Chapter 2

Laplace Transform Review

Dr. Franjo Cecelja
1
Process and Information Systems Engineering Research Centre
Faculty of Engineering and Physical Sciences
University of Surrey
2.1 The Need for Laplace Transform
The use the Laplace Transform is strongly motivated by real engineering problems and especially
by problems in the area of Control. In control, we are usually interested how the output of a
systems depends on its input. Wast majority of control systems are dynamic systems, which
means that the output not only depends on the value of the input, but also on the change of
value of input.
Example 2.1: Lets take the control systems in Figure 2.1 represented in a block diagram
form and which has input u(t) and the output y(t) which depends on the derivative of the
input value.
Figure 2.1: Example of a simple dynamic control system
It is evident from Figure 2.1 that the output is calculated as derivative of the input as
y(t) =
d
dt
u(t) (2.1)
In Example 2.1, the output y(t) of the system entirely depends on the change in the value of the
input u(t), not actual value; if the input is constant, the output is 0. Such a system is called a
dynamic system.
1
These lecture notes have been compiled from the literature stated in the Bibliography Section
15
16 CHAPTER 2. LAPLACE TRANSFORM REVIEW
Example 2.2: Lets take an example of a moving car. Assume that the engine force u
moves the car with mass m in x direction at a speed v, as shown in Figure 2.2, and that the
rotation inertia of the wheels is negligible [2]. We are interested to see how the velocity of the
car v changes with the engine force u. The car is then modeled using free body diagram as
shown in Figure 2.3, with the force u generated by the engine being input (or manipulated)
variable and the car speed v is obviously output variable of the system.
Figure 2.2: A car in motion
Note that the friction b x acts in opposite direction from the force u as much as the inertia
of the car m x, that the velocity is the rst derivation of the position v = dx/dt = x and
that the acceleration is the rst derivative of the velocity, hence the second derivative of the
position a = dv/dt = v = d
2
x/dt
2
= x. Here, b [Ns/m] is the viscous friction constant
between the car and the surface of the road. As a consequence, the friction is entered as
negative force and using the second Newtons law the force balance on the car is:
u b x = m x (2.2)
or
x +
b
m
x =
u
m
(2.3)
Figure 2.3: Free body diagram for cruise control system
For the reason of the cruise control, it is the car speed v (= x) that is of interest here,
hence the equation (2.3) can be expressed in terms of velocity and written as:
v +
b
m
v =
u
m
(2.4)
This is obviously the rst order dierential equation, the solution of which, as you have
already experienced, is not easy to calculate. Yet, for the input in the form of step function
starting at t = 0
u(t) =
_
1000 t 0
0 t < 0
2.1. THE NEED FOR LAPLACE TRANSFORM 17
and zero initial conditions (v(0) = 0 [m/s]) the solution is
v(t) = 1000
_
1
b

1
b
e
t
b
m
_
(2.5)
As you recognise, in equation (2.5) the term
1
b
is the particular integral and the term
1
b
e
t
b
m
is the homogeneous solution. Numerical solution
2
for two cars weighting 1000 [kg]
and 2000 [kg], with the same friction of b = 26 [Ns/m] (this data is obtained experimentally
as shown in ref. [2], chapter 2) is shown in Figure 2.4. At the beginning, the car is stationary,
and then at the moment of observation (t = 0 [s]) the engine provides step force of 1000 [N].
The results clearly show that the velocity changes in an exponential manner and also that
it takes longer for a heavier car to achieve stationary speed than for a lighter one, both for
the same force of the engine u = 1000 [N].
Figure 2.4: Computer solution of dierential equation
The things become even more complicated if the order of the dierential equation is higher
or if we want to combine more systems together. Yes, there is a signicant number of computer
tools to help you with this, but then we loose inside view into the system, the part which is
so important in designing an appropriate control system. To compromise between the ease of
nding the solution and still heaving good understanding of the system, numerous mathematical
methods have been developed; the Laplace transform is probably the most commonly used.
So, the Laplace transform is one of the mathematical tools used to solve linear ordinary
dierential equations, and has the following features:
1. The homogeneous equation and the particular integral of the solution of dierential equa-
tions are obtained in one operation;
2. The Laplace transform converts the dierential equation into an algebraic equation in a
new domain, the s-domain. It is then possible to manipulate the algebraic equation by
2
MatLab example Ex1.
18 CHAPTER 2. LAPLACE TRANSFORM REVIEW
simple algebraic rules to obtain the solution in time or simply t-domain. The nal solution
is then obtained by taking the inverse Laplace transform.
Before we go to the denition of Laplace transform, it is important to note that the author of
these notes assumes that you are fully familiar with the complex numbers and complex variables.
If not, you will have to make an eort to familiarise yourself with that mathematical area fully.
2.1.1 The Concept of the Laplace Transform
Normal form of dierential equations is in so called input-output form as demonstrated by
eq. (2.4). Lets break the system input u(t) into simpler inputs consisting of sum of impulses
and then determine the system response, the output y(t), to each of these impulses. By using
superposition principle, it is possible then to obtain the system response to the actual input.
For a typical system input as shown in Figure 2.5 a), we approximate this input by the
square pulse train so that
u(t)

