 10s + 15) = (K1 s + K2 )(s + 10)(s + 15)

2 2 2 2
   
+K3 (s + 15) (s + 11) + 4 + K4 (s + 10) (s + 11) + 4

▶ To calculate K1 expand both sides and substitute K2 , K3 and K4

s3 + 13s2 + 45s + 45 = (K1 + 5.515)s3 + (25K1 + 125.9)s2
+(150 K1 + 722.3)s + 45.08
differentiating both sides
3 s2 + 26s + 45 = 3s2 (K1 + 5.515) + 2s(25K1 + 125.9) + (150K1 + 722.3)
substituting s = 0 and solving for K1
45 = 150 K1 + 722.3
K1 = −4.515

Al-Motasem Aldaoudeyeh, PhD Slide 62 of 242 BasmahWantsMeHappy

Partial Fraction Expansion with Complex Poles [IN15; p. 565]
▶ Substituting numbers
(−4.515)s + (−44.43) −1.235)
( 6.75
F (s) = 2 2
+ +
(s + 11) + 4 s + 10 s + 15
expanding
4.515s 44.43 −1.235)
( 6.75
= − − + +
s + 11)2 + 42 (s + 11)2 + 42
( s + 10 s + 15
rewriting
s + 11 − 11 44.43 ( −1.235) 6.75
= −4.515 2 2
− 2 2
+ +
(s + 11) + 4 (s + 11) + 4 s + 10 s + 15

Al-Motasem Aldaoudeyeh, PhD Slide 63 of 242 BasmahWantsMeHappy

Partial Fraction Expansion with Complex Poles [IN15; p. 565]
▶ Rewriting in standard forms
s + 11) − 11
( 44.43 (−1.235) 6.75
F (s) = −4.515 − + +
s + 11)2 + 42 (s + 11)2 + 42
( s + 10 s + 15
(s + 11) −11 44.43
= −4.515 − 4.515 −
2
(s + 11) + 4
2 2
(s + 11) + 4
2 2
(s + 11) + 4
2

(−1.235) 6.75
+ +
s + 10 s + 15
(s + 11) 5.235 4 1.235 6.75
= −4.515 + − +
2
(s + 11) + 4
2 4
2
(s + 11) + 4
2 s + 10 s + 15

Al-Motasem Aldaoudeyeh, PhD Slide 64 of 242 BasmahWantsMeHappy

Partial Fraction Expansion with Complex Poles [IN15; p. 565]
▶ Thus
(s + 11) 4
F (s) = − 4.515 2 2
+ 1.309
( s + 11) + 4 s + 11)2 + 42
(

1.235 6.75
− +
s + 10 s + 15
from Laplace transform tables
(s + a)
A e−at cos (ω t) u(t)⇔A
s + a)2 + ω 2
(
ω
B e−at sin (ω t) u(t)⇔B 2 2
(s + a) + ω

thus f (t)
f (t) = − 4.515e−11t cos (4t) u(t) + 1.309e−11t sin (4t) u(t)
− 1.235e−10t u(t) + 6.75e−15t u(t)

Al-Motasem Aldaoudeyeh, PhD Slide 65 of 242 BasmahWantsMeHappy

Try It in MATLAB
syms s t

F_s = ((s^2 + 10*s + 15)*(s + 3))/((s + 10)*(s + 15)*(s^2 + 22*s + 137));

pretty(F_s);
f_t = ilaplace(F_s, s, t);
f_t = vpa(f_t, 5);
pretty(f_t);

Al-Motasem Aldaoudeyeh, PhD Slide 66 of 242 BasmahWantsMeHappy

 The Transfer Function
▶ A transfer function is a viable definition which algebrically relates the output of the
system to its input [Nis11; p. 44]
d it also allows us to algebrically combine representations of subsystems to yield a total system
representation

Al-Motasem Aldaoudeyeh, PhD Slide 67 of 242 BasmahWantsMeHappy

Definition of a Transfer Function [Nis11; pp. 44–45]
▶ Assume we want to write an nth-order, linear, time-invariant differential equation as
follows
dn c(t) dn−1 c(t) dm r(t) dm−1 r(t)
an + an−1 + · · · + a0 c(t) = bm + bm−1 + · · · + b0 r(t)
d tn d tn−1 d tm d tm−1
(17)

where
c(t) the output
r(t) the input
ai , bi constants which depend on the system

Al-Motasem Aldaoudeyeh, PhD Slide 68 of 242 BasmahWantsMeHappy

Definition of a Transfer Function [Nis11; pp. 44–45]
▶ Taking the Laplace transform of both sides in Eq. (17)

an sn C (s) + an−1 sn−1 C (s) + · · · + a0 C (s)+ Initial Conditions

Involving c(t)
m
= bm s R(s) + bm−1 s
m−1
R(s) + · · · + b0 R(s)+ Initial Conditions (18)
Involving r(t)

if all initial conditions are zero, then Eq. (18) becomes purely algebraic
( an sn + an−1 sn−1 + · · · + a0 )C (s) = (bm sm + bm−1 sm−1 + · · · + b0 )R(s) (19)

Al-Motasem Aldaoudeyeh, PhD Slide 69 of 242 BasmahWantsMeHappy

Definition of a Transfer Function [Nis11; pp. 44–45]
▶ From Eq. (19) we calculate the ratio of the output C (s) to the input R(s) as follows

C (s) bm sm + bm−1 sm−1 + · · · + b0 )

(
= G(s) = n n−1 + · · · + a )
(20)
R(s) (an s + an−1 s 0

▶ The ratio G(s) in Eq. (20) is known as the transfer function and we evaluate it at zero
initial conditions
▶ Given certain input R(s), we calculate the output C (s)

C (s) = R(s)G(s) (21)

Al-Motasem Aldaoudeyeh, PhD Slide 70 of 242 BasmahWantsMeHappy

Solving a Differential Equation Using Transfer Function [Nis11; p. 46]
Assuming a system has the following differential equation

d3 c d2 c dc d2 r dr
+ 3 + 7 + 5 c= + 4 + 3 r
d t3 d t2 dt d t2 dt

▶ Calculate the transfer function assuming the input is r and the output is c
▶ Assuming the input to be r(t) = t2 u(t), what is the output c(t) in time domain?

Al-Motasem Aldaoudeyeh, PhD Slide 71 of 242 BasmahWantsMeHappy

Solving a Differential Equation Using Transfer Function [Nis11; p. 46]
▶ The differential equation
d3 c d2 c dc d2 r dr
+ 3 + 7 + 5 c= + 4 + 3 r
d t3 d t2 dt d t2 dt
dn
to convert such equation to s-domain replace every with sn , every c with C (s)
d tn
and every r with R(s)
s3 C (s) + 3s2 C (s) + 7sC (s) + 5C (s) = s2 R(s) + 4sR(s) + 3R(s)
collecting C (s) and R(s)
s3 + 3s2 + 7s + 5 C (s) = s2 + 4s + 3 R(s)

thus, we get G(s) as

C (s) s2 + 4 s + 3
G(s) = =
R(s) s3 + 3s2 + 7s + 5

Al-Motasem Aldaoudeyeh, PhD Slide 72 of 242 BasmahWantsMeHappy

Solving a Differential Equation Using Transfer Function [Nis11; p. 46]
▶ The input in time domain is
r(t) = t2 u(t)
from Laplace transform tables
n!
tn u(t)⇔
sn+1
thus the input in s domain
2! 2
R(s) = 2+1
=
s s3

Al-Motasem Aldaoudeyeh, PhD Slide 73 of 242 BasmahWantsMeHappy

Solving a Differential Equation Using Transfer Function [Nis11; p. 46]
▶ The output in s domain is calculated using Eq. (21)
C (s) = R(s)G(s)
2 s2 + 4 s + 3
=
s3 s3 + 3s2 + 7s + 5

Al-Motasem Aldaoudeyeh, PhD Slide 74 of 242 BasmahWantsMeHappy

Solving a Differential Equation Using Transfer Function [Nis11; p. 46]
▶ The output in s domain is calculated using Eq. (21)
C (s) = R(s)G(s)
s2 + 4s + 3
2
=
s3 s3 + 3 s2 + 7 s + 5
3 2
solving the factor s + 3s + 7s + 5 yields s1, 2, 3 = {−1, −1 ± j 2}. Thus, using
Eq. (16) we rewrite C (s) as
2 s2 + 4 s + 3
=
s3 (s + 1) [(s + 1)2 + 22 ]

Al-Motasem Aldaoudeyeh, PhD Slide 75 of 242 BasmahWantsMeHappy

Solving a Differential Equation Using Transfer Function [Nis11; p. 46]

2 s2 + 4 s + 3
C (s) = 3 (
s s + 1) [(s + 1)2 + 22 ]

▶ Expanding C (s)
K1 K2 K3 K4 K5 s + K6
C (s) = + + + +
s3 s2 s s + 1 (s + 1)2 + 22
multiply both sides by s3
s2 + 4s + 3 K4 K5 s + K6
s3 C (s) = 2 2 2
= K1 + K2 s + K3 s +
2
s+ 2 2
s
(s + 1) [(s + 1) + 2 ] s+1 (s + 1) + 2

we calculate K1 , K2 and K3 using Eq. (15)


1 di−1 Ĉ (s) 
Ki = ; i = 1, 2, 3; 0! = 1
(i − 1)! d si−1 

s=−p1

Al-Motasem Aldaoudeyeh, PhD Slide 76 of 242 BasmahWantsMeHappy

Solving a Differential Equation Using Transfer Function [Nis11; p. 46]
▶ To calculate K1 substitute s = 0 in s3 C (s)
3
2 × = K1
(1) × [(1)2 + 22 ]
K1 = 1.2

Al-Motasem Aldaoudeyeh, PhD Slide 77 of 242 BasmahWantsMeHappy

Solving a Differential Equation Using Transfer Function [Nis11; p. 46]
▶ To calculate K2 differentiate s3 C (s), substitute s = 0 and divide over (2 − 1)!
s2 +4s+3

d2 (s+1)[( 2 2

1 s+1) +2 ] 
K2 = = −0.08
(2 − 1)! ds 

s=0
to calculate K3 take the second derivative of s3 C (s), substitute s = 0 and divide
over (3 − 1)!
s2 +4s+3

2
d 2 2

1 2
(s+1)[(s+1) +2 ] 
K3 =  = −0.208
(3 − 1)! d s2 

s=0

Al-Motasem Aldaoudeyeh, PhD Slide 78 of 242 BasmahWantsMeHappy

Solving a Differential Equation Using Transfer Function [Nis11; p. 46]
▶ To calculate K4 use Eq. (14)

