Professional Documents
Culture Documents
An Introduction To The Laplace Transform and The Z Transform - Anthony C.grove (PH 1991 138s)
An Introduction To The Laplace Transform and The Z Transform - Anthony C.grove (PH 1991 138s)
A. C. Grove
Senior Lectumr in Mathematics, Nottingham Polytechnic
Prentice Hall
New York London Toronto Sydney Tokyo Singapore
First published 1991 by
Prentice Hall International (UK) Ltd
66 Wood Lane End, Hemel Hempstead
Hertfordshire HP2 4RG
A division of
Simon & Schuster International Group
Preface vii
7 Convolution
Answers to exercises
Bibliography
Index
Preface
vii
The Laplace transform and the
inverse Laplace transform
Introduction
One of the most efficient methods for solving certain ordinary and
partial differential equations is the use of Laplace transforms. The effec-
tiveness of the transform is its ability to convert many differential
equations into algebraic equations. Although other methods provide sol-
utions of equations where the input or forcing function is an exponential
and/or sine function the Laplace transform is able to deal with other
inputs such as pulses, square waves, point loads, etc., in exactly the
same way. The differential equations that arise in the analysis of control
systems are particularly suitable for solution by this means.
The Laplace transform is defined in order to develop the necessary
techniques but a more detailed introduction is given in Appendix 1.
0
e-Stf(t) dt
Also
Functions such as t-' and tan(t) do not have Laplace transforms since
the integrals diverge.
The following example illustrates the use of Table 1 . 1 .
I EXAMPLE 1 . 1
EXERCISE 1.1
Use standard formulae to derive the Laplace transform of the following
functions:
EXAMPLE 1.2
Determine the inverse Laplace transform of
2s+3 S-6
(a) (b)
s 2 + 4 s + 13 (S - I)(s - 2)
4s - 9
(c)
(s - 1)*(S- 2)
2s+ 3 2s+ 3
s+212+9
on completing the square. Then
2s+ 3 2(s + 2) 1
(c) 9 - l [ 4s- 9 5
(s - I ) ~ ( -
s 2) ( s - 1)2 s - 2
by partial fractions. Thus
2-1 - [ -
= et + 5te1- e2'
I
6 Laplace transform and z transform
EXERCISE 1.2
Determine the inverse Laplace transform of each of the following functions
of s:
The transforms of derivatives and
the application to differential
equations
Introduction
Electrical circuits
or on elimination of q gives
Particle dynamics
equations
Servomechanisms
The negative sign is required because the torque must oppose the error.
Structures
Economics
The transforms of the first, second, third and fourth derivatives are
given in Table 2.1 and are numbered 2, 3, 4 and 5 respectively in
Similarly
sx- xo
Method
1. Transform the differential equation using all the necessary rules and
standard forms.
2. Solve the resulting equation (usually algebraic) for the unknown as
a function of s.
3. Invert the function of s to obtain the required solution.
EXAMPLE 2.1
Solve the equation
E X A M P L E 2.2
Solve the equation
from which
Inverting,
x = Ae-' + Be - 2 t - te-2'
If non-initial conditions are known such as x = 1, dxldt = 3 when t = 4,
then the constants A and B can be determined by substitution.
EXAMPLE 2.3
Solve the simultaneous equations
EXAMPLE 2.4
Solve the simultaneous differential equations
The transforms of derivatives 15
Note that both equations contain dxldt and dyldt. In this case
generate a third differential equation by eliminating either dxldt or
dyldt (free choice). For example, eliminating dyldt by multiplying
equation (2.6) by 2 and subtracting equation (2.5) gives
Rearranging we have
Inverting, we obtain
EXAMPLE 2.5
If an electron is projected into a uniform magnetic field perpendicular
to its direction of motion its path is given by
He sJ - s2X =
- -
cm
Now eliminate J
Therefore
ucm
x = -sin t)
He
Similarly
y= -=He [I - c o s ( e t ) ]
EXERCISE 2
Solve the following equations:
The transforms of derivatives 17
Y=-
P
[
G sin(nx)-nx
n1
+ 1 - cos(nx)
I
where n = PIEI, and hence that
Some useful theorems
9?{e-"'f ( t ) )= 100
0
e-"e-"' f ( t ) dt
since F ( s ) = 100
0
e-"'f ( t ) d t . Similarly
'
9-( F ( s+ a ) ) = e-"' f ( t )
EXAMPLE 3.1
From Y { t n )= n . / s( n + l )
n!