k=
u(t, t
k
, t) =

k=
u(t
k
)p(t t
k
, t)t (2.6)
where p(t t
k
, t) is the pulse of height 1/t and width t and centered at t = t
k
. This pulse
is called unit pulse since its area is unity.
Lets dene the response of initially unexcited linear and time invariant system to a unit
pulse p(t , t) as h(t, , t) where is the time the centre of the unit pulse occurs and t is
its width. Now, if we let the width of the pulse approach zero, than the unit impulse is dened
as
(t ) = lim
t0
p(t , t) (2.7)
which is called the Dirac delta function , or simply function. The corresponding system
Figure 2.5: Decomposition of input signal u(t)
output for this unit impulse is then
h(t ) = lim
t0
h(t, , t) (2.8)
2.2. DEFINITION OF THE LAPLACE TRANSFORM 19
The actual system input time is not just the unit pulse but the unit pulse multiplied by u(t
k
)t.
The response due to the complete pulse train, by applying the superposition principle and for
t 0, is then the integral expressed as:
y(t) =
_

h(t, )u()d (2.9)

The integral given by eq. (2.9) is called the convolution integral. Note that for t 0, the
impulse function (t) has innitely short duration and innitely high amplitude to maintain the
unity area.
The interpretation of the integral 2.9 is that the response of the system to the string
of impulses is just the sum of responses to individual impulses. In eq. (2.9), the function
h(t, ) = h(t ) represent a stream of delayed impulses (delayed by ).
One of the uses of the convolution integral (2.9) is to nd the response of the system as
a sum of individual responses to individual impulse functions. If we take a real system which
can not respond to the input before it occurs (there is no prediction built in the system), then
the convolution integral (2.9) can be written in so called restricted form
y(t) =
_

0
h(t, )u()d (2.10)
which assumes that the system excitation starts at t = 0 and which serves as an idea for the
Laplace transform which has more practical use.
2.2 Denition of the Laplace Transform
Given the real function f(t) that satises the condition
_

f(t)e
t

dt < (condition of
continuity) for some nite and real , the Laplace transform of a time dependent function f(t)
is dened as
F(s) = L{f(t)} =
_

0
f(t)e
st
dt (2.11)
The variable s is referred to as Laplace operator, and it is a complex variable
s = +j (2.12)
where is the real component, is the imaginary component and j =

1. In control terms,
(= 2f) is the frequency and that is the reason why it is commonly said that the Laplace
transform converts the time domain function into the frequency domain. Here, f (= 1/T)
denotes the actual frequency which is measured in Hertz [Hz] (1 [Hz] is a value of one period
per second) as shown in Figure 2.6. Note in eq. (2.11) that the integration goes between 0 and
. That means that all information in f(t) that date before t = 0 is ignored. As explained for
the convolution integral, this limitation does not impose any limitation for real systems including
control systems. Control systems are always considered from t = 0. Even if the excitation starts
at t = 0, the function response can never starts before t = 0, so the equation (2.11) is always
fully valid. Also note that in equation (2.11) we use low case letters for time domain function
(f(t)) and capital letters for functions in s-domain (F(s)), which is an established convention
in control (and mathematics).
20 CHAPTER 2. LAPLACE TRANSFORM REVIEW
Figure 2.6: Denition of frequency
Example 2.3: Find Laplace transform for the unit step function dened as:
f(t) =
_
1 t 0
0 t < 0
and which is shown in Figure 2.7.
Figure 2.7: Unity step function
The Laplace transform is the obtained as:
F(s) = L{f(t)} =
_

0
e
st
dt =
1
s
e
st

0
= 0 +
1
s
=
1
s
Example 2.4: Find the Laplace transform for a ramp function of the form
f(t) =
_
bt t 0
0 t < 0
where b denes the slope and which is shown in Figure 2.8.
The Laplace transform is obtained by applying the partial integration
3
of the form
_
u
dv
dx
dx = uv
_
v
dv
dx
where both functions u and v are functions of t, we have
F(s) =
_

0
bte
st
dt =
_

bte
st
s

be
st
s
2
_

0
=
b
s
2
3
If we have a function y = uv, then by dierentiating both sides we get
dy
dx
= u
dv
dx
+v
du
dx
. Now, by integrating
this equation we get

dy
dx
dx =

u
dv
dx
dx +

v
du
dx
dx. The left hand side just returns to y, which writes uv =

u
dv
dx
dx +

v
du
dx
dx. By rearranging we get

u
dv
dx
dx = uv

v
du
dx
dx
2.2. DEFINITION OF THE LAPLACE TRANSFORM 21
Figure 2.8: Ramp function
Example 2.5: Find the Laplace transform of an exponential function of the form
f(t) =
_
e
at
t 0
0 t < 0
The Laplace transform is
F(s) =
_