N (s) 
K4 = 
D̂(s) s=−1
s2 + 4s + 3 

2
=
s3 [(s + 1)2 + 22 ] s=−1
= 0

Al-Motasem Aldaoudeyeh, PhD Slide 79 of 242 BasmahWantsMeHappy

Solving a Differential Equation Using Transfer Function [Nis11; p. 46]
▶ Now we substitute numbers in the expanded C (s)
s2 + 4 s + 3
2
C (s) =
s3 (s + 1) [(s + 1)2 + 22 ]
1.2 −0.08 −0.208 0 K5 s + K6
= + + + +
s 3 s2 s s + 1 (s + 1)2 + 22
multiply both sides by s3 (s + 1) (s + 1)2 + 22
 

s3 (s + 1) (s + 1)2 + 22 C (s) = 2(s2 + 4s + 3)

 

2 2
 
= 1.2(s + 1) (s + 1) + 2

2 2
 
+(−0.08)s (s + 1) (s + 1) + 2

2 2 2
 
+(−0.208)s (s + 1) (s + 1) + 2

+[ K5 s + K6 ] s3 (s + 1)

Al-Motasem Aldaoudeyeh, PhD Slide 80 of 242 BasmahWantsMeHappy

Solving a Differential Equation Using Transfer Function [Nis11; p. 46]
▶ To calculate K6 we take the third derivative of s3 (s + 1) (s + 1)2 + 22 C (s) and
 

substitute s = 0
0 = (60 K5 − 12.48)s2 + (24K5 + 24K6 − 16.9)s + 6K6 − 2.976
0 = 6 K6 − 2.976
K6 = 0.496

to calculate K5 we take the fourth derivative of s3 (s + 1) (s + 1)2 + 22 C (s) and

 

substitute s = 0
0 = 24 K5 + 24K6 + s (120K5 − 24.96) − 16.9
substituting K6 = 0.496

0 = 24 K5 − 4.992
K5 = 0.208

Al-Motasem Aldaoudeyeh, PhD Slide 81 of 242 BasmahWantsMeHappy

Solving a Differential Equation Using Transfer Function [Nis11; p. 46]
▶ We finally get the following
2 s2 + 4 s + 3
C (s) =
s3 (s + 1) [(s + 1)2 + 22 ]
1.2 −0.08 −0.208 0 s + 0.496
0.208
= + + + +
s 3 s 2 s s+1 (s + 1)2 + 22
rewriting
2−0.08 −0.208 (s + 1 − 1) 0.496
= 0.6 + + + 0.208 +
s 3 s2 s 2
(s + 1) + 2
2 ( s + 1)2 + 22
2 −0.08 −0.208 (s + 1)
= 0.6 + + + 0.208
s 3 s2 s 2
(s + 1) + 2
2

(−1) 0.496
+0.208 +
2
(s + 1) + 2
2 2
(s + 1) + 2
2

Al-Motasem Aldaoudeyeh, PhD Slide 82 of 242 BasmahWantsMeHappy

Solving a Differential Equation Using Transfer Function [Nis11; p. 46]
▶ Now we substitute numbers in the expanded C (s)
2 −0.08 −0.208 (s + 1) 0.288
C (s) = 0.6 + + + 0.208 +
s3 s2 s s + 1)2 + 22
( s + 1)2 + 22
(

2 −0.08 −0.208 (s + 1) 0.288 2

= 0.6 + + + 0.208 +
s 3 s2 s 2
(s + 1) + 2
2 2 (s + 1)2 + 22

▶ From Laplace transform tables

n!
tn u(t)⇔
sn+1
(s + a)
A e−at cos (ω t) u(t)⇔A
s + a)2 + ω 2
(
ω
B e−at sin (ω t) u(t)⇔B 2 2
(s + a) + ω

Al-Motasem Aldaoudeyeh, PhD Slide 83 of 242 BasmahWantsMeHappy

Solving a Differential Equation Using Transfer Function [Nis11; p. 46]
▶ Thus, c(t) is
c(t) = 0.6t2 − 0.08t − 0.208 + 0.208 e−t cos (2t) + 0.144 e−t sin (2t) u(t)
 

Al-Motasem Aldaoudeyeh, PhD Slide 84 of 242 BasmahWantsMeHappy

Try It in MATLAB
syms s t

G_s = (s^2 + 4*s + 3)/(s^3 + 3*s^2 + 7*s + 5);

R_s = 2/s^3;
C_s = R_s*G_s;

c_t = ilaplace(C_s, s, t);

c_t = vpa(c_t, 5);
pretty(c_t)

Al-Motasem Aldaoudeyeh, PhD Slide 85 of 242 BasmahWantsMeHappy

 Electrical Networks Transfer Functions
▶ Electrical networks contain three main passive components [Nis11; pp. 47–48]
d resistors
d inductors
d capacitors

▶ Passive means no internal source of energy

Al-Motasem Aldaoudeyeh, PhD Slide 86 of 242 BasmahWantsMeHappy

▶ When obtaining the transfer functions we could follow one of the following methods
[Nis11; pp. 47–48]
Write an Integrodifferential Equation
Use The Concept of Impedances
▶ With either method, we need to clearly specify the input and the output

Al-Motasem Aldaoudeyeh, PhD Slide 87 of 242 BasmahWantsMeHappy

▶ It is possible to obtain the transfer function by [Nis11; pp. 47–48]
Kirchoff’s Voltage Law which involves writing voltages in terms of currents
Kirchoff’s Current Law which involves writing currents in terms of voltages
▶ The suitability of either method depends on the amount of effort and algebraic
manipulation required

Al-Motasem Aldaoudeyeh, PhD Slide 88 of 242 BasmahWantsMeHappy

Voltage-Current Relationships for Electrical Components [Nis11; pp. 47–48]

Time Domain Frequency Domain

Voltage-Current Voltage-Current Impedance

Component Component V (s)
Relationship Relationship Z (s) =
I (s)

v (t) = Ri(t) V (s) = RI (s) R

di(t) V (s) = sLI (s) sL

v (t) = L
dt
v (t) =
1 1
Zt V (s) = I (s)
1
i(τ ) dτ sC sC
C
0
Current-Voltage Relationships for Electrical Components [Nis11; pp. 47–48]

Time Domain Frequency Domain

Current-Voltage Current-Voltage Admittance
Component Component I (s)
Relationship Relationship Y (s) =
V (s)
1 1 1
i(t) = v (t) I (s) = V (s) G=
R R R
i(t) =
1 1
Zt I (s) = V (s)
1
v (τ ) dτ sL sL
L
0

dv (t) I (s) = sCV (s) sC

i(t) = C
dt
Transfer Functions of Single Loop Electrical Network [Nis11; pp. 48–49]
For the circuit shown in the adjacent figure
calculate
▶ The transfer function by writing
integrodifferential equation of the
system +
-
▶ The transfer function by converting the
network to the s domain
Assume the input is v (t) and the output is
vC (t)

Al-Motasem Aldaoudeyeh, PhD Slide 91 of 242 BasmahWantsMeHappy

Transfer Functions of Single Loop Electrical Network [Nis11; pp. 48–49]
▶ Summing voltages around the loop
Zt
di(t) 1
−v (t) + L + Ri(t) + i(τ ) dτ = 0
dt C
0

dq (t)
current is is defined as i(t) = . Thus
dt
d2 qC (t) dqC (t) 1
L + R + qC (t) = v (t)
d t2 dt C

Al-Motasem Aldaoudeyeh, PhD Slide 92 of 242 BasmahWantsMeHappy

Transfer Functions of Single Loop Electrical Network [Nis11; pp. 48–49]
▶ The charge of the capacitor is given as
qC = CvC (t)
substituting qC in the previous equation
d2 vC (t) dvC (t)
LC 2
+ RC + vC (t) = v (t)
dt dt
converting the previous equation into s domain
LCs2 VC (s) + RCsVC (s) + VC (s) = V (s)
collecting VC (s)
LCs2 + RCs + 1 VC (s) = V (s)

Al-Motasem Aldaoudeyeh, PhD Slide 93 of 242 BasmahWantsMeHappy

Transfer Functions of Single Loop Electrical Network [Nis11; pp. 48–49]
▶ The transfer function is thus
VC (s) 1
=
V (s) LCs 2 + RCs + 1
1
LC
=
R 1
s2 + L s + LC

Block Diagram of the Transfer Function

from V (s) to VC (s)

Al-Motasem Aldaoudeyeh, PhD Slide 94 of 242 BasmahWantsMeHappy

Transfer Functions of Single Loop Electrical Network [Nis11; pp. 48–49]
▶ To convert the network to the s domain,
we replace components with their
impedances
d R remains R +
-
d L is replaced with sL
1
d C is replaced with
sC
d all voltages and currents notations are
replaced with capital letters

+
-

Al-Motasem Aldaoudeyeh, PhD Slide 95 of 242 BasmahWantsMeHappy

Transfer Functions of Single Loop Electrical Network [Nis11; pp. 48–49]
▶ Summing voltages around the loop
1
−V (s) + RI (s) + sLI (s) + I (s) = 0
sC
rearranging and collecting I (s)
 
1
R + sL + I (s) = V (s)
sC
V (s)
I (s) = 1
R + sL + sC

Al-Motasem Aldaoudeyeh, PhD Slide 96 of 242 BasmahWantsMeHappy

Transfer Functions of Single Loop Electrical Network [Nis11; pp. 48–49]
▶ VC (s) is calculated as
1
VC (s) = I (s)
sC
substituting I (s)
V (s)1
VC (s) = 1
sC R + sL + sC
dividing both sides over V (s) and rearranging
VC (s) 1 1
=
1
V (s) sC R + sL + sC

Al-Motasem Aldaoudeyeh, PhD Slide 97 of 242 BasmahWantsMeHappy

Transfer Functions of Single Loop Electrical Network [Nis11; pp. 48–49]
▶ The transfer function becomes
VC (s) 1
=
V (s) sRC + s2 LC + 1
1
LC
=
R 1
s2 + L s + LC

Al-Motasem Aldaoudeyeh, PhD Slide 98 of 242 BasmahWantsMeHappy

Transfer Functions of Single Loop Electrical Network [Nis11; pp. 48–49]
For the circuit shown in the following fig-
ure, calculate the transfer function G(s) =
Vo (s)
Vi (s)
+
-

Al-Motasem Aldaoudeyeh, PhD Slide 99 of 242 BasmahWantsMeHappy

Transfer Functions of Single Loop Electrical Network [Nis11; pp. 48–49]
▶ Define mesh currents in the circuit

+
-

Al-Motasem Aldaoudeyeh, PhD Slide 100 of 242 BasmahWantsMeHappy

Transfer Functions of Single Loop Electrical Network [Nis11; pp. 48–49]
▶ Convert the circuit to s domain
d R remains R
d L is replaced with sL
d C is replaced with 1 +
sC -
d all voltages and currents notations are
replaced with capital letters

Al-Motasem Aldaoudeyeh, PhD Slide 101 of 242 BasmahWantsMeHappy

Transfer Functions of Two Loop Electrical Network [Nis11; p. 99]
▶ Summing voltages for loop 1
−Vi (s) + 2sI1 (s) + 1 × (I1 (s) − I2 (s)) = 0
rearranging
s + 1)I1 (s) − I2 (s) = Vi (s)
(2

summing voltages for loop 2

1 × (I2 (s) − I1 (s)) + 3sI2 (s) + 2sI2 (s) = 0
rearranging
2
−I1 (s) + (3s + I s
+ 1) 2 ( ) = 0
s