Lf(e-artn)=
( S + a)"+'
EXAMPLE 3.2
Determine
Y- 21.1
If F ( s + a ) = l / ( s+ 214 then F ( s ) = 1/s4 and f ( t ) = 9 - ' ( F ( s ) ]
9-
1
+ 2)4) = e - 2 t 9 ] t
= e-Zt -
3!
If F ( s ) = ~2?{
f ( t ) ) then
Some useful theorems 21
Similarly
EXAMPLE 3.3
Determine
EXAMPLE 3.4
Determine
i dt] =! g l i l
,[[: 1 -
= -s
where i = Y ( i ) .
22 Laplace transform and z transform
EXAMPLE 3.5
Evaluate
EXAMPLE 3.6
Determine
Using (6)
Note that in Examples 3.5 and 3.6 the same results could have been
obtained by other methods. For example
EXAMPLE 3.7
Solve the differential equation
di t
L - + R ~ + ~ idt=E
dt c 0
1
for the current i in a circuit if L, C, R and E are constants and i = 0
when t = 0.
(
i Ls+R+-
is)-:--
- E
I =
(LS' + R s + l/C)
If we assume that R' < 4L/C, a = R/2L and 0' = 1/LC- ~ ' 1 4 ~then '
P' > 0 and the solution on inversion becomes
E
i = - e-"' sin(@)
LP
However, if R 2 > 4L/C, a = R/2L, y 2 = ~ 1 then4 Y2 > ~0.
~ - l/LC ~
Hence
- E
I =
L [ ( s + a ) ' - Y2]
and the solution becomes
I=- e-"' sinh(?t)
LY
Differentiation of transforms
I f F ( s ) = 9 {f ( t ) )then
d
91- t f ( t ) )= - 1F(s)l
ds
and more generally
24 Laplace transform and z transform
or in inverse form
EXAMPLE 3.8
Find 6e(t2 sin(2t)) .
EXAMPLE 3.9
Find y if
Some useful theorems 25
EXAMPLE 3.10
Find
= [tan-l(i)]a=--
.s
7r
2 tan- (i)
k
Figure 3.1
EXAMPLE 3.1 1
Find y if J = s / ( s 2 - 112.
26 Laplace transform and z transform
EXAMPLE 3.12
Solve the differential equation
Since xo = 0, xd = A ,
x = Cte -''
Some useful theorems 27
I EXAMPLE 3.13
I Determine the Laplace transform of the square wave defined by
-
-
1-e-~T
[I
'I2
0
~ e - , dt - 1
T
T/2
Ee-" dt]
28 Laplace transform and z transform
F(s) =
S
tanh (T)
EXAMPLE 3.14
Solve the equation
where f(t) is the square wave defined in Example 3.13 and given that
x = dxldt = 0 when t = 0.
Transforming, we have
Therefore
The only standard form which might enable us to deal with F(s)
multiplied by an exponential is standard form 8 of Appendix 2 which
is yet to be discussed. We shall therefore leave the formal inverse until
later; however, we need the exponential function to be in the
numerator. Therefore
Some useful theorems 29
PROOF 9 (f'(t)) = ] 0
e -st f t (t) dt. Integration by parts gives
fa-
limit
s-00
J0 eaStf'(t) d t = O
Therefore
limit [sF(s) - f (O)] = 0
s+a
0
e-"ft (t) dt
00
= f ' (t) dt
= limit
t-00
1f 0
t
'(t) dt
Hence
limit f (t) = limit sF(s)
t-'- s-0
Applying the final value theorem, both limits must be zero. Therefore
A = 0. Hence
EXAMPLE 3.16
If sy = - tan-'(s) + C and y = 0 when t = 0 determine y.
Therefore C = 4 2 . Hence
sjj = - tan-'(s) + -T2 = tan-'
Since
Since
EXERCISE 3.1
Determine the Laplace transforms of each of the following functions where
O,a, b, k and T are constant:
32 Laplace transform and z transform
Period 27r
12 y = k t O<r<T
Period T
Period 2
Introduction
The Heaviside step function (unit step function) denoted by u(t) or H(t)
is defined as (Figure 4.1)
only. From the graph it can be seen that an alternative name is the 'unit
step function' which may be interpreted as a model of a switch that is
switched on at time t = T.
The combination of the step function and any other function f ( t ) , i.e.
EXAMPLE 4.1
Sketch the function
y = 3 4 1 - 1 ) - 2 u ( t - 2)
and determine its Laplace transform.