0
e
at
e
st
dt =
_

0
e
(s+a)t
dt = 0 +
1
s +a
=
1
s +a
In literature you can nd derived Laplace transform for numerous functions, especially
those most commonly in use today. Some of those are summarised in Table 2.1, at least those
which will be used in the course of this module
4
.
f(t) F(s) = L{f(t)} f(t) F(s) = L{f(t)}
1
1
s
1 cos(at)
a
2
s(s
2
+a
2
)
t
1
s
2
at sin(at)
a
3
s
2
(s
2
+a
2
)
t
n n!
s
n+1
tcos(at)
a
2
a
2
(s
2
+a
2
)
2
e
at 1
sa
tsin(at)
2as
s(s
2
+a
2
)
te
at 1
(sa)
2
e
at
cos(bt)
s+a
(s+a)
2
+b
2
1
(n1)!
t
n1
e
at 1
(sa)
n e
at
sin(bt)
b
(s+a)
2
+b
2
e
at
e
bt
ab
(a = b)
1
(sa)(sb)
u(t)
1
s
ae
at
be
bt
ab
(a = b)
s
(sa)(sb)
u(t a)
e
as
s
cos(at)
s
s
2
+a
2
(t) 1
sin(at)
a
s
2
+a
2
(t a) e
as
cosh(at)
s
s
2
a
2
sinh(at)
a
s
2
a
2
Table 2.1: Laplace transform of the most commonly used functions
4
Data prepared by Dr. N. Rockli
22 CHAPTER 2. LAPLACE TRANSFORM REVIEW
2.3 Properties of Laplace Transforms
2.3.1 Superposition (linearity)
The Laplace transform is linear: this is one of the most important properties. Given two time-
domain functions f
1
(t) and f
2
(t) and two real coecients and , the linearity property of the
Laplace transform can be written as
L{f
1
(t) +f
2
(t)} =
_

0
(f
1
(t) +f
2
(t)) e
st
dt =
_

0
f
1
(t)e
st
dt +
_

0
f
2
(t)e
st
dt
(2.13)
Hence
L{f
1
(t) +f
2
(t)} = F
1
(s) +F
2
(s) (2.14)
The scaling property follows naturally from equation (2.14) as a special case, that is
L{f(t)} = F(s) (2.15)
2.3.2 Transforms of Derivatives
The Laplace transform of derivative of a time domain function f(t)
L{df(t)/dt} =
_

0
df(t)
dt
e
st
dt
can be shown in the following way: applying the integration by parts on function
5
f(t), with
u = f(t) and dv/dt = e
st
we can show that from the denition of the Laplace transform (2.11):
F(s) =
_

0
f(t)e
st
dt = f(t)
e
st
s

_

0
df(t)
dt
e
st
s
dt (2.16)
Hence
F(s) =
f(0)
s
+
1
s
L
_
df(t)
dt
_
(2.17)
which then gives
L
_
df(t)
dt
_
= sF(s) f(0) (2.18)
Here, f(0) is the initial condition of the function f(t), which means the value it takes at t = 0.
Similarly, we obtain a general formula for transform of derivatives of second order as
L
_
d
2
f(t)
dt
2
_
= s
2
F(s) sf(0)
df(0)
dt
(2.19)
Or, in general for derivative of any order the Laplace transform is
L
_
d
n
f(t)
dt
n
_
= s
n
F(s) s
n1
f(0) s
n2
df(0)
dt

d
n1
f(0)
dt
n1
(2.20)
5
If we have a function y = uv, then by dierentiating both sides we get
dy
dx
= u
dv
dx
+v
du
dx
. Now, by integrating
this equation we get

dy
dx
dx =

u
dv
dx
dx +

v
du
dx
dx. The left hand side just returns to y, which writes uv =

u
dv
dx
dx +

v
du
dx
dx. By rearranging we get

u
dv
dx
dx = uv

v
du
dx
dx
2.3. PROPERTIES OF LAPLACE TRANSFORMS 23
Example 2.6: Find the Laplace transform of the dierential equation
d
2
y(t)
dt
2
+ 3
dy(t)
dt
+ 2y(t) = 5u(t)
with u(t) as input function and with initial conditions y(0) = 1 and y(0) =
dy(t)
dt

t=0
= 2.
Using the derivative property of the Laplace transform, as per eq. (2.20), we get
s
2
Y (s) sy(0) y(0) + 3sY (s) 3y(0) + 2Y (s) = 5U(s)
or
Y (s)
_
s
2
+ 3s + 2
_
+s + 1 = 5U(s)
2.3.3 Transform of the Integral of a Function
For the Laplace transform of an integral we have
F
1
(s) = L
__
t
0
f()d
_
=
_

0
__
t
0
f()d
_
e
st
dt (2.21)
and employing the integral by parts where u =
_
t
0
f()d and dv = e
st
dt we get
F
1
(s) =
e
st
s
_
t
0
f()d

_

0

e
st
s
f(t)dt =
F(s)
s
(2.22)
Example 2.7: Calculate the Laplace transform of the integral of a time domain function
f(t) =
_
tdt assuming the initial condition f(0) = 0.
From eq. (2.22) the Laplace transform is:
F(s) = L
__
tdt
_
=
L{t}
s
=
1
s s
2
=
1
s
3
Alternatively and to prove validity of the Laplace transform of integration, we derive the
integral of the given function as:
f(t) =
_
tdt =
t
2
2
+f(0) =
t
2
2
then the Laplace transform from the Table 2.1 is:
F(s) = L{f(t)} = L
_
t
2
2
_
=
2
2s
3
=
1
s
3
which proves the validity.
2.3.4 Delayed (in time) functions: the second shift theorem
Suppose that the function f(t) is delayed by > 0 units of time as shown in Figure 2.9. Then,
the Laplace transform is F(s) =
_