Al-Motasem Aldaoudeyeh, PhD Slide 102 of 242 BasmahWantsMeHappy

Transfer Functions of Two Loop Electrical Network [Nis11; p. 99]
▶ Rewriting both equations in matrix format
 
(2s + 1) −1 
 I1 (s)
  
 V i (s)
 2  =
−1 3s + + 1 I2 (s) 0
s

Al-Motasem Aldaoudeyeh, PhD Slide 103 of 242 BasmahWantsMeHappy

Transfer Functions of Two Loop Electrical Network [Nis11; p. 99]
▶ Using Cramer’s rule to calculate I2 (s)
s + 1) Vi (s)
(2

−1 0
I2 (s) =
s + 1) 
(2 −1 
2
−1 3s + + 1
s
s
= Vi (s)
3 2
6s + 5s + 4s + 2

Al-Motasem Aldaoudeyeh, PhD Slide 104 of 242 BasmahWantsMeHappy

Transfer Functions of Two Loop Electrical Network [Nis11; p. 99]
▶ The value of Vo (s) is
Vo (s) = 3s × I2 (s)
s
= 3 s × Vi (s)
s3 + 5s2 + 4s + 2
6 +
2 -
3s
= Vi (s)
3 2
6s + 5s + 4s + 2

Vo (s)
▶ The transfer function G(s) = is
Vi (s)
thus
Vo (s) 3 s2
G(s) = =
Vi (s) s3 + 5s2 + 4s + 2
6

Al-Motasem Aldaoudeyeh, PhD Slide 105 of 242 BasmahWantsMeHappy

Try It in MATLAB
syms s I_1 I_2 V_i V_o

eq1 = 2*s*I_1 + 1*(I_1 =

I_2) == V_i;
eq1 = collect(eq1, {'I_1', 'I_2'});
pretty(eq1)

eq2 = 1*(I_2 =
I_1) + 3*s*I_2 + (2/s)*I_2 == 0;
eq2 = collect(eq2, {'I_1', 'I_2'});
pretty(eq2)

eq_sys = [eq1, eq2];

solve(eq_sys, V_i)

num = det([2*s+1, V_i; ...

=1, 0]);
den = det([2*s+1, 1; ...=
=_1,(3*s + 2/s + 1)]);
I 2 = num/den;

V_o = I_2*3*s;
V_o_V_i = V_o/V_i;
pretty(V_o_V_i)

Al-Motasem Aldaoudeyeh, PhD Slide 106 of 242 BasmahWantsMeHappy

Al-Motasem Aldaoudeyeh, PhD Slide 107 of 242 BasmahWantsMeHappy
A Short-Cut to Obtain Loop Equations [Nis11; p. 53]
▶ In the circuit of the previous example it is possible to obtain exactly the same system
of equations as follows
     
Sum of
Sum of
 I1 (s) −  Impedances  I2 (s) = Sum of Voltages
     
+ 
 Impedances   Common to  Around Mesh 1
Around Mesh 1
Mesh 1 and 2
     
Sum of
 Impedances  Sum of
 I1 (s) +  Impedances  I2 (s) = Sum of Voltages
   
−  Common to    Around Mesh 2
Around Mesh 2
Mesh 1 and 2

Al-Motasem Aldaoudeyeh, PhD Slide 108 of 242 BasmahWantsMeHappy

Transfer Functions of Three Loop Electrical Network [Nis11; p. 99]
For the circuit shown in the following fig-
ure, calculate the transfer function G(s) =
Vo (s)
Vi (s)

+
-

Al-Motasem Aldaoudeyeh, PhD Slide 109 of 242 BasmahWantsMeHappy

Transfer Functions of Three Loop Electrical Network [Nis11; p. 99]
▶ Define mesh currents in the circuit

+
-

Al-Motasem Aldaoudeyeh, PhD Slide 110 of 242 BasmahWantsMeHappy

Transfer Functions of Three Loop Electrical Network [Nis11; p. 99]
▶ Convert the circuit to s domain
d R remains R
d L is replaced with sL
d C is replaced with 1
sC
d all voltages and currents notations are
replaced with capital letters
+
-

Al-Motasem Aldaoudeyeh, PhD Slide 111 of 242 BasmahWantsMeHappy

Transfer Functions of Three Loop Electrical Network [Nis11; p. 99]
▶ Since we have four loops, we need 4
equations to solve for Vo (s)
1
▶ However, the impedance s and are in
s
parallel, their equivalent is
!−1
1 1
Z1 (s) = +
1 +
s s
-

s
=
s2 + 1

Al-Motasem Aldaoudeyeh, PhD Slide 112 of 242 BasmahWantsMeHappy

Transfer Functions of Three Loop Electrical Network [Nis11; p. 99]
▶ Redrawing the circuit again
▶ We thus need only 3 equations

+
-

Al-Motasem Aldaoudeyeh, PhD Slide 113 of 242 BasmahWantsMeHappy

Transfer Functions of Three Loop Electrical Network [Nis11; p. 99]
▶ For loop 1
   
s s
s+ 2 I1 (s) − 2 I2 (s) − [s]I3 (s) = Vi (s)
s +1 s +1
for loop 2
   
s s 1
− 2 I1 (s)+ 2 + 1 + I2 (s) − [1]I3 (s) = 0
s +1 s +1 s
for loop 3
− [s]I1 (s) − [1]I2 (s)+ [s + s + 1]I3 (s) = 0

Al-Motasem Aldaoudeyeh, PhD Slide 114 of 242 BasmahWantsMeHappy

Transfer Functions of Three Loop Electrical Network [Nis11; p. 99]
▶ Rewriting the three equations in matrix format
 s s    
s+ 2 − 2 −s I1 (s) Vi (s)
s +1 s +1    
 s s 1   
= 

 − 2 −1

2
+ 1 +  I2 (s) 0 

s +1 s +1 s   
−s −1 2s + 1 I3 (s) 0

Al-Motasem Aldaoudeyeh, PhD Slide 115 of 242 BasmahWantsMeHappy

Transfer Functions of Three Loop Electrical Network [Nis11; p. 99]
▶ We solve for I2 (s) using Cramer’s rule
s
s+ Vi (s) −s
s2 + 1
s
− 2 0 −1
s +1
−s 0 2s+1
I2 (s) = s s
s+ − 2 −s
s2 + 1 s +1
s s 1
− 2 2 +1
+ 1 + −1
s +1 s s
−s −1 s+1
2

Al-Motasem Aldaoudeyeh, PhD Slide 116 of 242 BasmahWantsMeHappy

Transfer Functions of Three Loop Electrical Network [Nis11; p. 99]
▶ The solution of I2 (s) is
s s2 + 2 s + 2


I2 (s) = Vi (s) 4
s + 2s3 + 3s2 + 3s + 2

Al-Motasem Aldaoudeyeh, PhD Slide 117 of 242 BasmahWantsMeHappy

Transfer Functions of Three Loop Electrical Network [Nis11; p. 99]
▶ From the circuit we see that Vo (s) is
given as
1
Vo (s) = × I2 (s)
s
substituting I2 (s)
s s2 + 2 s + 2


1
= × Vi (s) 4
s s + 2 s3 + 3 s2 + 3 s + 2
s2 + 2 s + 2
= Vi (s) +
s4 + 2s3 + 3s2 + 3s + 2 -

Al-Motasem Aldaoudeyeh, PhD Slide 118 of 242 BasmahWantsMeHappy

Transfer Functions of Three Loop Electrical Network [Nis11; p. 99]
Vo (s)
▶ The transfer function G(s) = is
Vi (s)
thus
Vo (s) s2 + 2s + 2
G(s) = =
Vi (s) s4 + 2s3 + 3s2 + 3s + 2

+
-

Al-Motasem Aldaoudeyeh, PhD Slide 119 of 242 BasmahWantsMeHappy

Try It in MATLAB
syms s I_1 I_2 I_3 V_i V_o

eq1 = (s + s/(s^2 + 1))*I_1 =

(s/(s^2 + 1))*I_2 = (s)*I_3 == V_i;
eq1 = collect(eq1, {'I_1', 'I_2', 'I_3'});
pretty(eq1)

=
eq2 = (s/(s^2 + 1))*I_1 + (s/(s^2 + 1) + 1 + 1/s)*I_2 = (1)*I_3 == 0;
eq2 = collect(eq2, {'I_1', 'I_2', 'I_3'});
pretty(eq2)

=
eq3 = (s)*I_1 =
(1)*I_2 + (2*s + 1)*I_3 == 0;
eq3 = collect(eq3, {'I_1', 'I_2', 'I_3'});
pretty(eq3)

num = det([(s + s/(s^2 + 1)), V_i, =(s); ...

=(s/(s^2 + 1)), 0, =(1);...
=(s), 0, (2*s + 1)]);
den = det([(s + s/(s^2 + 1)), = (s/(s^2 + 1)), =(s); ...
=(s/(s^2 + 1)), (s/(s^2 + 1) + 1 + 1/s), =(1);...
=_(s), =(1), (2*s + 1)]);
I 2 = num/den;

Al-Motasem Aldaoudeyeh, PhD Slide 120 of 242 BasmahWantsMeHappy

pretty(I_2)

V_o = I_2*(1/s);
V_o_V_i = V_o/V_i;
pretty(V_o_V_i)

Al-Motasem Aldaoudeyeh, PhD Slide 121 of 242 BasmahWantsMeHappy

 Translational Mechanical Systems Transfer Functions
▶ Mechanical networks contain three main passive components [Nis11; pp. 61–69]
d dampers (which account for friction)
d springs
d masses

▶ Passive means no internal source of energy

Al-Motasem Aldaoudeyeh, PhD Slide 122 of 242 BasmahWantsMeHappy

Force-Displacement Relationships for Mechanical Components [Nis11; pp. 61–69]

Time Domain Frequency Domain

Force- Force- Impedance
Component Displacement Component Displacement F (s)
Relationship Relationship ZM (s) =
X (s)

dx(t) F (s) = s fv X (s) s fv

f (t) = fv
dt

f (t) = Kx(t) F (s) = KX (s) K

Force-Displacement Relationships for Mechanical Components [Nis11; pp. 61–69]

Time Domain Frequency Domain

Force- Force- Impedance
Component Displacement Component Displacement F (s)
Relationship Relationship ZM (s) =
X (s)

d2 x(t) F (s) = s2 M X (s) s2 M

f (t) = M
d t2
Transfer Functions of One Degree-of-Freedom Mechanical Systems
[Nis11; pp. 63–64]
For the mass-spring-damper system shown
in the adjacent figure calculate
▶ The transfer function by writing
integrodifferential equation of the
system
▶ The transfer function by converting the
network to the s domain
Assume the input is f (t) and the output is
x(t)