EXAMPLE 4.2
Determine an expression for the function
a<t<b
elsewhere
This function is considered in two parts: the operation at t = a (switch
on), and the operation at t = b (switch off). At t = a there is a step of
magnitude + 5, i.e. 5u(t - a). At t = b there is a step of magnitude
- 5, i.e. - 5u(t - b). Therefore
EXAMPLE 4.3
Determine an expression for
EXAMPLE 4.4
Determine the Laplace transform of (t - 2)u(t - 2).
1
g ( ( t - 2)u(t - 2)) = e - 2 s S ( t ) = e - 2 L
s
since T = 2 and f (t - 2) = t - 2, so that f(t) = t and F ( s ) = l/s2.
EXAMPLE 4.5
Find the Laplace transform of
f(t) =
-e
O-'[ t<T
t>T
f(t) = -e-'u(t - T ) = -e -(I-T+T) u(t - T )
-
--e - Te-(t-T) u (t - T )
Note that t is written as t - T + T and that the laws of indices are used
to write e - ( I - T + T ) as e -(I-T) e - T .
9 ( f ( t ) ] = S ( - e - Te -(I-T) u(t - TI1
- 1
- - e - T e -ST
s+ 1
-(s+ 1) T
-
- -e
s+ 1
using 8 and 17 of Appendix 2.
9 Laplace transform and z transform
EXERCISE 4.1
Sketch the graph of each of the following functions and find their Laplace
transforms:
Write the following as a single equation using the unit step function; sketch
the graphs and find the Laplace transforms:
Express the following as a single equation using the step function and find
the Laplace transform:
9
The Heaviside step function 39
12 Show that
cos(t - 4 3 ) t )4 3
elsewhere
t>l
elsewhere
IS f ( t ) = r 03 ' t>l
elsewhere
40 Laplace transform and z transform
y= [ E
-E
E
O<t<T
T<t<2T
2T<t<3Tetc.
Show that
20 Determine jj if y = 1,2,3, ... for the intervals (0, T), (T, 2T),
(2T, 3 T), ... (the staircase function).
EXAMPLE 4.6
Determine
EXAMPLE 4.7
Solve the differential equation
where
and y = 1 when t = 0 .
The Heaviside step function 41
Since
the inverse of p is
Hence
I EXAMPLE 4.1
I
The solution of Example 3.14 of a square wave forcing function now
becomes
42 Laplace transform and z transform
EXAMPLE 4.9
A light horizontal beam of length I is clamped at each end and carries
a load w Nm-' over the length x = a to x = b measured from the left-
hand end of the beam. Determine the deflection of the beam.
- x 3 +-
y =A x 2 + B [(x - a)'u(x - a ) - ( x - b14u(x - b)]
2 6 24EI (4.1)
Figure 4.5 (a) The step function y = u(t); (b) the function
y=u(-t)
44 Laplace transform and z transform
4. u ( T - t ) = u [ - ( t - T ) ]= 1 - u ( t - T ) .
5 . Scaling: # ( a t )= u ( t )
i.e. the step at t = 0 is independent of a;
EXERCISE 4.2
Determine the current in the circuit if both current and charge are
initially zero, i.e. solve the equations
Introduction
a
p(x) d x
density for the total point mass (charge) to be equal to unity, i.e.
However, 6(x) = 0 for all x except x = 0 and 6(0) must be infinite. For
a particle of unit mass at xo, the corresponding density is 6(x-xo).
Although the impulse function and associated functions, referred to as
generalized functions, are idealized functions, the operations involving
them can be made mathematically rigorous. These functions can be
combined algebraically, integrated and differentiated. Their value in the
analysis of engineering and physical systems is considerable.
A more general property is that
Figure 5.3 (a) The combination y =f(t) and y = 6(t); (b) the
combination y =f(t) and y = 6(t - a )
EXAMPLE 5.1
Solve the differential equation
Transforming we have
(s2p- sxo - xd) + ~ ( S -
P XO)+ 2P = 5e-2s
Since xo = 4, xd= 0,
P(s2 + 3s + 2) = 5e-2Q 4s + 12
From Appendix 2
1 -t -21
~ - ~ [ ( ~ + l ) ( s + 2= )e) - e
and
The impulse function 51
Therefore
X = 5 ie-('-2) - e-2(f-2)]u(t - 2) + 4(2e-" - e-')
+ 12(e-' - e-2')
= 5 ie-('-2) - e-2(f-2) 1~ (-t2) + 8e-I - 4e-2'
EXAMPLE 5.2
Solve for i the equations
EXERCISE 5
Transfer functions
where a , b, c are constants and all the initial conditions are zero. In a
stable system the roots of the auxiliary equation must have negative real
parts so that the ratio bla is positive.