0
f(t )e
st
dt. If we dene a = t , then da = dt since
is a constant, and f(t) = 0 for t < 0. Thus
F(s) =
_

0
f(a)e
s(a+)
da =
_

0
f(a)e
sa
e
s
da (2.23)
24 CHAPTER 2. LAPLACE TRANSFORM REVIEW
Figure 2.9: Delayed unity step function
Since e
s
is independent of time, it can be taken out of the integral, so
F
1
(s) = e
s
_

0
f(a)e
sa
da = e
s
F(s) (2.24)
So, the delay in the time-domain by units of time of a time domain function f(t) is equal to
multiplication by exponent function e
s
in s-domain.
Example 2.8: Determine the Laplace transform of a ramp function f(t) = t delayed by 2
sec, as shown in Figure 2.11.
Figure 2.10: Delayed ramp function
The delayed function in time domain is f(t 2) = t 2. Because F(s) = L{f(t)} =
1
s
2
,
the Laplace transform of the delayed function according to the eq. (2.24) is
L{f(t 2)} =
e
2s
s
2
Example 2.9: Represent the function in Figure 2.11 in analytical form and nd its Laplace
transform.
The analytical format is
f(t) = u(t 1) u(t 3)
where u(t) is the unity step function and u(t a) is the unity step function u(t) shifted by
a. Therefore, the Laplace transform is, using the linearity theorem (2.14) and delay theorem
2.3. PROPERTIES OF LAPLACE TRANSFORMS 25
Figure 2.11: A pulse function
(2.24):
F(s) = L{f(t)} = L{u(t 1)} L{u(t 3)} =
1
s
_
e
s
e
3s
_
2.3.5 The rst shift theorem: frequency shift
Multiplication (modulation) of f(t) by a time dependent exponential e
at
expression in the
time-domain corresponds to a shift in frequency by a in the s-domain:
F
1
(s) =
_

0
f(t)e
at
e
st
dt =
_

0
f(t)e
(a+s)t
dt = F(s +a) (2.25)
Example 2.10: Calculate the Laplace transform of the function f(t) = t
3
e
4t
.
It is obvious that the function f(t) can be decomposed to a function f
1
(t) = t
3
shifted
in frequency by -4 (e
4t
). Hence, the Laplace transform per Table 2.1 is
F
1
(s) =
6
s
4
Now, from eq. (2.25) the Laplace transform for the function f(t) is
F(s) = F
1
(s 4) =
6
(s 4)
4
2.3.6 Multiplication by time: dierentiation of the Laplace transform
Multiplication by time in the time-domain corresponds to dierentiation in the frequency do-
main:
d
ds
F(s) =
d
ds
_

0
f(t)e
st
dt =
_

0
te
st
f(t)dt =
_

0
t(f(t))e
st
dt = L{tf(t)} (2.26)
Then
L{tf(t)} =
d
ds
F(s) (2.27)
Example 2.11: Find the Laplace transform of a time-domain function f(t) = te
3t
.
Taking that f
1
(t) = e
3t
, the the Laplace transform of it is
L{f
1
(t)} = F
1
(s) =
1
s + 3
26 CHAPTER 2. LAPLACE TRANSFORM REVIEW
The derivative of F
1
(s) is
d
ds
F
1
(s) =
d
ds
_
1
s + 3
_
=
0 1
(s + 3)
2
=
1
(s + 3)
2
So, the solution is
L{f(t)} = L
_
te
3t
_
=
d
ds
F
1
(s) =
1
(s + 3)
2
2.3.7 Division by time: integration of the Laplace transform
The Laplace transform of a time function f(t) divided by time t is equivalent to the integration
is s-domain:
L
_
f(t)
t
_
=
_

s
F( s)d s (2.28)
To prove this, lets start from integral of the Laplace transform
_

s
F( s)d s. This yields
_

s
F( s)d s =
_

0
__

0
e
st
f(t)dt
_
d s =
_

0
__

0
e
st
d s
_
f(t)dt =
_

0
1
t
e
st
f(t)dt =
= L
_
f(t)
t
_
(2.29)
2.3.8 The Laplace transform of convolution
For the convolution of two functions of time domain f
1
(t) and f
2
(t)
f
1
(t) f
2
(t) =
_
t
0
f
1
()f
2
(t )d (2.30)
it can be shown that by permutation of variables, the convolution is a symmetric operation, so
that f
1
(t) f
2
(t) = f
2
(t) f
1
(t), hence
_
t
0
f
1
()f
2
(t )d =
_
t
0
f
2
()f
1
(t )d (2.31)
Then, the Laplace transform is
L{f
1
(t) f
2
(t)} = L
__
t
0
f
1
()f
2
(t )d
_
=
_

t=0
_
t
=0
s
st
f
1
()f
2
(t )ddt (2.32)
Substituting = t and d = dt and using the valid extension of the upper bounds of
integration to yields
L{f
1
(t) f
2
(t)} =
_