Al-Motasem Aldaoudeyeh, PhD Slide 125 of 242 BasmahWantsMeHappy

Transfer Functions of One Degree-of-Freedom Mechanical Systems
[Nis11; pp. 63–64]
▶ Draw the free body diagram of the system

Al-Motasem Aldaoudeyeh, PhD Slide 126 of 242 BasmahWantsMeHappy

Transfer Functions of One Degree-of-Freedom Mechanical Systems
[Nis11; pp. 63–64]
▶ The sum of forces must be zero. Thus,
d2 x(t) dx(t)
f (t) − M 2
− fv − Kx(t) = 0
dt dt
rearranging
d2 x(t) dx(t)
2
M
+ fv + Kx(t) = f (t)
dt dt
taking the Laplace transform assuming zero initial conditions
s2 M X (s) + s fv X (s) + K X (s) = F (s)
collecting X (s)
s2 M


+ s fv + K X (s) = F (s)

Al-Motasem Aldaoudeyeh, PhD Slide 127 of 242 BasmahWantsMeHappy

Transfer Functions of One Degree-of-Freedom Mechanical Systems
[Nis11; pp. 63–64]
X (s)
▶ The transfer function G(s) =
F (s)
X (s) 1
G(s) = =
2 +
F (s) Ms fv s + K
Block Diagram of the Transfer Function
from V (s) to VC (s)

Al-Motasem Aldaoudeyeh, PhD Slide 128 of 242 BasmahWantsMeHappy

Transfer Functions of One Degree-of-Freedom Mechanical Systems
[Nis11; pp. 63–64]
▶ Convert the free body diagram into the frequency domain

Al-Motasem Aldaoudeyeh, PhD Slide 129 of 242 BasmahWantsMeHappy

Transfer Functions of One Degree-of-Freedom Mechanical Systems
[Nis11; pp. 63–64]
▶ The sum of forces must be zero. Thus,
F (s) − s2 M X (s) − s fv X (s) − K X (s) =
rearranging
s2 M X (s) + s fv X (s) + K X (s) = F (s)
collecting X (s)
s2 M


+ s fv + K X (s) = F (s)

Al-Motasem Aldaoudeyeh, PhD Slide 130 of 242 BasmahWantsMeHappy

Transfer Functions of One Degree-of-Freedom Mechanical Systems
[Nis11; pp. 63–64]
X (s)
▶ The transfer function G(s) =
F (s)
X (s) 1
G(s) = =
2 +
F (s) Ms fv s + K
Block Diagram of the Transfer Function
from V (s) to VC (s)

Al-Motasem Aldaoudeyeh, PhD Slide 131 of 242 BasmahWantsMeHappy

Transfer Functions of One Degree-of-Freedom Mechanical Systems
[Nis11; pp. 63–64]
In the following system, calculate the trans-
fer function. Assume the input is f (t) and
the output is x2 (t)

Al-Motasem Aldaoudeyeh, PhD Slide 132 of 242 BasmahWantsMeHappy

Transfer Functions of One Degree-of-Freedom Mechanical Systems
[Nis11; pp. 63–64]
▶ Convert the system to the frequency
domain
d fv is replaced with s fv
d K remains K
d M is replaced with s2 M
d all displacement and force notation are
replaced with capital letters

Al-Motasem Aldaoudeyeh, PhD Slide 133 of 242 BasmahWantsMeHappy

Transfer Functions of Two Degrees-of-Freedom Mechanical Systems [Nis11; p. 69]
▶ Draw the free body diagram of the system for x1 (t) = 0 (motion of mass 2 only) and
x2 (t) = 0 (motion of mass 1 only)

Al-Motasem Aldaoudeyeh, PhD Slide 134 of 242 BasmahWantsMeHappy

Transfer Functions of Two Degrees-of-Freedom Mechanical Systems [Nis11; p. 69]
▶ The sum of forces on mass 1 must be zero. Thus,
F (s) − s X1 (s) − s X1 (s) − s2 X1 (s) − 1X1 (s) − s X1 (s)
+ s X2 (s) + s X2 (s) + 1X2 (s) + s X2 (s) = 0

+ s X1 (s) + s X1 (s) + s2 X1 (s) + 1X1 (s) + s X1 (s)

− s X2 (s) − s X2 (s) − 1X2 (s) − s X2 (s) = F (s)

Al-Motasem Aldaoudeyeh, PhD Slide 135 of 242 BasmahWantsMeHappy

Transfer Functions of Two Degrees-of-Freedom Mechanical Systems [Nis11; p. 69]
▶ Rearranging
 2 
+ s + 3s + 1 X1 (s) − [3s + 1] X2 (s) = F (s)

Al-Motasem Aldaoudeyeh, PhD Slide 136 of 242 BasmahWantsMeHappy

Transfer Functions of Two Degrees-of-Freedom Mechanical Systems [Nis11; p. 69]
▶ Draw the free body diagram of the system for x2 (t) = 0 (motion of mass 1 only) and
x1 (t) = 0 (motion of mass 2 only)

Al-Motasem Aldaoudeyeh, PhD Slide 137 of 242 BasmahWantsMeHappy

Transfer Functions of Two Degrees-of-Freedom Mechanical Systems [Nis11; p. 69]
▶ The sum of forces on mass 2 must be zero. Thus,
− s X2 (s) − s X2 (s) − s2 X2 (s) − 1X2 (s) − s X2 (s) − s X2 (s)
+ s X1 (s) + s X1 (s) + 1X1 (s) + s X1 (s) = 0

+ s X2 (s) + s X2 (s) + s2 X2 (s) + 1X2 (s) + s X2 (s) + s X2 (s)

− s X1 (s) − s X1 (s) − 1X1 (s) − s X1 (s) = 0

Al-Motasem Aldaoudeyeh, PhD Slide 138 of 242 BasmahWantsMeHappy

Transfer Functions of Two Degrees-of-Freedom Mechanical Systems [Nis11; p. 69]
▶ Rearranging
− [3s + 1] X1 (s) + s2 + 4s + 1 X2 (s) = 0
 

Al-Motasem Aldaoudeyeh, PhD Slide 139 of 242 BasmahWantsMeHappy

Transfer Functions of Two Degrees-of-Freedom Mechanical Systems [Nis11; p. 69]
▶ Rewriting the two equations in matrix format
 2     
s + 3s + 1  − [3s + 1]  X1 (s) F (s)
=
− [3s + 1] s2 + 4 s + 1 X2 (s) 0

Al-Motasem Aldaoudeyeh, PhD Slide 140 of 242 BasmahWantsMeHappy

Transfer Functions of Two Degrees-of-Freedom Mechanical Systems [Nis11; p. 69]
▶ We solve for X2 (s) using Cramer’s rule
 2 
s + 3s + 1 F (s)
− [3s + 1] 0
X2 (s) =  2 
s + 3s + 1  − [3s + 1] 
− [3s + 1] s2 + 4s + 1

Al-Motasem Aldaoudeyeh, PhD Slide 141 of 242 BasmahWantsMeHappy

Transfer Functions of Two Degrees-of-Freedom Mechanical Systems [Nis11; p. 69]
▶ The solution of X2 (s) is
s+1
3
X2 (s) = F (s) 3 +7 2 +5
s s
( s s + 1)

Al-Motasem Aldaoudeyeh, PhD Slide 142 of 242 BasmahWantsMeHappy

Transfer Functions of Two Degrees-of-Freedom Mechanical Systems [Nis11; p. 69]
X2 (s)
▶ The transfer function G(s) = is
F (s)
thus
X2 (s) 3 s+1
G(s) = =
3 +7 2 +5
F (s) s s
( s s + 1)

Al-Motasem Aldaoudeyeh, PhD Slide 143 of 242 BasmahWantsMeHappy

Try It in MATLAB
syms s X_1 X_2 F_s

=
eq1 = s*X_1 =
s*X_1 =
s^2*X_1 =
1*X_1 = s*X_1 ...
+s*X_2 + s*X_2 + 1*X_2 + s*X_2 == F_s; =
=
eq1 = 1*eq1;
eq1 = collect(eq1, {'X_1', 'X_2'});
pretty(eq1)

=
eq2 = s*X_2 =
s*X_2 =
s^2*X_2 =
1*X_2 = s*X_2 = s*X_2 ...
+s*X_1 + s*X_1 + 1*X_1 + s*X_1 == 0;
=
eq2 = 1*eq2;
eq2 = collect(eq2, {'X_1', 'X_2'});
pretty(eq2)

num = det([(s^2 + 3*s + 1), F_s; ...

=(3*s + 1), 0]);
den = =
det([(s^2 + 3*s + 1), (3*s + 1); ...
=_(3*s+ 1), (s^2 + 4*s + 1)]);
X 2 = num/den;
pretty(X_2)

X_2_F = X_2/F_s;

Al-Motasem Aldaoudeyeh, PhD Slide 144 of 242 BasmahWantsMeHappy

pretty(X_2_F)

Al-Motasem Aldaoudeyeh, PhD Slide 145 of 242 BasmahWantsMeHappy

A Short-Cut to Equations of Translational Motion [Nis11; p. 67]
▶ In the mechanical system of the previous example it is possible to obtain exactly the
same system of equations as follows
     
Sum of
Sum of
 X1 (s) −  Impedances  X2 (s) =  Sum of Forces 
     
+ 
 Impedances   Common to X1   Applied at X1 
Connected to X1
and X2
     
Sum of
 Impedances  Sum of
 X1 (s) +  Impedances  X2 (s) =  Sum of Forces 
   
− Common to X1     Applied at X2 
Connected to X2
and X2

Al-Motasem Aldaoudeyeh, PhD Slide 146 of 242 BasmahWantsMeHappy

A Short-Cut to Equations of Translational Motion [Nis11; p. 67]

Al-Motasem Aldaoudeyeh, PhD Slide 147 of 242 BasmahWantsMeHappy

A Short-Cut to Equations of Motion [Nis11; p. 67]
▶ For X1 (s)
s2 + 3s + 1 X1 (s)
 
− [3s + 1]X2 (s) = F (s)
for X2 (s)
− [3s + 1]X1 (s)+ s2 + 4s + 1 X2 (s) = 0
 

Al-Motasem Aldaoudeyeh, PhD Slide 148 of 242 BasmahWantsMeHappy

Transfer Functions of Three Degrees-of-Freedom Mechanical Systems [Nis11; p. 101]
▶ In the following system, calculate the transfer function. Assume the input is f (t)
and the output is x3 (t). Also, assume no friction at below all masses is zero.