Taking Laplace transforms of both sides produces the algebraic
equation
where s is the Laplace transform variable and X(s),T(s) are the Laplace
transforms of x and f(t) respectively. The function as2+ bs + c is called
the characteristic function and the equation
The function on the right-hand side is the transfer function relating X(s)
and T(s) and is the ratio
Laplace transform of output
Laplace transform of input
The usual notation for a transfer function is Y(s), G(s). A block
diagram (Figure 6.1) can be used to represent such a relationship.
A transfer function can also be obtained by one of the following
methods.
The function on the right-hand side is called the transfer function of the
system and is the ratio of output to input.
This form will be encountered in the discussion of the Fourier trans-
form.
D operator
If D = d/dt and D 2 = d2/dt2 then equation (6.1) can be written
so that
Therefore
6
5
= Y4 ( s )Y3 ( s )Y2 ( s )Yl ( s )
1
(6.6)
Hence the transfer functions of blocks in series are multiplied and the
system can be represented by a block diagram as shown in Figure 6.3.
In practice it may be necessary to compare the output 9, to a par-
ticular value or set of values, say Bi, so that suitable correction voltages
E are applied to the amplifier. The equation for an error sensor is then
Typical examples
1. Consider the system shown in Figure 6.5. From (6.6) and (6.7)
above the equations are
Transfer functions, block diagrams, stability 57
Figure 6.5
Therefore
2. A system (Figure 6.6) may also have a transfer function in the feed-
back loop. The equations are
80 = YI (s) Y2(s)
-
E
-x
80
= Y3(s)
Figure 6.6
58 Laplace transform and z transform
EXAMPLE 6.1
Determine the steady state gain of the system whose transfer function
is
lOO(1 + 2s)
= limit
S-o ( s 2 + 8 s + 6 ) ( s + 3)
The prime reason for using feedback is to minimize the error between
the actual system output and the desired system output. The magnitude
of the steady state error, the value to which the error signal tends as the
transient disturbance from any change of input dies away, is a measure
of the accuracy of the system.
Consider the system defined in Figure 6.9. The equations are
By definition
steady state error = limit e ( t )= limit si?
t+- s + 0
Figure 6.9
60 Laplace transform and z transform
EXAMPLE 6.2
- 1 ) ( ~- 0.5)
S(S
G ( s )=
( s + 1 ) 2 (s 2 + S + 1 )
The, zeros are at s = 0, s = 1, s = 0.5. The poles are at s = - 1 and
+
s = - 112 j&2. Since G ( s )+ 0 as s + + 0 0 , G ( s ) is also considered
to have a zero at infinity.
Since s may be complex all finite values of poles and zeros can then
be plotted on an Argand diagram (Figure 6.10) where x and 0 mark
the poles and zeros respectively.
* Im(s)
-- 1
X
*
+
f\ f3 A
*
\/ w w
-1 1 Re@)
X
--- 1
Figure 6.10
Transfer functions, block diagrams, stability 61
where G(s) is the transfer function of the system and X(s) is dependent
on the input to that system. Hence the poles of Y(s) can arise from
1. the poles of G (s) and
2. the poles of X(s).
The portion due to the poles of G(s) is defined as the natural response
and that due to the poles of X(s) the forced response. If the system has
all its poles in the left-hand half of the s plane these values of s have
negative real parts so that the natural response is a combination of nega-
tive exponential terms which tend to zero as t 0 0 , i.e. a transient
+
response.
Stability considerations
By a stable system we mean a system that yields a finite output for every
finite input. Stability considerations can be represented by the transfer
function G (s).
A system is stable if all the poles of G(s) lie in the left-hand half of
the s plane since again the response will contain only negative
exponential terms, e.g.
+
The first-order complex conjugate pair s = j a defines an oscillatory
system.
Note that the coefficients in the numerator which determine the zeros
only affect the constants in the calculation of partial fractions and not
the form of those fractions. The zeros therefore contribute to amplitude
and phase but do not influence the form of the time function and hence
stability.
since
arg[Y(jo)l = arg[G(jo)X(jo)l
= arg [G(jw)l + arg [X(jo)l
(the multiplication of complex expressions in polar form).
