Chapter 8: System Analysis Using Laplace Transforms

• Recall that we used transfer functions to represent the system dynamics in Chapters 5-7

– We derived the transfer functions using the D operator without formally using Laplace transform methods

• This chapter presents a brief overview of Laplace transform theory and its use in obtaining the

response of LTI dynamic systems

• Laplace transformation offers a systematic approach for solving an LTI differential equation by

transforming its variables in time t to an algebraic equation in the domain of the complex

Laplace variable s

– Any existing initial conditions are handled in a systematic manner, and the system’s time response is finally

obtained by determining the inverse Laplace transform

 8.2 Laplace Transformation
• The Laplace transform converts the function f(t) from the time domain to the domain of the
complex variable s, and is defined by

L{ f(t) }  F ( s)  0 f (t )e dt
 st

• The Laplace transform variable s    j is a complex variable where s and ω are the real
and imaginary parts, respectively

• The “inverse Laplace transform of F(s)” converts the complex function F(s) into the time
function f(t)

L-1{ F(s) } = f(t)

 Laplace Transforms of Common Time Functions: Example 8.1

• Compute the Laplace transform of the exponential function

f (t )  Ae  at for t > 0 where f(t) = 0 for t < 0

• Using the definition of the Laplace transform

 
L { Ae } 0 0
( a  s )t
 at  st
  at
Ae e dt  A e dt
( s  a )t t 
e Ae   Ae 0
A  
 ( s  a ) t 0  ( s  a )  ( s  a )

A
The Laplace transform is F ( s) 
sa
 Laplace Transform Tables

• In a similar fashion we can

compute Laplace transforms of

other common time functions

Example 8.1
(step, sinusoid, etc.) and

assemble a table of Laplace

transforms
 Laplace Transform Using MATLAB

• MATLAB’s Symbolic Math Toolbox can be used to compute the Laplace transform of a given
time function f(t)

• As an example, the following MATLAB commands will determine the Laplace transform of

f (t )  3 sin 2t

>> syms t % defines variable t (time) as a symbolic object

>> f = 3*sin(2*t) % defines function f(t) as a symbolic object
>> F = laplace(f) % computes the Laplace transform F(s)
>> pretty(F) % displays F in a format similar to typeset mathematics

6
Typing all commands we obtain -------- (See Laplace table,
2 entry #8)
s +4
 Poles and Zeros of Laplace Transforms

• A Laplace transform can be expressed using the form

a s  z1 s  z 2  s  z m 
F ( s) 
s  p1 s  p2 s  pn 
• The valuess   z1 , s   z2 , , s   zm that make F(s) = 0 are called the zeros of the transform
F(s)

• The values s   p1 , s   p2 , , s   pn that make F (s )   (or, the denominator of F(s)

equals zero) are called the poles of F(s)

1 No zeros
Simple example: F1 ( s )  1 pole at s = –3 (determines time fct)
s3

Note: corresponding time function is f 1 ( t )  e 3 t

 Laplace Transform Properties and Theorems

• Superposition (linearity):

• Differentiation:
 Laplace Transform Properties and Theorems (2)

• Integration:

• Multiplication f(t) by e–at in the time domain:

 Laplace Transform Properties and Theorems (3)

• Final-value theorem:

f ( )  lim f (t )  lim sF ( s )
t  s 0

• Therefore, we can use the Laplace transform F(s) to compute the steady-state value of the
corresponding time response f(t)

• The final-value theorem only holds when the function f(t) reaches a steady-state (constant) value as
time t  
– For example, the final-value theorem does NOT hold for the sinusoidal function
f (t )  sin 3t
 Laplace Transform Properties and Theorems (4)

• Initial-value theorem:

f (0 )  lim f (t )  lim sF ( s )

t 0  s 

• Therefore, we can use the Laplace transform F(s) to compute the initial value of the

time function, i.e., f(0+)

 Laplace Transforms Example 8.4

• Compute the Laplace transform of the time function

f (t )  0.2  2e 3t  5e 6t  8 sin 4t

• Take the Laplace transform of every term (use the Laplace transform table) and use the linearity
property (b):

0.2 2 5 32
F (s)     2
s s  3 s  6 s  16
 Laplace Transforms Example 8.5

• Compute the final value f () (if it exists) and the initial value f(0+) from the given Laplace
transform
7( s  3)
F ( s)  2
s  2 s  10

• The poles are the roots of s 2  2s  10  0 , or s  1  j 3

• Because the real part of the two poles is negative we can use the final-value theorem