Al-Motasem Aldaoudeyeh, PhD Slide 149 of 242 BasmahWantsMeHappy

Transfer Functions of Three Degrees-of-Freedom Mechanical Systems [Nis11; p. 101]
▶ Convert all quantities into frequency domain. fv is replaced with s fv , K remains K,
M is replaced with s2 M . All displacement and force notation are replaced with
capital letters

Al-Motasem Aldaoudeyeh, PhD Slide 150 of 242 BasmahWantsMeHappy

Transfer Functions of Three Degrees-of-Freedom Mechanical Systems [Nis11; p. 101]
▶ For X1 (s)
 2 
4s + 2s + 6 X1 (s) − [2s]X2 (s) − [0]X3 (s) = 0
for X2 (s)
s2 + 2s + 2s + 6 X2 (s)
 
− [2s]X1 (s)+ 4 − [6]X3 (s) = F (s)
for X3 (s)
s2 + 2s + 6 X3 (s) = 0
 
− [0]X1 (s) − [6]X2 (s)+ 4

Al-Motasem Aldaoudeyeh, PhD Slide 151 of 242 BasmahWantsMeHappy

Transfer Functions of Three Degrees-of-Freedom Mechanical Systems [Nis11; p. 101]
▶ Rewriting the three equations in matrix format
 2     
4s + 2s + 6
 2 − [2s]  − [0] X1 (s) 0
 − [2s] 4s + 4s + 6
 2 − [6]  X2 (s) = F (s)
   
− [0] − [6] 4s + 2s + 6 X3 (s) 0

Al-Motasem Aldaoudeyeh, PhD Slide 152 of 242 BasmahWantsMeHappy

Transfer Functions of Three Degrees-of-Freedom Mechanical Systems [Nis11; p. 101]
▶ We solve for X3 (s) using Cramer’s rule
 2 
4s + 2s + 6
 2 − [2s]  0

− [2s] 4s + 4s + 6 F (s)
− [0] − [6] 0
X3 (s) =  2 
4s + 2s + 6
 2 − [2s]  − [0]
− [2s] 4s + 4s + 6
 2 − [6] 
− [0] − [6] 4s + 2s + 6

Al-Motasem Aldaoudeyeh, PhD Slide 153 of 242 BasmahWantsMeHappy

Transfer Functions of Three Degrees-of-Freedom Mechanical Systems [Nis11; p. 101]
▶ The solution of X3 (s) is
2


12 2s + s + 3
X3 (s) = F (s)
s (64s5 + 128s4 + 352s3 + 392s2 + 384s + 216)
3
= F (s) 3 + 12 2 + 26
s s
(8 s s + 18)

Al-Motasem Aldaoudeyeh, PhD Slide 154 of 242 BasmahWantsMeHappy

Transfer Functions of Three Degrees-of-Freedom Mechanical Systems [Nis11; p. 101]
X3 (s)
▶ The transfer function G(s) = is thus
F (s)
X3 (s) 3
G(s) = =
3 + 12 2 + 26
F (s) s s
(8 s s + 18)

Al-Motasem Aldaoudeyeh, PhD Slide 155 of 242 BasmahWantsMeHappy

Try It in MATLAB
syms s X_1 X_2 X_3 F_s

eq1 = (4*s^2 + 2*s + 6)*X_1 =

2*s*X_2 0*X_3 == 0; =
eq1 = collect(eq1, {'X_1', 'X_2', 'X_3'});
pretty(eq1)

=
eq2 = 2*s*X_1 + (4*s^2 + 2*s + 2*s + 6)*X_2 = 6*X_3 == F_s;
eq2 = collect(eq2, {'X_1', 'X_2', 'X_3'});
pretty(eq2)

eq3 = 0*X_1 =
6*X_2 + (4*s^2 + 2*s + 6)*X_3 == 0;
eq3 = collect(eq3, {'X_1', 'X_2', 'X_3'});
pretty(eq3)

=
num = det([(4*s^2 + 2*s + 6), ( 2*s), 0; ...
=
( 2*s), (4*s^2 + 4*s + 6), F_s; ...
= = =
( 0), ( 6), ( 0)]);
=
den = det([(4*s^2 + 2*s + 6), ( 2*s), ( 0); ... =
= =
( 2*s), (4*s^2 + 4*s + 6), ( 6); ...
= =
( 0), ( 6), (4*s^2 + 2*s + 6)]);

X_3 = num/den;

Al-Motasem Aldaoudeyeh, PhD Slide 156 of 242 BasmahWantsMeHappy

X_3 = simplify(X_3);
pretty(X_3)

X_3_F = X_3/F_s;
pretty(X_3_F)

Al-Motasem Aldaoudeyeh, PhD Slide 157 of 242 BasmahWantsMeHappy

 Rotational Mechanical Systems Transfer Functions
▶ Rotational networks contain three main passive components [Nis11; pp. 69–74]
d dampers (which account for friction)
d springs
d inertia

▶ Passive means no internal source of energy

Al-Motasem Aldaoudeyeh, PhD Slide 158 of 242 BasmahWantsMeHappy

Torque-Angular Displacement Relationships for Mechanical Components [Nis11;
pp. 69–74]

Time Domain Frequency Domain

Torque-Angular Torque-Angular Impedance
Component Displacement Component Displacement T (s)
Relationship Relationship ZM (s) = θ(s)

dθ(t) T (s) = s Dθ(s) sD

T (t) = D
dt

T (t) = Kθ(t) T (s) = Kθ(s) K

Torque-Angular Displacement Relationships for Mechanical Components [Nis11;
pp. 69–74]

Time Domain Frequency Domain

Torque-Angular Torque-Angular Impedance
Component Displacement Component Displacement T (s)
Relationship Relationship ZM (s) = θ(s)

d2 θ(t) T (s) = s2 Jθ(s) s2 J

T (t) = J
d t2
Transfer Functions of Two Degrees-of-Freedom Rotational Mechanical Systems
[Nis11; p. 74]
▶ In the following system, calculate the transfer function. Assume the input is T (t)
and the output is θ2 (t)

Al-Motasem Aldaoudeyeh, PhD Slide 161 of 242 BasmahWantsMeHappy

Transfer Functions of Two Degrees-of-Freedom Rotational Mechanical Systems
[Nis11; p. 74]
▶ Redraw the circuit with an inertia of 0 kg·m2 at θ2 (t)

Al-Motasem Aldaoudeyeh, PhD Slide 162 of 242 BasmahWantsMeHappy

Transfer Functions of Two Degrees-of-Freedom Rotational Mechanical Systems
[Nis11; p. 74]
▶ Convert all quantities into frequency domain. D is replaced with s D, K remains K,
J is replaced with s2 J. All angular displacement and force notation are replaced
with capital letters

Al-Motasem Aldaoudeyeh, PhD Slide 163 of 242 BasmahWantsMeHappy

Transfer Functions of Two Degrees-of-Freedom Rotational Mechanical Systems
[Nis11; p. 74]
▶ Draw the free body diagram for the inertia at θ1 (s)
showing all torques
▶ The sum of torques must be zero. Thus,
T (s) − s2 θ1 (s) − sθ1 (s) − 1 × θ1 (s)
+ sθ2 (s) + 1 × θ2 (s) = 0

rearranging
2
+ s θ1 (s) + sθ1 (s) + 1 × θ1 (s)

− sθ2 (s) − 1 × θ2 (s) = T (s)

Al-Motasem Aldaoudeyeh, PhD Slide 164 of 242 BasmahWantsMeHappy

Transfer Functions of Two Degrees-of-Freedom Rotational Mechanical Systems
[Nis11; p. 74]
▶ Collecting θ1 (s) and θ2 (s)
 2 
s + s + 1 θ1 (s) − [s + 1] θ2 (s) = T (s)

Al-Motasem Aldaoudeyeh, PhD Slide 165 of 242 BasmahWantsMeHappy

Transfer Functions of Two Degrees-of-Freedom Rotational Mechanical Systems
[Nis11; p. 74]
▶ Draw the free body diagram for the inertia at θ2 (s)
showing all torques
▶ The sum of torques must be zero. Thus,
− 0 s2 θ2 (s) − 2 × sθ2 (s) − 2 × 1 × θ2 (s)
+ sθ1 (s) + 1 × θ1 s = 0

rearranging
2
+ 0 s θ2 (s) + 2sθ2 (s) + 2 × θ2 (s)
− sθ1 (s) − 1 × θ1 (s) = 0

Al-Motasem Aldaoudeyeh, PhD Slide 166 of 242 BasmahWantsMeHappy

Transfer Functions of Two Degrees-of-Freedom Rotational Mechanical Systems
[Nis11; p. 74]
▶ Collecting θ1 (s) and θ2 (s)
− [s + 1] θ1 (s) + [2s + 2] θ2 (s) = 0

Al-Motasem Aldaoudeyeh, PhD Slide 167 of 242 BasmahWantsMeHappy

Transfer Functions of Two Degrees-of-Freedom Rotational Mechanical Systems
[Nis11; p. 74]
▶ Rewriting the two equations in matrix format
 2     
s +s+1 − [s + 1] θ1 (s) T (s)
=
− [s + 1] [2s + 2] θ2 (s) 0

Al-Motasem Aldaoudeyeh, PhD Slide 168 of 242 BasmahWantsMeHappy

Transfer Functions of Two Degrees-of-Freedom Rotational Mechanical Systems
[Nis11; p. 74]
▶ We solve for θ2 (s) using Cramer’s rule
 2 
s + s + 1 T (s)
− [s + 1] 0
θ2 (s) =  2 
s + s + 1 − [s + 1]
− [s + 1] [2s + 2]

Al-Motasem Aldaoudeyeh, PhD Slide 169 of 242 BasmahWantsMeHappy

Transfer Functions of Two Degrees-of-Freedom Rotational Mechanical Systems
[Nis11; p. 74]
θ2 (s)
▶ The transfer function G(s) = is thus
T (s)
θ2 (s) 1
G(s) = =
T (s) s2 + s + 1
2

Al-Motasem Aldaoudeyeh, PhD Slide 170 of 242 BasmahWantsMeHappy

Try It in MATLAB
syms s theta_1 theta_2 T_s

eq1 = (s^2 + s + 1)*theta_1 =

(s + 1)*theta_2 == T_s;
eq1 = collect(eq1, {'theta_1', 'theta_2'});
pretty(eq1)

eq2 = =
(s + 1)*theta_1 + (s + 1)*theta_2 == 0;
eq2 = collect(eq2, {'theta_1', 'theta_2'});
pretty(eq2)

num = det([+(s^2 + s + 1), T_s; ...