EXAMPLE 6.3
Determine the magnitude response and phase response of
l + s
G(s) =
1+s+s2
l + s
G(s) =
1+s+s2
1 +( j 4
G(jo) =
1 + +
(jw) ( j ~ ) ~
- l+jo
(1 - w2) j w +
magnitude response = I G(jo) (
= (1
-
I 1+jo
- w2) +jw
- Il+jwI
l(1 -w2)+jwl
2 1/2
-
- (l+o
[(I- w212 w 2] + 1/2
EXERCISE 6
Determine the poles, zeros, magnitude response and phase response of the
following functions of s:
64 Laplace transform and z transform
Convolution
EXAMPLE 7.1
Determine the convolution of f(t) and g(t) if f(t) = e', g(t) = t2.
I
U2el-rr
d u = 2e1- t 2 - 2t - 2
0
EXAMPLE 7.2
Determine f *g if
and f (t ) = sin(t )
66 Laplace transform and z transform
I so that
f*g= 1I
3
2 sin(t-u) du for 3 < t
I Hence
EXAMPLE 7.3
Determine
Convolution 67
1 t e a
e s + 1 ) , ] = 1 1o6 ( u - a ) ( t - u ) d u =
EXAMPLE 7.4
Using Laplace transforms and the convolution integral solve the
differential equation
EXAMPLE 7.5
Determine the convolution of the functions
t2 t > O 1 O<t<l
0 t<O 0 elsewhere
I
h *x =
So [ l - H(u - 1)] (t - u ) du
~
Figures 7.l(a) and 7.l(b) show graphs of h(u) and x(u). The iden-
tification of points A and B emphasizes the operations of folding
(Figure 7.1 (c), graph of x ( - u)) and shift (Figure 7.1 (d), graph of
<
x(t - u)). Figure 7.l(e) shows the graph of h(u)x(t - u) for 0 t < 1.
Hence
y= 1I
0
I r u 2 du = j t 3 O<t<l (7.2)
Figure 7.1 (a) Graph of h(u); (b) graph of x(u); (c) graph of
x(- u ) (folding); (d) graph of x(t - u) (shift); (e) graph of
h(u)x(t - u), 0 < t < 1; (f) graph of h(u)x(t - u), t 2 1
Convolution 69
where x(t) is an unknown function and f(t), h(t) are known functions.
Subsequently many problems in thermodynamics, electrical systems,
nuclear reactor theory and chemotherapy have been modeled with this
type of equation. Note that the integral in equation (7.4) is a convo-
lution integral so that taking Laplace transforms and using number 10
of Appendix 2 gives
EXAMPLE 7.6
Solve the Volterra integral equation
x(t)=e-'-4 cos[2(t-u)x](u)du
- 5 4 8
--- +
s+1 (s+2) ( ~ + 2 ) ~
by partial fractions. Inverting we obtain x(t) = 5e-' - 4e-2' - 8te- 2'.
>
Convolution 71
EXERCISE 7
K constant
'
which is a power series in z - or 112. F ( z ) is called the z transform of
the sequence or discrete-time signal f ( n T ) and the relationship is written
as
EXAMPLE 8.1
Determine the z transform of the sampled function f ( n T ) if f ( t ) is the
unit step function u ( t ) .
n=O
EXAMPLE 8.2
Determine the z transform of the sampled function f ( n T ) if f ( t ) is the
ramp function f ( t )= t .
EXAMPLE 8.3
Determine the z transform of the sampled function f(n) if f(t) = e-"
where a is a constant.
- 1 - Z
1 - e-az- 1 - e-a
which is convergent for I z 1 > e-". Therefore
These results can be proved from the definition (Doetsch, 1971; Karni
and Byatt, 1982).
1. Addition:
,yv1(n) + f2(n)1 = r v 1 ( n ) ) + yrf2cn)r
Sampled-data control systems 77
Multiplication by a constant:
These are shown in the table of standard forms (23 and 19-22
respectively).
Initial value theorem:
and
78 Laplace transform and z transform
EXAMPLE 8.4
Using the table of standard forms determine the z transforms of the
following sampled functions:
(a) f ( n ) = n (b) f (n ) = Ke - O n - 2 sin(bn) if a, b and K
are constants.
(c) f ( n ) = e - 2 ( n - 2 ) ~ ( n - 2 ) (d) f ( n ) = 2 " ( n 2 - n )
(a) f(n)=n:
2
F(2) =
(2 - 1)2
using number 4.
(b) f(n)=Ke-On- 2 sin(bn):
Kz - 22 sin(b)
F(2) =
z - e-" z2 - 22 cos(b) + 1
using numbers 6 and 8.