=
_

=0
e
s(+)
f
1
()f
2
()dd (2.33)
2.4. INVERSE LAPLACE TRANSFORM 27
Because both functions f
1
(t) and f
2
(t) have zero values for t < 0, it follows with respect to lower
limit of integration that
L{f
1
(t) f
2
(t)} =
_

t=0
e
s
f
1
()d
_

0
s
s
f
2
()d = F
1
(s)F
2
(s) (2.34)
In consequence, the Laplace transform of the convolution of two functions is the product
of the Laplace transforms of individual functions:
L{f
1
(t) f
2
(t)} = F
1
(s)F
2
(s) (2.35)
2.3.9 The Laplace transform of periodic functions
To nd the Laplace transform of basic periodic functions sin(t) and cos(t), we can use their
properties of periodicity. Lets take for example the function f(t) = sin(t). For this function
we have

f(t) = cos(t),

f(t) =
2
sin(t) where initial conditions are f(0) = 0 and

f(0) = .
So, we can now write

2
L{f(t)} = L
_

f(t)
_
= s
2
L{f(t)} sf(0)

f(0)
By rearranging this equation we get
L{f(t)}
_
s
2
+
2
_
=
which nally gives the solution
L{f(t)} = L{sin(t)} =

s
2
+
2
In the similar way we can determine the Laplace transform for all other periodic functions.
The summary of the Laplace transform properties
6
is given in Table 2.2.
2.4 Inverse Laplace Transform
The inverse Laplace transform refers to the process of nding the time-domain function f(t) for
the corresponding Laplace transform (s-domain function) F(s). The inverse Laplace transform
of a function F(s), denoted as L
1
{F(s)}, is given by:
L
1
{F(s)} = f(t) =
_
c+j
cj
F(s)e
st
ds (2.36)
where c is a real constant that is greater than the real parts of all singularities (poles) of F(s).
Now, we will not be solving the integral given by the eq. (2.36) as it is a very dicult job. Instead,
several easier methods are available for nding the inverse Laplace transforms. The simplest of
all, of course, is to use tables (Table 2.1) of Laplace transforms to nd the corresponding function
f(t). However, given s-domain function F(s) is not always in a table form, so several methods
are available to prepare F(s) for direct table reading. One of the most useful, especially with
rational functions, and hence the most commonly used is the Partial-fraction Expansion the
use of which is explained in the follow-on section.
6
The data is prepared by Dr. N. Rockli
28 CHAPTER 2. LAPLACE TRANSFORM REVIEW
1. Denition Laplace transform of f(t) is F(s) = L{f(t)} =
_

0
f(t)e
st
dt
2. Linearity L{af(t)bg(t)} = aL{f(t)} +bL{g(t)}
3. Derivative 1
st
: L{f

(t)} = sL{f(t)} f(0)

2
nd
: L{f

(t)} = s
2
L{f(t)} sf(0) f

(0)
n
th
: L
_
f
(n)
(t)
_
= s
n
L{f(t)} s
(n1)
f(0) s
(n2)
f

(0)
f
(n1)
(0)
4. Integral L
_
_
t
0
f()d
_
=
1
s
L{f(t)}
5. 1
st
shift If L{f(t)} = F(s) then L
_
e
at
f(t)
_
= F(s a)
alternatively L
1
{F(s a)} = e
at
f(t)
6. 2
nd
shift If L{f(t)} = F(s) then L{u(t a)f(t a)} = e
as
F(s)
alternatively L
1
{e
as
F(s)} = u(t a)f(t a)
where u(t a) denotes the shift step function
7. Dierentiation If L{f(t)} = F(s) then L{tf(t)} = F

(s)
8. Integration If L{f(t)} = F(s) then L
_
f(t)
t
_
=
_

s
F( s)d s
9. Convolution For f and g, con.is (f g) (t) =
_
t
0
f()g(t )d
If L{f(t)} = F(s) and L{g(t)} = G(s) then the Laplace
transform is L
_
_
t
0
f()g(t )d
_
= F(s)G(s)
10. Periodic func. The Laplace transform of a periodic function f with
period p is L{f(t)} =
1
1e
ps
_
p
0
f(t)e
st
dt
Table 2.2: Properties of the Laplace transform
2.4. INVERSE LAPLACE TRANSFORM 29
2.4.1 Partial-Fraction Expansion
When the Laplace transform solution of a dierential equation is a rational function in s (from
Table 2.1 it is apparent that most of functions of the Laplace transform are rational functions),
it can be written in a general form as
G(s) =
Q(s)
P(s)
=
b
m
s
m
+b
m1
s
m1
+ +b
0
s
n
+a
n1
s
n1
+ +a
0
(2.37)
where Q(s) and P(s) are polynomials of s. It is assumed that the order n of P(s) is greater
than the order m of Q(s), n > m, the assumption which holds for most of practical engineering
problems. The polynomial P(s) (= s
n
+a
n1
s
n1
+ +a
1
s +a
0
), where a
n1
, . . . , a
1
, a
0
are
real coecients, has solutions obtained by setting
P(s) = s
n
+a
n1
s
n1
+ +a
1
s +a
0
= 0 (2.38)
The method of partial-fraction is based on manipulating solutions of the polynomial P(s), which
are in control normally called the poles: solutions of the polynomial P(s) are values for which
P(s) = 0 and hence G(s) = . The ultimate outcome of applying the partial-fraction expansion
is to convert the function G(s) into the form which is available from Table 2.1. We will consider
here the use of partial-fraction expansion for cases of simple poles, multiple-order poles and
complex-conjugate poles of G(s).
Simple poles
If all poles of G(s) are simple and real, the equation (2.37) can be written as
G(s) =
Q(s)
P(s)
=
Q(s)
(s +s
1
) (s +s
2
) (s +s
n
)
(2.39)
where s
1
= s
2
= = s
n
. By applying the partial-fraction expansion, the eq. (2.39) can be
further expanded into the form
G(s) =
Q(s)
(s +s
1
) (s +s
2
) (s +s
n
)
=
K
s1
s +s
1
+
K
s2
s +s
2
+ +
K
sn
s +s
n
(2.40)
The coecients K
si
(i = 1, 2, , n) are determined by multiplying both sides of the equation
(2.39) by the factor (s +s
i
) and then setting s equal to s
i
such as:
K
si
=
_
(s +s
i
)
Q(s)
P(s)
_