=(s + 1), 0]);
den = =
det([+(s^2 + s + 1), (s + 1); ...
=(s + 1), +(2*s + 2)]);

theta_2 = num/den;
theta_2 = simplify(theta_2);
pretty(theta_2)

theta_2_T = theta_2/T_s;
pretty(theta_2_T)

Al-Motasem Aldaoudeyeh, PhD Slide 171 of 242 BasmahWantsMeHappy

A Short-Cut to Equations of Rotational Motion [Nis11; p. 73]
▶ In the rotational mechanical system of the previous example it is possible to obtain
exactly the same system of equations as follows
     
Sum of
Sum of
 θ1 (s) −  Impedances  θ2 (s) = Sum of Torques
     
+ 
 Impedances   Common to θ1   Applied at θ1 
Connected to θ1
and θ2
     
Sum of
 Impedances  Sum of
 θ1 (s) +  Impedances  θ2 (s) = Sum of Torques
   
−  Common to θ1     Applied at θ2 
Connected to θ2
and θ2

Al-Motasem Aldaoudeyeh, PhD Slide 172 of 242 BasmahWantsMeHappy

 Rotational Mechanical Systems Transfer Functions with
Gears
▶ It is common for rotational mechanical systems to contain gear trains (commonly
referred to as gears) [Nis11; pp. 74–77]
d gears allow us to obtain higher torques or different rotational speed

Al-Motasem Aldaoudeyeh, PhD Slide 173 of 242 BasmahWantsMeHappy

Torque and Angle Conversion in Gears [Nis11; pp. 74–77]
▶ Assuming the gear train is ideal (i.e., no
mechanical losses at gear 1 or gear 2),
the input andm output power are equal.
Thus
T1 ω1 = T2 ω2
thus
T2 ω1
=
T1 ω2 Input Output
the number of teeth is inversely proportional Gear 1 Gear 2
(Drive Gear) (Driven Gear)
to ω1 and ω2 . Thus
T2 ω1 N2
= = (22)
T1 ω2 N1

Al-Motasem Aldaoudeyeh, PhD Slide 174 of 242 BasmahWantsMeHappy

Torque and Angle Conversion in Gears [Nis11; pp. 74–77]
▶ As gears rotate, the angular distance
traveled by each gear is proportional to the
number of teeth. Thus,
r1 θ1 = r2 θ2
thus
θ1 r2
=
θ2 r1
the number of teeth is proportional to r1 and Input Output
r2 . Thus Gear 1 Gear 2
(Drive Gear) (Driven Gear)
θ1 r2 N2
= = (23)
θ2 r1 N1

Al-Motasem Aldaoudeyeh, PhD Slide 175 of 242 BasmahWantsMeHappy

Torque and Angle Conversion in Gears [Nis11; pp. 74–77]
▶ Eqs. (22) and (23) yields

T2 θ1 N2
= = (24)
T1 θ2 N1

Input Output
Transfer Functions of Angular Displace- Gear 1 Gear 2
(Drive Gear) (Driven Gear)
ment and Torque for Gear Trains

Al-Motasem Aldaoudeyeh, PhD Slide 176 of 242 BasmahWantsMeHappy

Reflecting Mechanical Impedances in Gears [Nis11; pp. 74–77]
▶ The equation of motion at θ2 (s) is
s2 J + s D + K θ2 (s) = T2 (s)

N2
according Eq. (24), T2 (s) = T1 (s)
N1
N2
s2 J + s D + K θ2 (s) =


T1 (s)
N1
N1
according Eq. (24), θ2 (s) = θ1 (s)
N2

N1 N2
s2 J + s D + K θ1 (s) = T1 (s)
N2 N1

Al-Motasem Aldaoudeyeh, PhD Slide 177 of 242 BasmahWantsMeHappy

Reflecting Mechanical Impedances in Gears [Nis11; pp. 74–77]
N1
▶ Multiplying both sides by , we get
N2
"  2  2
2 N1 N1
s J + sD
N2 N2
 2 #
N1
+K θ1 (s) = T1 (s)
N2

Al-Motasem Aldaoudeyeh, PhD Slide 178 of 242 BasmahWantsMeHappy

Reflecting Mechanical Impedances in Gears [Nis11; pp. 74–77]
▶ Mechanical impedances are thus converted as follows
2
Number of Teeth


ZM, new = ZM, old 

in the Destination Shaft  (25)
Number of Teeth
in the Source Shaft

Al-Motasem Aldaoudeyeh, PhD Slide 179 of 242 BasmahWantsMeHappy

Summary of Conversion of Quantities Across Gears [Nis11; pp. 74–77]
Quantity Conversion Rule Notes
Number of Teeth of Gear
  
Torque increases

Torque Tnew = Told  in the Destination Shaft 
 Number of Teeth of Gear with teeth in gears
 in the Source Shaft
Number of Teeth of Gear 2
 
 Mechanical
Mechanical

ZM, new = ZM, old  in the Destination Shaft  impedance increases
Impedance  Number of Teeth of Gear

in the Source Shaft with teeth in gears




Number of Teeth of Gear
 Angular
Angular 
in the Source Shaft displacement
θnew = θold  
Displacement 
 Number of Teeth of Gear decreases with teeth
 in the Destination Shaft in gears
Transfer Functions of Three Degrees-of-Freedom Rotational Mechanical Systems with
Gears [Nis11; p. 102]
▶ In the following system, obtain the transfer function from T (t) to θL (t)

Al-Motasem Aldaoudeyeh, PhD Slide 181 of 242 BasmahWantsMeHappy

Transfer Functions of Three Degrees-of-Freedom Rotational Mechanical Systems with
Gears [Nis11; p. 102]
▶ Convert all quantities into frequency domain. D is replaced with s D, K remains K,
J is replaced with s2 J. All angular displacement and force notation are replaced
with capital letters

Al-Motasem Aldaoudeyeh, PhD Slide 182 of 242 BasmahWantsMeHappy

Transfer Functions of Three Degrees-of-Freedom Rotational Mechanical Systems with
Gears [Nis11; p. 102]
▶ Convert input torque to the region in the middle using Eq. (24)
Number of Teeth of Gear
 

Tnew = Told =  in the Destination Shaft 

Number of Teeth of Gear
 Source Shaft
in the
33
Tnew = T (s) = 3T (s)
11

we convert the impedance 0.04s using Eq. (25)

Number of Teeth of Gear 2
 

ZM, new = ZM, old  in the Destination Shaft 

Number of Teeth of Gear
in the Source Shaft
 2
50
= 0.04s = s
10

Al-Motasem Aldaoudeyeh, PhD Slide 183 of 242 BasmahWantsMeHappy

Transfer Functions of Three Degrees-of-Freedom Rotational Mechanical Systems with
Gears [Nis11; p. 102]
▶ Redraw the diagram
▶ We add masses of 0 s2 inertia to conveniently write equations of motion

Al-Motasem Aldaoudeyeh, PhD Slide 184 of 242 BasmahWantsMeHappy

Transfer Functions of Three Degrees-of-Freedom Rotational Mechanical Systems with
Gears [Nis11; p. 102]
▶ For inertia at θ3
 2 
s + 2s θ3 (s) − [2s]θ2 (s) − [0]θ1 (s) = 3T (s)
for inertia at θ2
s2 + 2s + 3 θ2 (s)
 
− [2s]θ3 (s)+ 0 − [3]θ1 (s) = 0
for inertia at θ1
s2 + 3 + s θ1 (s) = 0
 
− [0]θ3 (s) − [3]θ2 (s)+ 0

Al-Motasem Aldaoudeyeh, PhD Slide 185 of 242 BasmahWantsMeHappy

Transfer Functions of Three Degrees-of-Freedom Rotational Mechanical Systems with
Gears [Nis11; p. 102]
▶ Rewriting all equations in matrix format
  2     
s + 2s  2− [2s]  − [0] θ3 (s) 3T (s)
 − [2s] 0 s + 2s + 3
 2− [3] 
θ2 (s) =  0 
− [0] − [3] 0s + 3 + s θ1 (s) 0

Al-Motasem Aldaoudeyeh, PhD Slide 186 of 242 BasmahWantsMeHappy

Transfer Functions of Three Degrees-of-Freedom Rotational Mechanical Systems with
Gears [Nis11; p. 102]
▶ Using Cramer’s rule to calculate θ2 (s)
 2 
s + 2s  − [2s]  3T (s)
2
− [2s] 0 s + 2s + 3 0

− [0] − [3] 0
θ1 (s) =  2 
s + 2s  − [2s]  − [0]
2
− [2s] 0 s + 2s + 3
 2− [3] 
− [0] − [3] 0s + 3 + s

18
= T (s) 2 +9
s (2 s s + 6)

Al-Motasem Aldaoudeyeh, PhD Slide 187 of 242 BasmahWantsMeHappy

Transfer Functions of Three Degrees-of-Freedom Rotational Mechanical Systems with
Gears [Nis11; p. 102]
▶ From the original figure and Eq. (24) we see that
Number of Teeth of Gear
 

θnew = θold  in the Source Shaft 

Number of Teeth of Gear
 in the Destination Shaft
50
θL (s) = θ1 (s) = 5θ1 (s)
10

T (s)
thus, the transfer function G(s) = becomes
θL (s)
θL (s) 90
=
T (s) s (2
2 +9
s s + 6)

Al-Motasem Aldaoudeyeh, PhD Slide 188 of 242 BasmahWantsMeHappy

Try It in MATLAB
syms s theta_3 theta_2 theta_1 T_s

N_1 = 11; N_2 = 33;

N_3 = 50; N_4 = 10;

eq1 = (s^2 + 2*s)*theta_3 =

(2*s)*theta_2 =
(0)*theta_1 == (N_2/N_1)*T_s;
eq1 = collect(eq1, {'theta_1', 'theta_2', 'theta_3'});
pretty(eq1)

=
eq2 = (2*s)*theta_3 + (0*s^2 + 2*s + 3)*theta_2 =
(3)*theta_1 == 0;
eq2 = collect(eq2, {'theta_1', 'theta_2', 'theta_3'});
pretty(eq2)

Z_M_new = 0.04*s*(N_3/N_4)^2;
eq3 = = (0)*theta_3 =
(3)*theta_2 + (0*s^2 + 3 + Z_M_new)*theta_1 == 0;
eq3 = collect(eq3, {'theta_1', 'theta_2', 'theta_3'});
pretty(eq3)

=
num = det([(s^2 + 2*s), (2*s), 3*T_s; ...
=(2*s), (2*s + 3), 0; ...
= (0), =(3), 0]);
=
den = det([(s^2 + 2*s), (2*s), (0); ... =
Al-Motasem Aldaoudeyeh, PhD Slide 189 of 242 BasmahWantsMeHappy
=(3), (2*s + 3), =(2*s); ...
=(0), =(3), (s + 3)]);
theta_1 = num/den;
theta_1 = simplify(theta_1);
pretty(theta_1)

theta_L = theta_1*(N_3/N_4);
pretty(theta_L)

theta_L_T = theta_L/T_s;
pretty(theta_L_T)

Al-Motasem Aldaoudeyeh, PhD Slide 190 of 242 BasmahWantsMeHappy

Transfer Functions of One Degree-of-Freedom Rotational Mechanical Systems with
Gears [Nis11; p. 102]
▶ In the following figure, calculate the
transfer function assuming the input is
T (t) and the output is θ2 (t)

Al-Motasem Aldaoudeyeh, PhD Slide 191 of 242 BasmahWantsMeHappy

Transfer Functions of One Degree-of-Freedom Rotational Mechanical Systems with
Gears [Nis11; p. 102]
▶ Convert all quantities into frequency
domain. D is replaced with s D, K
remains K, J is replaced with s2 J. All
angular displacement and force
notation are replaced with capital letters