(c) f(n)=e-2(n-2)u(n-2):
F(z) = z - 2 r { e - 2 n ]
Z
= 2-2
z - e-2
using numbers 23 and 6.
(d) f(n) = 2"(n2 - n). From number 27 it is necessary first to
determine the transform of n - n:
z(2 + 1) - 2 22
d(n2-n) = -
(Z - i 3 (Z - 112- (2 - 113
The required transform is
2(2/ 2) - 82
[(2/2) - 11) - ( z - 213
EXERCISE 8
Hence
E X A M P L E 9.1
Determine the inverse z transform of
z2+z
F(z)=
z3-3z2+3z-1
Z-' + 4z-2 + 9z-3 + 1 6 ~ - ~ . . .
z3-3z2+3z-1 1 z2+z
z2-3~+3-~-'
4~-3+Z-'
42 - 12 + 122-' - 4zW2
9- l l z - ' + 4 ~ - ~
'
9 - 272- + 27z-2 - 92- 3
162- ' - 2 3 ~ +- 9z-3
~
Hence
, F ( z ) = z- ' +4 ~ + 9z-3
- ~ + 1 6 ~ -..~ .
so that ao=O, a1 = 1 , a2 = 4 , a 3 = 9 , a4= 16, ..., and f(O)=O, f ( l ) = 1,
f(2) = 4, f(3) = 9, f (4)= 16, ..., which suggests that f ( n ) may be equal
to n2.
A disadvantage of this method is that, in general, only the sequence
values are obtained and additional techniques are necessary to obtain a
closed form version of f ( n ) .
Other methods can be used to derive the series expansion, e.g. the
binomial theorem can be used to expand z/(z + 1 ) in a series of
negative powers of z:
-- - ( l + Z - l ) - l = l + Z - 1 +...
z+ 1
-
The approach here parallels that used to obtain the inverse Laplace
transform of F ( s ) . In this case, however, we work with F ( z ) / z instead
of F ( s ) . Why F ( z ) / z and not F(z)? Examining our standard forms we
82 Laplace transform and z transform
EXAMPLE 9.2
Determine f(n) if
Hence
EXAMPLE 9.3
Determine f(n) if
The inverse z transform 83
EXAMPLE 9.4
Determine the inverse z transform of
and since
sinh(wo) = [cosh2(wo) - 11 'I2
Now
Therefore
2
f (n) = cosh(won) + - (i - k)sinh(oon)
J5
-1 3
where wo = cosh (z).
84 Laplace transform and z transform
Residue method
Pole
Z -
limit G(z) = 00
ZI
Residue
EXAMPLE 9.5
Determine the residues of
For z = - 3 we obtain
R-3 = 9/16
There is a pole of order 2 at z = 1:
EXAMPLE 9.6
Determine the inverse z transform of
4
EXAMPLE 9.7
Determine the inverse z transform of
Consider
For n 2 1
Hence
f(n) = (- I)"-' - (-2)"-' rial
The inverse z transform 87
Figure 9.1
Note that the function f(n) or f(nT) obtained using either of these
methods only describes the function at the sampling instants. There is
no information available between sampling instants. For example, both
of the functions f i and f2 in Figure 9.1 pass through the given data
points at tl, t2, t3, t4, t5 and ta.
EXERCISE 9
Determine the inverse z transforms of the following.
88 Laplace transform and z transform
where g(n) and f(n) are two sequences. Some typical examples follow.
Digital filter
If in is the current in the nth loop and V ,is the voltage at node An
(Figure 10.1) then the relationship between successive loop currents is
the voltages at two previous nodes. Similarly the loop equation (10.1)
yields a second-order equation in the loop currents.
Simultaneous equations can also be generated. For example
Bending moments
where w is the weight per unit length of the beam and a is the distance
between the supports.
Population growth
Savings account
Computing
Sampled-data system
1. additionlsubtraction,
2. multiplication/division,
3. summation,
4. single-sample delay,
then, since the inputs and outputs are simply strings of numbers, the
system can be implemented either as a set of equations or by the use of
special-purpose electronic circuits. In either case the system can be
defined in terms of elementary building blocks that specify the basic
mathematical operations.
From Figure 10.2 the equation is
E X A M P L E 10.1
Solve the difference equation
f ( n + 2 ) - 2 f ( n + 1)+ f ( n ) = 3n n>2
given that f(0) = 0 and f(1) = 2 .