s=s
i
=
Q(s
i
)
(s
i
+s
1
) (s
i
+s
2
) (s
i
+s
n
)
(2.41)
For K
s1
, for example, we have
K
s1
=
_
(s +s
1
)
Q(s)
P(s)
_

s=s
1
=
Q(s
1
)
(s
1
+s
2
) (s
1
+s
3
) (s
1
+s
n
)
(2.42)
30 CHAPTER 2. LAPLACE TRANSFORM REVIEW
Example 2.12: Find the partial-fraction expansion of the function G(s) =
5s+3
(s+1)(s+2)(s+3)
.
Given function G(s) can be written in the partial form as
G(s) =
K
1
s + 1
+
K
2
s + 2
+
K
3
s + 3
(2.43)
the coecients are:
K
1
= [(s + 1) G(s)]|
s=1
=
5 + 3
(1 + 2) (1 + 3)
= 1
K
2
= [(s + 2) G(s)]|
s=2
=
10 + 3
(2 + 1) (2 + 3)
= 7
K
3
= [(s + 3) G(s)]|
s=3
=
15 + 3
(3 + 1) (3 + 2)
= 6 (2.44)
So, from equations (2.43) and (2.44) the solution is
G(s) =
1
s + 1
+
7
s + 2

6
s + 3
(2.45)
The function (2.45) is now directly readable from the Laplace transform table (Table 2.1) as:
f(t) = e
t
+ 7e
2t
6e
3t
(2.46)
To test the validity of derived partial-fraction expansion, we can simply sum fractions in eq.
(2.45) and the result should be the same as the original function.
Simple and Repeated Poles
Similarly to the simple poles procedure (eq. (2.41)), we can also use partial fraction expansion
for the system G(s) which has repeated poles [4]. The procedure will be explained on the
following example.
Example 2.11: Find the partial-fraction expansion of the s-domain function
G(s) =
2
(s + 1) (s + 2)
2
(2.47)
where evidently the poles of (s + 2)
2
(the solutions of the denominator) are repeated since
the factor is raised to an integer power higher than 1. In this case the pole s = 2 is a
multiple pole of multiplicity 2.
We can write the partial-fraction expansion as a sum of terms, where each factor of the
denominator forms the denominator of each terms. In addition, each multiple pole generates
additional terms consisting of denominator factor of reduced multiplicity. So, for the rational
function given by eq. (2.47), we have
G(s) =
2
(s + 1) (s + 2)
2
=
K
1
(s + 1)
+
K
2
(s + 2)
2
+
K
3
(s + 2)
(2.48)
Then K
1
= 2, which can be found as described in the previous Section, whereas the coecient
K
2
can be isolated by multiplying equation (2.48) by (s + 2)
2
, yielding
2
s + 1
= (s + 2)
K
1
(s + 1)
+K
2
+ (s + 2) K
3
(2.49)
2.4. INVERSE LAPLACE TRANSFORM 31
Letting s = 2, we nd K
2
= 2. To nd K
3
, we see that if we dierentiate eq. (2.49) with
respect to s, we get
2
(s + 1)
2
=
(s + 2)
2
(s + 1)
2
K
1
+K
3
(2.50)
K
3
is isolated and we nd it if we take s = 2. Hence, K
3
= 2. Consequently, the
partial-fraction expansion of the rational function described by eq. (2.47) has the form
G(s) =
2
s + 1

