M.D.

Bryant ME 344 notes 03/25/08 1

Laplace Transforms &

Transfer Functions
Laplace Transforms: method for solving
differential equations, converts differential
equations in time t into algebraic equations
in complex variable s

Transfer Functions: another way to represent

system dynamics, via the s representation
gotten from Laplace transforms, or excitation
by est
M.D. Bryant ME 344 notes 03/25/08 2

Laplace Transforms
Purpose: Converts linear differential equation
in t into algebraic equation in s

• Forward transform, t → s: f(t) ⇒ F(s)

∞
F(s) = ⌡ f(t) e- st dt = L {f(t); t → s }
⌠

t= 0

Convergence/existence of integral:
|f(t)| < M e- at , a + Re(s) > 0
so that e-[a + Re(s)]t finite as t → ∞

• Inverse transform, s→ t: F(s) ⇒ f(t)

c+j∞
1
⌡ F(s) e st ds = L {F(s); s→ t }
⌠ -1
f(t) = 2πj
ω= c- j∞

Integral in complex plane, rarely do

M.D. Bryant ME 344 notes 03/25/08 3

Procedure:
• Convert Linear Differential Equations

x˙ = Ax + f (t)
dxn
d x dx n"1
dx 2

+a +!+ a +a + a x = f (t)
dt dt dt dt
n n"1 n"1 2 2 1 0

in time t into algebraic equations in

complex variable s, via forward Laplace
! transforms:

X(s)=L {x(t); t → s }, F(s)=L {f(t); t → s }

• Solve resulting algebraic equations in s,

for solution X = X(s)

• Convert solution X(s) into time function

x(t), via inverse Laplace transform:

x(t) = L-1{X(s); s→ t }
M.D. Bryant ME 344 notes 03/25/08 4

Example: Solve initial value problem,

differential equation with initial condition:
2
˙x˙ + 2 ζ ωn x˙ + ωn x = Ho us(t)

x(0) = xo , x˙ (0) = x1
• Apply Laplace transform
∞
L= ⌠⌡ { . } e- st dt to all terms in equation:
t= 0
# # #

$
t= 0
x˙˙(t)e dt + $ 2%& x˙ (t)e dt + $ & x(t)e dt
" st

t= 0 n
" st

t= 0
2

n
" st

∞
= ⌠
⌡ Ho us(t) e- st dt
t= 0
• factor constants and apply definition for
! us(t) :
∞ ∞
⌠
 2
d x - st ⌠ dx 2 ∞
e dt + 2ζω  e- st dt +ω ⌠ ⌡ x e- st dt

⌡ dt 2 n ⌡ dt n
t= 0 t= 0
t= 0
∞
= Ho ⌠
⌡ e- st dt
t= 0
M.D. Bryant ME 344 notes 03/25/08 5

∞
• Define X(s) = x(t) e- st dt
⌠
⌡
t= 0
• Integrate by parts
∞
dx - st ∞ dx - st
⌠

- st ∞
dt e |t=0 +s 
dt
⌡
e dt +2 ζωn {x(t) e |t=0
t= 0
∞ 2 ∞
⌠
+s x(t) e
⌡ - st dt } + ωn X(s) = H e- st|t=0 O

t= 0 ("s)
• Rearrange, note lim e = 0
t "#
$ st

- x˙ (0) + s (- x(0) + s !
X(s) ) + 2ζωn {- x(0) +
2 Ho
s X(s) } + ωn X(s) = s
!
 Note: L {dx/dt} = - x(0) + s X(s)
L {d2x/dt2} = L {d x˙ /dt} =- x˙ (0) + sL {dx/dt}

• Result: algebraic equation in s

! ! =
H
(s + 2"#
2
s +# 2
)X(s) O
+ x + (s + 2"# )x
s
n n 1 n 0

Note: x˙ (0) = x1 , x(0) = xo

!
M.D. Bryant ME 344 notes 03/25/08 6

• Solve
Ho
s + x1 + ( s + 2 ζ ωn) xo
X(s) = 2
2
s + 2 ζ ωn s + ω n

• Note:
o characteristic equation in denominator
o terms from initial conditions x1 ,
H
xo grouped with excitation term O

!
M.D. Bryant ME 344 notes 03/25/08 7

• Apply inverse Laplace: x(t) = L-1{X(s); s→ t }

2
Ho ωn
x(t) = 2 L { -1
2 }
ωn s (s2 + 2 ζ ωn s + ωn)
2
xo s ωn
+ 2 L { -1
2 }
ωn s2 + 2 ζ ωn s + ωn
2
x1 + 2 ζ ωn xo ωn
+ 2 L {
-1
2 }
ωn s2 + 2 ζ ωn s + ωn

o Use Laplace transform tables for L-1:

Ho e-ζ ωn t sin(ωn 1 - ζ2 t + arccos ζ )
x(t) = 2 {1 - }
1 - ζ2
ωn
2 -ζ ω t
xo ωn e n sin(ωn 1 - ζ2 t - arccos ζ )
+ 2 {- }
1-ζ2
ωn
x1 + 2 ζ ωn xo ωn e-ζ ωn t sin(ωn 1 - ζ2 t )
+ 2 { }
1-ζ 2
ωn
M.D. Bryant ME 344 notes 03/25/08 8

Transfer Functions
• Method to represent system dynamics, via
s representation from Laplace transforms.
Transfer functions show flow of signal
through a system, from input to output.

• Transfer function G(s) is ratio of output x

to input f, in s-domain (via Laplace trans.):
X(s)
G(s) =
F(s)
• Method gives system dynamics
representation equivalent to
!  Ordinary differential equations
 State equations

• Interchangeable: Can convert transfer

function to differential equations
M.D. Bryant ME 344 notes 03/25/08 9

Transfer Function G(s)

 Describes dynamics in operational sense
 Dynamics encoded in G(s)
 Ignore initial conditions (I.C. terms are
“transient” & decay quickly)
 Transfer function, for input-output
operation, deals with steady state terms

h(t) x(t)
G(s)
H(s) X(s)

 Example: Speed of automobile

o Output: speed x(t) expressed as X(s)
o Output determined by input, system
dynamics & initial conditions
o Input: gas pedal depression h(t)
o Dynamics: fuel system delivery,
motor dynamics & torque,
transmission, wheels, car translation,
mass, wind drag, etc.
o Initial conditions’ (initial speed)
influence on current speed
diminishes over time, thus ignore
M.D. Bryant ME 344 notes 03/25/08 10

Transfer Function Example: 2nd order

system
• Differential equation at steady state (can
ignore initial conditions)
2
˙x˙ + 2 ζ ωn x˙ + ωn x = h(t)

• Apply Laplace transform, define

X(s) =L{x(t); t → s}, H(s) =L{h(t); t → s }

• Result: algebraic equation

2
(s2 + 2ζωn s + ωn )X(s) = H(s)

• Transfer function G(s): ratio of output

X(s) to input H(s)

X(s) 1
G(s) = H(s) = 2
s2 + 2 ζ ωn s + ωn
• Note:
o characteristic equation in denominator
o denominator roots => eigenvalues
o transfer function’s eigenvalues called
poles
M.D. Bryant ME 344 notes 03/25/08 11

Example: First order system

"x˙ + x = f (t)
• Apply Laplace transform, define

X(s) =L{x(t); t → s}, F(s) =L{f(t); t → s }

L{τ x˙ ; t → s} +L{x(t); t → s} =L{f(t); t → s }

!
• Result: algebraic equation

sτX(s) + X(s) = F(s)

• Transfer function
1
G(s) =X(s)/F(s) =
"s +1
• Note:
o characteristic equation in denominator
o transfer function’s eigenvalues = poles
o p1 = λ1 = - 1/τ
!
M.D. Bryant ME 344 notes 03/25/08 12

Transfer Function from State Equations

• Matrix form

x˙ = Ax + f (t)
• Explicit equation form
n

x˙ = " a x + f (t)
k
j=1
jk j k

! • Define

Xk(s)=L {xk(t); t → s }, Fk(s)=L {fk(t); t → s }

! • Apply Laplace transform

sX(s) = AX(s) + F(s)

or
n

sX k
(s) = " a jk
X (s) + F (s)
j k

! j=1

!
M.D. Bryant ME 344 notes 03/25/08 13

• Rearrange matrix form:

[ sI " A] X(s) = F(s)

• Solve, via Cramer’s rule:

det{[ sI " A] }
! X (s) = kth column replaced by F ( s )

det[ sI " A]
k

• Transfer function, define which output

Xk(s) and which input Fj(s):

! X (s)
G (s) = k

F (s)
kj

• Solve, via previous result with all

components of F(s) zero, except Fj(s)

!
M.D. Bryant ME 344 notes 03/25/08 14

Example: differential equations

from transfer function:
X(s) 2s + 5
G(s) = =
F(s) s + 3s + 2s + 2s +1
4 3 2

 Cross multiply ratios:

! ( s + 3s + 2s + 2s +1) X(s) = (2s + 5) F(s)