Al-Motasem Aldaoudeyeh, PhD Slide 192 of 242 BasmahWantsMeHappy

Transfer Functions of One Degree-of-Freedom Rotational Mechanical Systems with
Gears [Nis11; p. 102]
▶ Reflecting impedances to the region in
the middle. From Eq. (25)
ZM, new = ZM, old
2
Number of Teeth


×  in the Destination Shaft 

Number of Teeth
in the Source Shaft

Al-Motasem Aldaoudeyeh, PhD Slide 193 of 242 BasmahWantsMeHappy

Transfer Functions of One Degree-of-Freedom Rotational Mechanical Systems with
Gears [Nis11; p. 102]
▶ For the 200 s2 impedance above
" #2
5
ZM , 1 = 200 s2 = 8 s2
25

for the 1000 s impedance above

" #2
5
ZM , 2 = 1000 s = 40 s
25

Al-Motasem Aldaoudeyeh, PhD Slide 194 of 242 BasmahWantsMeHappy

Transfer Functions of One Degree-of-Freedom Rotational Mechanical Systems with
Gears [Nis11; p. 102]
▶ For the 200 s2 impedance below
" #2
5
ZM, 3 = 200 s2 = 2 s2
50

for the 250 impedance below

" #2
5
ZM, 4 = 250 = 2.5
50

Al-Motasem Aldaoudeyeh, PhD Slide 195 of 242 BasmahWantsMeHappy

Transfer Functions of One Degree-of-Freedom Rotational Mechanical Systems with
Gears [Nis11; p. 102]
▶ Redraw the diagram

Al-Motasem Aldaoudeyeh, PhD Slide 196 of 242 BasmahWantsMeHappy

Transfer Functions of One Degree-of-Freedom Rotational Mechanical Systems with
Gears [Nis11; p. 102]
▶ The impedances 3 s2 , 8 s2 , and 2 s2 are
all inertias connected directly to each
other (thus, they rotate together). The
equivalent impedance of them is
3 s2 + 8 s2 + 2 s2 = 13 s2

▶ Redraw the diagram

Al-Motasem Aldaoudeyeh, PhD Slide 197 of 242 BasmahWantsMeHappy

Transfer Functions of One Degree-of-Freedom Rotational Mechanical Systems with
Gears [Nis11; p. 102]
▶ Writing one equation of motion at θ1
2
 
13 s + 40 s + 3 + 2.5 θ1 (s) = T (s)

the angle θ1 (s) is thus

T (s)
θ1 (s) = 2 + 40
13s s + 5.5

Al-Motasem Aldaoudeyeh, PhD Slide 198 of 242 BasmahWantsMeHappy

Transfer Functions of One Degree-of-Freedom Rotational Mechanical Systems with
Gears [Nis11; p. 102]
▶ From the original figure and Eq. (24) we see that
Number of Teeth of Gear
 

θnew = θold  in the Source Shaft 

Number of Teeth of Gear
 in the Destination Shaft
5 1
θ2 (s) = θ1 (s) = θ1 (s)
50 10

T (s)
thus, the transfer function G(s) = becomes
θ2 (s)
θ2 (s) 1
=
T (s) 130
2 + 400
s s + 55

Al-Motasem Aldaoudeyeh, PhD Slide 199 of 242 BasmahWantsMeHappy

 Electromechanical Systems Transfer Functions
▶ Thus far we studied electrical and mechnical systems independently
d it is, however, common for control systems to contain hybrid electrical and mechanical
components
d such type of systems is referred to as electromechanical systems [Nis11; pp. 79–84]

Al-Motasem Aldaoudeyeh, PhD Slide 200 of 242 BasmahWantsMeHappy

Electric Motor [Nis11; pp. 79–84]
▶ From control point of view, an electric motor is an electromechanical component
which yields an angular displacement for a voltage input
d this means the electrical input is converted into mechanical output

▶ The type of motor we will study is the armature-controlled DC servomotor

Al-Motasem Aldaoudeyeh, PhD Slide 201 of 242 BasmahWantsMeHappy

Derivation of Transfer Function of Armature-Controlled DC Motor [Nis11; pp. 79–84]
▶ In the following equivalent circuit of DC motor, how do we obtain the transfer
function G(s)?

Armature Circuit Field Circuit

{ (a) schematic
R

{
ot
or
Fi ield
xe
F

(b) block diagram

Derivation of Transfer Function of Armature-Controlled DC Motor [Nis11; pp. 79–84]
▶ In the adjacent figure a field is
established by stationarry magnet or

Fi ield
xe
F

d
stationary electromagnet called fixed
field
▶ The rotating circuit is called the
armature, in which a current denoted
ia (t) flows

R
ot
or
Al-Motasem Aldaoudeyeh, PhD Slide 203 of 242 BasmahWantsMeHappy
Derivation of Transfer Function of Armature-Controlled DC Motor [Nis11; pp. 79–84]
▶ The current flowing in the armature
produces a torque given as

Fi ield
xe
F

d
Tm (t) = Kt ia (t) (26)
where Kt is a constant which depends
on the construction of the machine.
Taking the Laplace transform of Eq. (26)
Tm (s) = Kt Ia (s) (27)

R
rewriting Eq. (27)

ot
or
1
Ia (s) = Tm (s) (28)
Kt

Al-Motasem Aldaoudeyeh, PhD Slide 204 of 242 BasmahWantsMeHappy

Derivation of Transfer Function of Armature-Controlled DC Motor [Nis11; pp. 79–84]
▶ Due to the torque produced in the
armature it rotates. The back

Fi ield
xe
F

d
electromotive force (back emf) (vb (t)) is
given as
dθm (t)
vb (t) = Kb (29)
dt
where Kb is a constant which depends
on the construction of the machine.

R
ot
Taking the Laplace transform of Eq. (29)

or
Vb (s) = Kb s θm (s) (30)

Al-Motasem Aldaoudeyeh, PhD Slide 205 of 242 BasmahWantsMeHappy

Derivation of Transfer Function of Armature-Controlled DC Motor [Nis11; pp. 79–84]
▶ We now convert all quantities to
frequency domain

Fi ield
xe
F

d
R
ot
or
Al-Motasem Aldaoudeyeh, PhD Slide 206 of 242 BasmahWantsMeHappy
Derivation of Transfer Function of Armature-Controlled DC Motor [Nis11; pp. 79–84]
▶ The relationship between E (s) and Ia (s) is given as
Ra Ia (s) + s La Ia (s) + Vb (s) = Ea (s) (31)
1
substituting Eq. (28) (Ia (s) = Tm (s)) and Eq. (30) (Vb (s) = Kb s θm (s)) in Eq. (31)
Kt
yields
Ra + s La )Tm (s)
(
+ Kb s θm (s) = Ea (s) (32)
Kt

Al-Motasem Aldaoudeyeh, PhD Slide 207 of 242 BasmahWantsMeHappy

Derivation of Transfer Function of Armature-Controlled DC Motor [Nis11; pp. 79–84]
▶ A typical mechanical load is shown below
d in this equivalent circuit Jm is the total inertia of the motor and load while Dm is the total friction
of the motor and load

▶ The torque Tm (s) is given as

Tm (s) = s2 Jm + s Dm θm (s)


(33)

Al-Motasem Aldaoudeyeh, PhD Slide 208 of 242 BasmahWantsMeHappy

Derivation of Transfer Function of Armature-Controlled DC Motor [Nis11; pp. 79–84]
▶ Now we substitute Eq. (33) in Eq. (32)
 2

(Ra + s La ) s Jm + s Dm θm (s)
+ Kb s θm (s) = Ea (s) (34)
Kt
when La is very small compared to Ra (usually valid assumption for DC motors),
Eq. (34) becomes
 
Ra
(s Jm + Dm ) + Kb s θm (s) = Ea (s) (35)
Kt

Al-Motasem Aldaoudeyeh, PhD Slide 209 of 242 BasmahWantsMeHappy

Derivation of Transfer Function of Armature-Controlled DC Motor [Nis11; pp. 79–84]
▶ We now convert all quantities to frequency

Fi ield
domain

xe
F

d
θm
G(s) =
Ea (s)
Kt
 a Jm
R
(36)

R
=

ot
 

or
1 Kt Kb
s s+ Dm +
Jm Ra

Al-Motasem Aldaoudeyeh, PhD Slide 210 of 242 BasmahWantsMeHappy

Derivation of Transfer Function of Armature-Controlled DC Motor [Nis11; pp. 79–84]
▶ It is also possible to rewrite Eq. (36) as follows

Fi ield
xe
F

d
θm (s) K
G(s) = = (37)
Ea (s) s [s + α ]
Kt
K = (38)
Ra Jm

R
 
1 Kt Kb

ot
or
α= Dm + (39)
Jm Ra

Al-Motasem Aldaoudeyeh, PhD Slide 211 of 242 BasmahWantsMeHappy

Derivation of Transfer Function of Armature-Controlled DC Motor [Nis11; pp. 79–84]
▶ The rotor itself in the DC machine has an impedance for inertia and friction s2 Ja and
s Da
▶ The impedance of the inertia and friction of the load are s2 JL and s DL

Fi ield
xe
F

d
Rotor

{
Derivation of Transfer Function of Armature-Controlled DC Motor [Nis11; pp. 79–84]
▶ It is frequent that we connect the motor to load by gear train

Fi ield
xe
F

d
Rotor

{
Derivation of Transfer Function of Armature-Controlled DC Motor [Nis11; pp. 79–84]

 2 )
N1 The total inertia
Jm = Ja + JL (40)
N2 as seen by the rotor

Fi ield
xe
F

d
Rotor

{
Derivation of Transfer Function of Armature-Controlled DC Motor [Nis11; pp. 79–84]

 2 )
N1 The total friction
Dm = Da + DL (41)
N2 as seen by the rotor

Fi ield
xe
F

d
Rotor

{
Derivation of Transfer Function of Armature-Controlled DC Motor [Nis11; pp. 79–84]
▶ In Eqs. (38) and (39), we need to determine the

Fi ield
constants Kt and Kb

xe
F

d
θm K
G(s) = =
Ea (s) s [s + α ]
Kt
K =
Ra Jm

R
ot
or
 
1 Kt Kb
α= Dm +
Jm Ra

Al-Motasem Aldaoudeyeh, PhD Slide 216 of 242 BasmahWantsMeHappy

Derivation of Transfer Function of Armature-Controlled DC Motor [Nis11; pp. 79–84]
▶ To determine Kt and Kb , we rewrite Eq. (32)
Ra + s La )Tm (s)
(
+ Kb s θm (s) = Ea (s)
Kt
for La = 0 , we get
Ra
Tm (s) + Kb s θm (s) = ea (t)
Kt
taking the inverse Laplace transform
Ra
Tm (t) + Kb s θm (t) = ea (t)
Kt
since s θm (t) = ωm (t)
Ra
Tm (t) + Kb ωm (t) = ea (t)
Kt