2
(s + 2)
2

2
s + 2
(2.51)
Shortly, the above procedure of partial fraction expansion of function (2.47) could be
shortened as follows:
G(s) =
2
(s + 1) (s + 2)
2
=
K
1
(s + 1)
+
K
2
(s + 2)
2
+
K
3
(s + 2)
which gives
K
1
= [(s + 1) G(s)]|
s=1
=
2
(1 + 2)
2
= 2
K
2
=
_
(s + 2)
2
G(s)
_

s=2
=
2
(2 + 1)
= 2
K
3
=
_
d
ds
(s + 2)
2
G(s)
_

s=2
=
2
(s + 1)

s=2
=
2
(2 + 1)
= 2
Complex-Conjugate Poles
If the rational function (2.39) contains a pair of complex-conjugate poles s
1
= + j and
s
2
= j (with a system which has a real-life implementation, complex poles always appear
as a complex-conjugate pairs), then the corresponding coecients of these poles can be found
by using the equation (2.41) as:
K
+j
= (s + j) G(s)|
s=+j
K
j
= (s + +j) G(s)|
s=j
(2.52)
Example 2.13: Write the function G(s) =
1
s
2
+2s+5
in the fractional form.
The given function can be written in the following format:
G(s) =
1
(s + 1 + 2j) (s + 1 2j)
=
K
1
s + 1 + 2j
+
K
2
s + 1 2j
(2.53)
where coecients K
1,2
, according to the equation (2.52), are:
K
1
= (s + 1 + 2j) G(s)|
s=12j
=
1
1 2j + 1 2j
=
1
4j
K
2
= (s + 1 2j) G(s)|
s=1+2j
=
1
1 + 2j + 1 + 2j
=
1
4j
(2.54)
Now, the original function G(s) in a fractional form is:
G(s) =

1
4j
s + 1 + 2j
+
1
4j
s + 1 2j
(2.55)
32 CHAPTER 2. LAPLACE TRANSFORM REVIEW
The imaginary form of equation (2.40) does not make any dierence. Taking from the
Laplace transform table (Table 2.1), the one for F(s) =
b
sa
in s-domain with the time
domain response is f(t) = be
at
, the time-domain response for equation (2.55) is:
f(t) =
1
4j
e
(1+2j)t
+
1
4j
e
(12j)t
=
1
4j
_
e
(12j)t
e
(1+2j)t
_
(2.56)
Using the identity e
(a+jb)t
= e
at
(cos(bt) +jsin(bt)), the equation (2.56) becomes
f(t) =
e
t
4j
(cos(2t) +jsin(2t) cos(2t) sin(2t)) =
=
e
t
4j
(cos(2t) +jsin(2t) cos(2t) +jsin(2t)) =
=
e
t
2
sin(2t) (2.57)
2.4.2 Inverse Laplace Transform Process
Once the partial-fraction form of the function is obtained, it is very likely that corresponding
time-domain function can be found directly from the Laplace transform table (Table 2.1), as
shown by eq. (2.46) and (2.57).
2.4.3 Solving Linear, Time-Invariant Dierential Equations
Linear and time-invariant dierential equations can be solved by using the Laplace transform
method with the aid of the theorems of Laplace transform, the partial-fraction expansion and
the Table 2.1 of Laplace transforms. The procedure is as follows:
1. Transform the dierential equation to the s-domain using Laplace transform;
2. Manipulate the transformed algebraic equation and solve for output variable;
3. If the obtained relationship is not directly readable from the table, perform the partial-
fraction expansion to the transformed algebraic equation;
4. Obtain inverse Laplace transform from the Laplace transform table.
Example 2.14: Lets take the dierential equation
d
2
y(t)
dt
2
+ 3
dy(t)
dt
+ 2y(t) = 5u(t) (2.58)
where u(t) is the unit step function, and initial conditions are y(0) = 1 and y(0) =
dy(t)
dt

t=0
= 2. Now, the rst step is to nd the Laplace transform of equation (2.58).
Using the derivative property of the Laplace transform, as given in the Section 2.3.2, and
the Laplace transform of the unity step input function u(t) =
_
1 t0
0 t<0
as U(s) =
1
s
, we get
s
2
Y (s) sy(0) s y(0) + 3sY (s) 3y(0) + 2Y (s) =
5
s
(2.59)
2.5. SKILL-ASSESSMENT EXERCISE 33
which, solving it for the output Y (s) gives
Y (s)
_
s
2
+ 3s + 2
_
sy(0) y(0) 3y(0) =
5
s
(2.60)
Y (s)
_
s
2
+ 3s + 2
_
=
5
s
s + 2 3 =
s
2
s + 5
s
(2.61)
Thus
Y (s) =
s
2
s + 5
s (s
2
+ 3s + 2)
=
s
2
s + 5
s (s + 1) (s + 2)
=
K
1
s
+
K
2
s + 1
+
K
3
s + 2
(2.62)
Now, applying partial-fraction expansion we get K
1
= 5/2, K
2
= 5 and K
3
= 3/2,
K
1
= sY (s)|
s=0
=
5
(0 + 1)(0 + 2)
=
5
2
K
2
= (s + 1)Y (s)|
s=0
=
1 + 1 + 5
1(1)
= 5
K
3
= (s + 2)Y (s)|
s=0
=
4 + 2 + 5
2(1))
=
3
2
hence the equation (2.62) becomes
Y (s) =
5
2s