7 s ( s  3) (0)(3)
f ()  lim sF ( s)  lim 2  0
s 0 s  0 s  2s  10 10

Hence the final value is zero

 Laplace Transforms Example 8.5 (2)

• Next, apply the initial-value theorem

7 s( s  3) 7s 2
f (0 )  lim sF ( s )  lim 2  lim 2  7
s  s  s  2 s  10 s  s

Hence the initial value is 7

 8.3 Inverse Laplace Transformation

• Our overall goal is to use Laplace transforms to obtain the response of a dynamic system that is
subjected to initial conditions and/or a known forcing or input function

• Systematic approach for solving the ODE and obtaining the dynamic response:

1. Take the Laplace transform of every term in the ODE and incorporate the initial conditions using
the differentiation properties

2. Using the result from step 1, solve for the Laplace transform of the dynamic variable, Y(s)

3. Obtain the system’s dynamic response by taking the inverse Laplace transform, y (t ) = L-1{ Y(s) }.
 Inverse Laplace Transform Example 8.6

• Given the following I/O equation (mathematical model)

2 y  10 y  12 y  u (t ) with y (0)  1 , y (0)  0.5

Determine the dynamic response y(t) if u(t) = 4U(t) (step input)

• Begin by taking the Laplace transform of each term on the left-hand side of the I/O equation
 Inverse Laplace Transform Example 8.6 (2)

• Next, take the Laplace transform of the right-hand side (input)

• Combining all Laplace transform results yields

2 4
2s Y ( s )  2 s  1  10sY ( s )  10  12Y ( s ) 
s

4 4  2s 2  9s
 2

2s  10s  12 Y ( s)   2 s  9 
s s
 Inverse Laplace Transform Example 8.6 (3)

• Solve for the Laplace transform Y(s)

 2s 2  9s  4  2s 2  9s  4
Y ( s)  2

2 s ( s  5s  6) 2s ( s  2)( s  3)
• Use partial-fraction expansion to express Y(s) as the sum of three fractions involving the three poles
at s = 0, –2, and –3

 2 s 2  9s  4 a1 a2 a3
Y (s)    
2s ( s  2)( s  3) s s  2 s  3

The corresponding constants are a1 = 1/3, a2 = –7/2, and a3 = 13/6

 Inverse Laplace Transform Example 8.6 (4)

• Taking the inverse Laplace transform of Y(s) yields the dynamic response

1 7 2t 13 3t
y (t )   e  e
3 2 6
• Check the initial value: y(0) = 1/3 – 7/2 + 13/6 = –1 (checks)

• The first time derivative is

y (t )  7e 2t  (13 / 2)e 3t

Therefore, y (0)  7  (13 / 2)  0.5 (checks)

 Partial-Fraction Expansion Method

• Partial fraction expansion is key to the previous example

• Partial-fraction expansion with distinct poles:

a1 a2 an
Y (s)   
s  p1 s  p2 s  pn

Evaluate each “residue” ai  ( s  pi )Y ( s ) s  p

i

MATLAB command residue computes the residues, poles,

and “direct” terms of the partial-fraction expansion

>> [a,p,k] = residue(numY,denY)

 Partial-Fraction Expansion Method Example 8.7

• Compute the inverse Laplace transform of

2s  5 2s  5
Y (s)    2 distinct poles: s = –2, –6
s 2  8s  12 ( s  2)( s  6)
2s  5 a a
• Partial-fraction expansion: Y ( s)   1  2
( s  2)( s  6) s  2 s  6

• Residues:
2s  5 1 2s  5 7
a1  ( s  2)Y ( s ) s 2    0.25 , a2  ( s  6)Y ( s ) s 6    1.75
s  6 s 2 4 s  2 s 6  4

• Take the inverse Laplace transform of both terms (see entry #6 in the Laplace transform table)

y (t )  0.25e 2t  1.75e 6t

 Partial-Fraction Expansion Method: Repeated Poles
• Partial-fraction expansion with repeated poles:
2s  8 2s  8
Example: Y (s)  3 2

s  7 s  16s  12 ( s  2)(s  2)(s  3)

Two repeated poles at s  2 and one distinct pole at s  3

• Partial-fraction expansion: a1 a2 a3
Y ( s)   
( s  2) 2 s  2 s  3

The inverse Laplace transform is

y (t )  a1te 2t  a2 e 2t  a3e 3t

 Residues for Repeated Poles
• Residues:
2s  8 4
a1  ( s  2) 2 Y ( s )   4
s  2 s  3 s 2 1

a2 
d
ds

( s  2) 2 Y ( s )  s  2

d  2s  8 

2

2s  8
ds  s  3  s 2 s  3 ( s  3) 2

2 4
  2
1 1
s  2

2s  8 2
a3  ( s  3)Y ( s ) s 3   2
( s  2) 2 s  3
1
4 2 2
• Partial-fraction expansion with residues: Y (s)   
( s  2) 2 s  2 s  3