Al-Motasem Aldaoudeyeh, PhD Slide 217 of 242 BasmahWantsMeHappy

Derivation of Transfer Function of Armature-Controlled DC Motor [Nis11; pp. 79–84]
▶ When e(t) is pure DC then Tm (t) and ωm (t) are constant. Thus we drop t
since s θm (t) = ωm (t)
Ra
Tm + Kb ωm = ea
Kt

▶ Finally, we solve for Tm to obtain

Kt Kb Kt
Tm = − ωm + ea (42)
Ra Ra

Al-Motasem Aldaoudeyeh, PhD Slide 218 of 242 BasmahWantsMeHappy

Derivation of Transfer Function of Armature-Controlled DC Motor [Nis11; pp. 79–84]
▶ Eq. (42) is actually a straight line. When ωm = 0 , we get the stall torque Tstall as
follows Eq. (42)
Kt
Tstall ea= (43)
Ra
when substituting Tm = 0 in Eq. (42), we get the no-load speed ωno−load
ea
ωno−load = (44)
Kb

Al-Motasem Aldaoudeyeh, PhD Slide 219 of 242 BasmahWantsMeHappy

Derivation of Transfer Function of Armature-Controlled DC Motor [Nis11; pp. 79–84]
▶ The parameters Kt and Kb may now be calculated
according to Eqs. (43) and (44) as follows

Kt Tstall
= (45)
Ra ea
ea
Kb = (46)
ωno−load

ea in Eqs. (45) and (46) is actually the rated voltage

Al-Motasem Aldaoudeyeh, PhD Slide 220 of 242 BasmahWantsMeHappy

Transfer Function of Armature-Controlled DC Motor [Nis11; p. 84]
For the electromechanical system in the next figure, assuming the torque-speed curve
for a given voltage of ea = 100 V is Tm = −8 ωm + 200, calculate
▶ Tstall and ωno−load
Kt
▶ The ratio and Kb
Ra
▶ The total inertia and friction as seen by the motor (Jm and Dm )
▶ The transfer function from ea (t) to θm (t)
▶ The transfer function from ea (t) to θL (t)
▶ The transfer function from ea (t) to TL (t)
▶ Assuming the input voltage is ea (t) = 37 V, does any of the results obtained before
change?

Al-Motasem Aldaoudeyeh, PhD Slide 221 of 242 BasmahWantsMeHappy

Transfer Function of Armature-Controlled DC Motor [Nis11; p. 84]

Fi ield
xe
F

d
Rotor

{
Al-Motasem Aldaoudeyeh, PhD Slide 222 of 242 BasmahWantsMeHappy
Transfer Function of Armature-Controlled DC Motor [Nis11; p. 84]
Tstall and ωno−load
▶ Tstall may be obtained by substituting ωm = 0 in the torque-speed curve
Tstall = −8 ωm + 200|ωm =0 = 200 N·m
ωno−load may be obtained by substituting Tm = 0 in the torque-speed curve and
solving for ωm
0 = −8 ωno−load + 200
ωno−load = 25 rad/s

Al-Motasem Aldaoudeyeh, PhD Slide 223 of 242 BasmahWantsMeHappy

Transfer Function of Armature-Controlled DC Motor [Nis11; p. 84]
Kt
The ratio and Kb
Ra
▶ From Eq. (45)
Kt Tstall
=
Ra ea
substituting Tstall = 200 and ea = 100

200
= = 2
100
from Eq. (46)
ea
Kb =
ωno−load
substituting ea = 100 and ωno−load = 25

100
= = 4
25

Al-Motasem Aldaoudeyeh, PhD Slide 224 of 242 BasmahWantsMeHappy

Transfer Function of Armature-Controlled DC Motor [Nis11; p. 84]
The total inertia and friction as seen by the motor (Jm and Dm )
▶ We refer the 400 kg·m2 inertia and 800 N·m·s/rad friction to the rotor of the motor

Fi ield
xe
F

d
Rotor

{
Al-Motasem Aldaoudeyeh, PhD Slide 225 of 242 BasmahWantsMeHappy
Transfer Function of Armature-Controlled DC Motor [Nis11; p. 84]
The total inertia and friction as seen by the motor (Jm and Dm )
▶ By using an expression similar to Eq. (40)
 2  2
N3 N1
Jm = Ja + JL
N4 N2
 2  2
25 20
= 1 + 400 ×
100 100

1
= 1 + 400 × = 2
400

Al-Motasem Aldaoudeyeh, PhD Slide 226 of 242 BasmahWantsMeHappy

Transfer Function of Armature-Controlled DC Motor [Nis11; p. 84]
The total inertia and friction as seen by the motor (Jm and Dm )
▶ By using an expression similar to Eq. (41)
 2  2
N3 N1
Dm = Da + DL
N4 N2
 2  2
25 20
= 5 + 800 ×
100 100

1
= 5 + 800 × = 7
400

Al-Motasem Aldaoudeyeh, PhD Slide 227 of 242 BasmahWantsMeHappy

Transfer Function of Armature-Controlled DC Motor [Nis11; p. 84]
The transfer function from ea (t) to θm
▶ From Eqs. (37) to (39)
θm K
=
Ea (s) s [s + α ]
Kt
K =
Ra Jm
 
1 Kt Kb
α= Dm +
Jm Ra

Al-Motasem Aldaoudeyeh, PhD Slide 228 of 242 BasmahWantsMeHappy

Transfer Function of Armature-Controlled DC Motor [Nis11; p. 84]
The transfer function from ea (t) to θm
▶ The constant K
Kt 1 1
K = × = 2 × = 1
Ra Jm 2

the constant α  
1 Kt
α= Dm + Kb
Jm Ra
1 15
= (7 + 2 × 4) =
2 2

Al-Motasem Aldaoudeyeh, PhD Slide 229 of 242 BasmahWantsMeHappy

Transfer Function of Armature-Controlled DC Motor [Nis11; p. 84]
The transfer function from ea (t) to θm
▶ the transfer function from ea (t) to θm thus becomes
θm (s) K
=
Ea (s) s [s + α ]
1
=  
15
s s+
2

Al-Motasem Aldaoudeyeh, PhD Slide 230 of 242 BasmahWantsMeHappy

Transfer Function of Armature-Controlled DC Motor [Nis11; p. 84]
The transfer function from ea (t) to θL (t)
▶ We obtain θL (s) by converting depending on turns ratio

Fi ield
xe
F

d
Rotor

{
Al-Motasem Aldaoudeyeh, PhD Slide 231 of 242 BasmahWantsMeHappy
Transfer Function of Armature-Controlled DC Motor [Nis11; p. 84]
The transfer function from ea (t) to θL (t)
▶ θm (s) is
1
θm (s) = Ea (s)  
15
s s+
2

the conversion of angles through gear trains is done using the relationship
Number of Teeth of Gear
 

θnew = θold  in the Source Shaft 

Number of Teeth of Gear
in the Destination Shaft
thus, we convert as follows
N1 N3
θL (s) = θm (s)
N2 N4

Al-Motasem Aldaoudeyeh, PhD Slide 232 of 242 BasmahWantsMeHappy

Transfer Function of Armature-Controlled DC Motor [Nis11; p. 84]
The transfer function from ea (t) to θL (t)
▶ θm (s) is
20 25
θL (s) = θm (s) × ×
100 100
1
= θm (s) ×
20

the angle θL (s) is thus

1 1
θL (s) = × Ea (s)  
20 15
s s+
2

Al-Motasem Aldaoudeyeh, PhD Slide 233 of 242 BasmahWantsMeHappy

Transfer Function of Armature-Controlled DC Motor [Nis11; p. 84]
The transfer function from ea (t) to θL (t)
▶ The transfer function from ea (t) to θL (t) thus becomes
1
θL (s) 20
=  
Ea (s) 15
s s+
2

Al-Motasem Aldaoudeyeh, PhD Slide 234 of 242 BasmahWantsMeHappy

Transfer Function of Armature-Controlled DC Motor [Nis11; p. 84]
The transfer function from ea (t) to TL (t)
▶ We obtain Tm (s) by multiplying θm (s) by the mechanical impedance seen by the
motor

Fi ield
xe
F

Al-Motasem Aldaoudeyeh, PhD Slide 235 of 242 BasmahWantsMeHappy

Transfer Function of Armature-Controlled DC Motor [Nis11; p. 84]
The transfer function from ea (t) to TL (t)
▶ Tm (s) becomes
1
s2 + 7 s
 
Tm (s) = Ea (s)  × 2
15
s s+
2
2
2s + 7s
= Ea (s)  
15
s s+
2

Al-Motasem Aldaoudeyeh, PhD Slide 236 of 242 BasmahWantsMeHappy

Transfer Function of Armature-Controlled DC Motor [Nis11; p. 84]
The transfer function from ea (t) to TL (t)
▶ We obtain TL (s) by converting depending on turns ratio

Fi ield
xe
F

d
Rotor

{
Al-Motasem Aldaoudeyeh, PhD Slide 237 of 242 BasmahWantsMeHappy
Transfer Function of Armature-Controlled DC Motor [Nis11; p. 84]
The transfer function from ea (t) to TL (t)
▶ Tm (s) is
1
s2 + 7 s
 
Tm (s) = Ea (s)  × 2
15
s s+
2

the conversion of torques through gear trains is done using the relationship
Number of Teeth of Gear
 

Tnew = Told  in the Destination Shaft 

Number of Teeth of Gear
in the Source Shaft
thus, we convert as follows
N2 N4
TL (s) = Tm (s)
N1 N3

Al-Motasem Aldaoudeyeh, PhD Slide 238 of 242 BasmahWantsMeHappy

Transfer Function of Armature-Controlled DC Motor [Nis11; p. 84]
The transfer function from ea (t) to TL (t)
▶ TL (s) is
100 100
TL (s) = Tm (s)
20 25

= 20 Tm (s)
substituting Tm (s)
2
2s + 7s
= 20 Ea (s)  
15
s s+
2

Al-Motasem Aldaoudeyeh, PhD Slide 239 of 242 BasmahWantsMeHappy

Transfer Function of Armature-Controlled DC Motor [Nis11; p. 84]
The transfer function from ea (t) to TL (t)
▶ The transfer function from ea (t) to TL (t) thus becomes
 2 
TL (s) 20 2 s + 7 s
=  
Ea (s) 15
s s+
2

Al-Motasem Aldaoudeyeh, PhD Slide 240 of 242 BasmahWantsMeHappy

Transfer Function of Armature-Controlled DC Motor [Nis11; p. 84]
Assuming the input voltage is ea (t) = 37 V, does any of the results obtained before
change?
▶ No. For linear, time-invariant system, the transfer function is entirely independent
of the input. Furthermore, ea = 100 V provided by the question is for the rated
voltage while ea (t) = 37 V is actually the input voltage

Al-Motasem Aldaoudeyeh, PhD Slide 241 of 242 BasmahWantsMeHappy

References - I

Al-Motasem Aldaoudeyeh, PhD Slide 242 of 242 BasmahWantsMeHappy