5
s + 1
+
3
2(s + 2)
(2.63)
Applying the inverse Laplace transform, the equation (2.63) takes the time-domain form as:
y(t) =
5
2
5e
t
+
3
2
e
2t
(2.64)
The rst term (
5
2
) in (2.64) is the steady-state solution or the particular integral; the last
two terms (5e
t
+
3
2
e
2t
) are transient, or homogeneous solution.
2.5 Skill-Assessment Exercise
2.5.1 Review Questions
1. Give the Laplace transform denition and explain its importance for control theory. What
is the inverse Laplace transform?;
2. Explain how the fact that the Laplace transform operates in the range between t = 0 and
t = aects the process of solving control problems;
3. Several properties of the Laplace transform are very useful for solving the control problem.
Show that for a time-domain function f(t) delayed by seconds such that f
1
(t ) the
Laplace transform is F
1
(s) = e
s
F(s);
4. Show that multiplication of a time domain function f(t) by e
at
corresponds to frequency
shift in frequency domain as F(s +a);
5. Show that multiplication by time t in time domain (tf(t)) corresponds to dierentiation
in frequency domain
d
ds
F(s).
34 CHAPTER 2. LAPLACE TRANSFORM REVIEW
2.5.2 Solving Problems
Task 2.1:
Derive the Laplace transform for time domain function f(t) = cos(t).
Hint: use the derivative property of Laplace transform of the form
L
_
d
n
f(t)
dt
n
_
= s
n
F(s) s
n1
f(0) s
n2
df(0)
dt

d
n1
f(0)
dt
n1
Task 2.2:
Find inverse Laplace transform, that is function g(t), for the following s-domain function:
G(s) =
s (s + + 5) + 3
s
3
+ 5s
2
+ 4s
where and are real coecients.
Task 2.3:
For the system in Figure 2.12 described by the dierential equation
y(t) + 5 y(t) + 4y(t) = u(t)
nd the time domain solution (the system output y(t)) for the initial conditions y(0) = 0,
y(0) = 0, and the input function u(t) = 2e
2t
. Sketch the response y(t) graphically.
Figure 2.12: A time-domain system
Task 2.4:
Given the dynamic model of a system as:
d
2
y(t)
dt
2
+ 2
dy(t)
dt
+ 2y(t) = u(t)
where y(t) is the output of the system and u(t) is the input of the system.
Find the time-domain response for zero initial conditions and impulse input u(t) = (t) and
sketch it in the time diagram. Derive the transfer function of the system in the form G(s) =
Y (s)
U(s)
.
2.5. SKILL-ASSESSMENT EXERCISE 35
Task 2.5:
For the system described by the transfer function
G(s) =
Y (s)
U(s)
=
1
s (s + 2)
with Y (s) being the output and U(s) being the input, nd the time domain response y(t) for
the step input function of the form
u(t) =
_
1 t 0
0 t < 0
Solution: The time domain response is
y(t) =
1
2

1
2
e
2t
Task 2.6:
Find the transfer function
7
G(s) =
Y (s)
U(s)
corresponding to the dierential equation
d
3
y(t)
dt
3
+ 3
d
2
y(t)
dt
2
+ 7
dy(t)
dt
+ 5y(t) =
d
2
u(t)
dt
2
+ 4
du(t)
dt
+ 3u(t)
Solution: The transfer function is
G(s) =
Y (s)
U(s)
=
s
2
+ 4s + 3
s
3
+ 3s
2
+ 7s + 5
Task 2.7:
Find the response
8
y(t) of the system described by the transfer function
G(s) =
Y (s)
U(s)
=
s
(s + 4) (s + 8)
for the ramp input of the form
u(t) =
_
t t 0
0 t < 0
Solution: the time-domain response is
y(t) =
1
32

1
16
e
4t
+
1
32
e
8t
7
This example is taken from the reference [4]
8
This example is taken from the reference [5]
36 CHAPTER 2. LAPLACE TRANSFORM REVIEW
Task 2.8:
Find the Laplace transform
9
of the function
f(t) = te
5t
Solution: The Laplace transform of t is
1
s
2
(directly from the Table ??). Applying the 1
st
shift
theorem (Table 2.2) we get the solution as:
F(s) =
1
(s + 5)
2
Task 2.9:
Find the inverse Laplace transform for the function
10
F(s) =
10
s (s + 2) (s + 3)
2
Solution: The inverse Laplace transform is
f(t) =
5
9
5e
2t
+
10
3
te
3t
+
40
9
e
3t
Task 2.10:
Find the inverse Laplace transform for the s-domain function
F(s) =
3
s (s
2
+ 2s + 5)
Solution: The inverse Laplace transform is
f(t) =
3
5

3
5
e
t
_
cos(2t) +
1
2
sin(2t)
_
Task 2.11:
Find the inverse Laplace transform for the s-domain function
G(s) =
s + 1
s
3
+ 3s
2
+ 2s
Solution: The inverse Laplace transform is
g(t) =
1
2
_
1 e
2t
_
9
This example is taken from the reference [5]
10
This example is taken from reference [4]
Bibliography
[1] Fraklin GF, Powell JD, Emami-Naeini A; Feedback Control of Dynamic Systems - fourt
edition, Prentince Hall 2002 - Chapter 3.
[2] Coughanowr DR; Process Systems Analysis and Controlm- second edition, McGraw-Hill In-
ternational 1991 - Chapter 3
[3] Katsuhiko, O; System Dynamics - third edition, Prentince-Hall inc. 1998 - Chapter 2
[4] Nise, NS; Control Systems Engineering - third edition, John Wiley & Sons, Inc., 2007
[5] Website www.wiley.com/college/nise
37