4 3 2

s X + 3s X + 2s X + 2sX + X = 2sF + 5F
4 3 2

 Treat: s => d/dt

!
! d4 x d3x d2x dx df
4
+ 3 3
+ 2 2
+ 2 + x = 2 + 5 f (t)
dt dt dt dt dt

!
M.D. Bryant ME 344 notes 03/25/08 15

Block Diagrams
h(t) x(t)
G(s)
H(s) X(s)

• Another way to represent system

dynamics pictorially

• Weakness: lacks causality information

• Shows signal flow through system

• Transfer function G(s) inside block

• Output:

X(s) = G(s) H(s)

transfer function times input H(s)

• Can assemble blocks into system model:

M.D. Bryant ME 344 notes 03/25/08 16

• Cascaded Blocks

H(s) X1(s) X2(s) X(s)

G1(s) G2(s) G3(s)

• Output = input to next

• Models “stringing” of components
M.D. Bryant ME 344 notes 03/25/08 17

• Example: stereo system

speaker

CD player amplifier

 CD player→ amplifier→ speakers

X (s)
 CD player: G (s) = 1

H (s)
1

• Input: H(s) from CD laser

reader
• Output: CD voltage X1(s)
! X (s)
 Power amplifier: G (s) = 2

X (s)
2

• Input: CD output voltage X1(s)

• Output: amp voltage X2(s)
X(s)
 Speakers: G (s) =
! 3
X (s) 2

• Input: amp voltage X2(s)

• Output: sound, acoustic
pressure X(s)
!
M.D. Bryant ME 344 notes 03/25/08 18

H(s) X1(s) X2(s) X(s)

G1(s) G2(s) G3(s)

• Overall transfer function is product of

block transfer functions:

X(s) X(s) X (s) X (s)

G(s) = = = G (s)G (s)G (s)
2 1

H (s) X (s) X (s) H (s)

1 2 3

2 1

Note:
X (s) X (s)
G (s) = 1
, G (s) = 2
,
H (s) X (s)
1 2

X(s)
G (s) =
X (s)
3

! !

!
M.D. Bryant ME 344 notes 03/25/08 19

Example
H(s) 3 2s !1 X(s)

s +1 s + 4s +1
2

• Transfer function:

X(s) 3 2s #1
G(s) = = G (s)G (s) = "
H (s) s +1 s + 4s +1
1 2 2

3(2s "1) 6s " 3

G(s) = =
(s +1)(s + 4s +1) s + 5s + 5s +1
2 3 2

• Poles = roots of denominator (values of s

such that transfer function becomes
infinite)
p1 = -1, p2,p3 = -2 ± √3

• Zeros = roots of numerator (values of s

such that transfer function becomes 0)
z1 = 1/2
M.D. Bryant ME 344 notes 03/25/08 20

Summer (Summing Junction)

X1(s)

-
X2(s) + X(s)

+
X3(s)

 Output = sum of inputs

 Sign on input => sign in equation
 Output X(s) = - X1(s) + X2(s) + X3(s)
M.D. Bryant ME 344 notes 03/25/08 21

Example: Feedback Control System

+ E(s) X(s)
R(s) G(s)

H(s)

 Goal: Closed loop transfer function

X(s)
G (s) =
R(s)
cl

 Formulate:
E(s) = R(s) " H (s)X(s)
X(s) = G(s)E(s)
!  Eliminate E(s):
X(s) = G(s)E(s) = G(s)[ R(s) " H (s)X(s)]
!
!  Solve for closed loop X(s)/R(s):

! X(s) G(s)
G (s) = =
R(s) 1+ G(s)H (s)
cl

!
M.D. Bryant ME 344 notes 03/25/08 22

Example: Feedback Controller with

Disturbance
D(s)
E(s) +
+ + X(s)
R(s) Gc(s) Gp(s)
-

H(s)

 Closed loop transfer function, reference

input (temporarily set D(s) = 0 )

X(s) G (s)G (s)

G (s) = = c p

R(s) 1+ G (s)G (s)H (s)

cl

c p

 Transfer function, disturbance (set R(s) = 0)

! X(s) G (s)
G (s) = = p

D(s) 1+ G (s)G (s)H (s)

d

c p

!
M.D. Bryant ME 344 notes 03/25/08 23

 Linear system, with both, sum outputs:

X(s) = G (s)R(s) + G (s)D(s)

cl d

G (s)G (s)
X(s) = c
R(s)p

1+ G (s)G (s)H (s)

c p

! G (s)
+ D(s)
p

1+ G (s)G (s)H (s)

c p