Using the residue method for the first term and standard form 4
(Appendix 3) for the second term we obtain
EXAMPLE 10.2
Solve the difference equation
2 f ( n ) - 3f ( n + 1 ) =(- I)"-'
if f ( 0 ) = 2.
z 3z +- 62
-- +
5(Z+ I ) 5 ( 3 z - 2) 3 2 - 2
in partial fractions. Thus
Hence
EXERCISE 10
In the same way that differential equations and the Laplace transform
enable continuous-time systems to be represented by transfer functions,
so can discrete-time systems be represented by transfer functions. We
are dealing, of course, with difference equations and the z transform,
which then allows discussion of frequency domain concepts such as
magnitude and phase response.
Consider a system with input x(n), output y(n) (Figure 11.1) deter-
mined by the coefficients ai, bj such that the input-output relationship
is defined by a difference equation
Note that the input x(n) is the output of a sampler which is assumed to
be ideal. Assuming that all initial conditions are zero, the z transform
of equation (1 1.1) yields
Figure 11.1
Hence the overall transfer function is the product of the two separate
transfer functions.
Figure 11.3 shows case 2. In this case the second element is driven at
the sampling instants and also between sampling instants. The z trans-
form of the output can be shown to be
Figure 11.2
Discrete transfer function 97
>
Figure 11.3
Now integrate:
I
S. T. R. Hancock, 'The Laplace transform', Mathematical Gazette, 361 (1963),
215-19.
99
100 Laplace transform and z transform
d X - p[xept
ept -
dt
-p 1 xept dt] - 6xept + 6 p 1 xept dt + 8 1 xept dt
(s2 - 6 s + 8) 1
00
0
9
e-"x dt = -+ 2 s - 3
s-5
1
o
e-"x dt =
2s2 - 13s + 24
(S - 5)(s2 - 6 s + 8)
Appendix 1 101
and that the method operating on the right-hand side of the original
differential equation (9e5') produced 91 ( s - 5).
The theorem implies that all bounded continuous functions and every
polynomial have Laplace transforms. Thus tn, n = 1,2,3, ... has a
Laplace transform but el2 does not. It should also be noted that the
102 Laplace transform and z transform
conditions are not necessary for the existence. There are functions, such
as ?-'I2, that do not satisfy the conditions but yet have a Laplace
transform.
7
Hence we assume that the Laplace transform and the inverse, for all
functions of practical interest, are unique.
Appendix 2
A table of Laplace transforms
3 Second derivative
4 Third derivative
5 Fourth derivative
(Continued)
41 (a) { < 1 and od=wo(l - { 2 ) 1/2 u(t) - e- loot
where a d is the frequency of free damped
{wo
oscillation. cos(wdt) + - sin(wdt)
ad oscilla
(b) t = 1 u(t) - e-@"[l + wet]
(c) { > 1 and ~ = w o ( { ~1)'12
- u(t) - e - toot
43 Rectangular pulse
44 Rectangular
periodic wave,
period T
45 Hal f-wave-rectified
sine, period
T = 2?r/w T/2<z<T (~~+
46 Full-wave-rectified
sine, period f(t) = I sin(wt) I
T = 2dw
Initial value theorem limit f(t) = limit sF(s) Final value theorem limi
1-0 S--00 t-tw
Appendix 3
A table of z transforms
1 f(O)=k Impulse at n = 0 k
f ( n ) = 0 , n = 1,2, ...
2 f(m)=k Impulse at n = m kz-"I
f(n) = 0 , n f m
kze -'
z 2 - 22 cos(o0)+ 1
(Continued)
112 Laplace transform and z transform
12 a nwhere a is constant
f l ( z ) - z f (0)
z 2 ~ ( z-)z2f(0)- zf(1)
z3~(-
z )z 3 f ( 0 )- z 2 f ( 1 )- z f ( 2 )
zU'F(z)- zn'f(0)- *..- zf(m - 1 )
(Continued)
Appendix 3 113
2- 1
Final value theorem limit f ( n ) = limit -F ( z )
-
Appendix 4
Iypes ot sampling
Natural sampling
Flat-topped sampling
Figure A4.1 (a) A signal y = m(t); (b) sampling waveform; (c) naturally
sampled signal
the merit that it simplifies the circuitry required to perform the sampling
operation. The distortion results from the fact that the original signal
was 'observed' through a finite rather than an infinitesimal time 'aper-
ture' and hence is referred to as the aperture efect. The spectrum is
found to be multiplied by the sampling function Sa(x).