 The inverse Laplace transform is

y (t )  4te 2t  2e 2t  2e 3t

Partial-Fraction Expansion Method: Complex Poles (Example 8.8)

• Compute the inverse Laplace transform of

2s  9
Y (s) 
s 2  6 s  25

Note the two poles are complex: s  3  j 4

• “Complete the square” on the denominator

2s  9
Y ( s) 
( s  3) 2  4 2

• This form matches entries #10 and 11 in the Laplace table

 Partial-Fraction Expansion Method: Example 8.8 (2)

• Re-write the Laplace transform so that it matches the Laplace transforms of exponentially damped
sine and cosine:
2s  9 2( s  3) (0.75)(4)
Y (s)   
( s  3) 2  4 2 ( s  3) 2  4 2 ( s  3) 2  4 2

• Take the inverse Laplace transform using entries #10 and 11 in the Laplace table

y (t )  2e 3t cos 4t  0.75e 3t sin 4t

Clearly the exp decay rate is the real part (a = –3) and the
frequency is the imaginary part (ω = 4) of the complex pole
 Inverse Laplace Transform Using MATLAB

• MATLAB’s Symbolic Math Toolbox can compute the inverse Laplace transform if we define the
Laplace transform as a symbolic object

• Re-work Example 8.8 using MATLAB:

>> syms s % defines Laplace variable s

>> Y = (2*s + 9)/(s^2 + 6*s + 25) % defines Laplace transform Y(s)
>> y = ilaplace(Y) % finds inverse Laplace transform
>> pretty(y) % displays y in math typeset

Typing all four line commands shown above yields

2 exp(-3 t) cos(4 t) + 3/4 exp(-3 t) sin(4 t)

which matches the solution on the previous slide

 8.4 Analysis of Dynamic Systems Using Laplace Transforms

• We can group the approaches for obtaining the dynamic response using Laplace methods into two

categories:

– Applying the Laplace transform to the system’s time-domain input-output (I/O) equation

– Using the system’s transfer function

• The first method yields the complete response because initial conditions can be incorporated

• The second method yields the response due to the input with zero initial conditions (due to the use

of the transfer function)

 Laplace Transform of the Input-Output Equation

• The basic steps for using the Laplace transform method to obtain the solution of the time-

domain I/O equation are

– Take the Laplace transform of every term in the I/O equation and include the initial conditions. This

step will convert an ODE into an algebraic equation in the Laplace variable s

– Solve the algebraic expression from step 1 for the Laplace transform of the output Y(s)

– Take the inverse Laplace transform of Y(s) to obtain the time response of the output y(t)

• See Examples 8.9 and 8.10 in the textbook

 Transfer-Function Analysis
• Let us now present the definition of the transfer function using Laplace methods:

• The transfer function G(s) is defined as the ratio of the Laplace transform of the output Y(s) to

the Laplace transform of the input U(s) with zero initial conditions

Quick example: y  8 y  10 y  4u

Take the Laplace transform of each term:

s 2Y ( s)  sy (0)  y (0)  8sY ( s )  y (0)   10Y ( s)  4sU ( s)  u (0) 

Set all initial conditions equal to zero, y (0)  y (0)  u (0)  0
s 2

 8s  10 Y ( s )  4 sU ( s)  Y (s) 4s
G(s)   2
Transfer function U ( s ) s  8s  10
 Transfer-Function Analysis

• The basic steps for using the transfer-function method to analytically obtain a system’s dynamic
response are
– Derive the system transfer function G(s) from the mathematical model (I/O equation)
– Multiply the transfer function G(s) by the Laplace transform of the given input function, U(s), to obtain
the Laplace transform of the output Y(s)
– Take the inverse Laplace transform of Y(s) to obtain the time response of the output y(t)

• See Examples 8.11-8.13 in the textbook

 Chapter 8: Summary
• We presented Laplace-transform methods for determining the response of dynamic systems
– Laplace-transform techniques can only be applied to systems represented by linear, time-invariant
(LTI) differential equations

• Determining the system response ultimately requires computing the inverse Laplace transform
which involves the partial-fraction expansion method
• When a system has zero initial conditions the Laplace transform of the output is

where G(s) is the system transfer function Yand

( s ) U(s)
 G (is the
s )U ( sLaplace
) transform of the system input