Answers to exercises
Exercise 1.I
118 Laplace transform and z transform
Exercise 1.2
1 it3
3 2t - 3 + 5e-'
Exercise 2
Answers to exercises 119
Exercise 3
120 Laplace transform and z transform
Exercise 4.1
1
1 - (4 - e - S )
S
13 (a) 2 ( t - l ) u ( t - 1 )
(b) cos(t) + [cos(2t)- cos(t)]u(t - n ) + [cos(3t)- cos(2t)l u ( t - 2 n )
Answers to exercises 121
Exercise 4.2
cos [3(t- I ) ] u ( t - 1 )
f
( e2"- " + cos[2(t- I ) ] + j sin[2(t- l ) ] ) u ( t - 1 )
3 -(t-2)
$ ( t- 2 ) e u(t - 2) - cos[2(t- l ) ] u ( t- 1 )
3 sin[3(t- 3)Ju(t- 3)
(t - 1)ze-2(t- 1) ~ (- t1 )
$ ( t - 5 ) 3e 3 ( r - 5 ) u(t - 5 )
[ l -cos(t- 2 ) J u ( t - 2 ) - [ l - c o s ( t - 3 ) ] u ( t - 3 )
i = [5e-(t-') + e 5 ( ' - ' ) - 61 u(t - 1 ) + eS' - e-'
x = t ( l - 2e-(r-2) + e - 2 ( ' - 2 ) ) u ( t - 2)
x = 2 - C O S ( ~ )- 2 [ l - C O S ( ~- 3)]u (t - 3)
y= [8e-(t-l) - gte-2(t-1) ] ~ ( t I-) - ( 1 + t)e-2t
i = $ [ I -cos(4t)] -:: 11 - c o s [ 4 ( t - Z ) ] ) u ( t - 2 )
x = 3 sin(2t) + :t sin(2t) - :(t - sin [2(t- u(t -
'A)] 'A)
16 i2 = !
!! Ie-2, sinh
J6
(di t) - e-2('- 0 sinh [& (t - I ) ] u(t - 1 ) )
Exercise 5
122 Laplace transform and z transform
9 x=- e sin[w(t-t~)Ju(t-to)
mu
where a = c/2m and w2 = k - c 2 / 4 m 2
Exercise 6
-+ tan-'(o)
7r
1 s=o
2
'
tan- (w/ 3 )
Answers to exercises 123
Exercise 7
(b) i = f(t) *-
1
b-a
(e-"' - e-") = 5 I
o
f(t - U)
e-'lr - - b/t
b-a
du
6 (a) x=-t+et-e-'
(b) x = 3(e1+ e-') * sin(t) = 5 e' + e-' - 3 cos(t)
Exercise 8
124 Laplace transform and z transform
Exercise 9
1
6 - sinh(w0n) where w o = cosh-'(2) 0, 1, 4, 15,...
J3
2
7 cosh(w0n) + - sinh(w0n) where wo = cosh- (2) ' 1, 4, 15, ...
P
8 ( n - 2)
sin(?)
10 - (n - 3)(- 2 ) ( n - 3 ) ~ (-
n 3) 0 , 0 , 0 , 0 , 2 , -8,24 ,...
11 -$(-1)"-n(-1)"+$(2)"
12 (n2+3n+2)(2)" 13 f(-3)"+$(4)"
Note that the sequence values may be obtained from the function f(n) by
substitution of successive values of n = 0, 1, 2, 3, ... .
Exercise 10
1
11 i(n) = - - sinh(o0n) where wo = cosh- (2) '
J3
6
13 i(n) = 2 cosh(o0n) +- sinh(o0n) where o o = cosh- ' (:)
45
14 f ( n ) = -!(-!)"+!-3n
Alternatively by division of the polynomials f (n) = 0, 1, - 4, - 6, - 9.5, ... .
Bibliography
Further reading
Doetsch, G. (1971) Guide to the Applications of the Laplace and z-transforms,
Van Nostrand Reinhold.
Erdklyi, A. (1954) Tables of Integral Transforms (Bateman Manuscript Project,
2 vols), McGraw-Hill.
Kaplan, W. (198 1) Advanced Mathematics for Engineers, Addison-Wesley .
Karni, S. and Byatt, W. J. (1982) Mathematical Methods in Continuous and
Discrete Systems, Holt , Rinehart and Winston.
Spiegel, M. R. (1965) Theory and Problems of Laplace Transforms, Schaum's
Outline Series, McGraw-Hill.
Stroud, K. A. (1 973) Laplace Transforms - Programmes and Problems, Stanley
Thornes.
Watson, E. J. (1981) Laplace Transforms and Applications, Van Nostrand
Reinhold.
Applications
The book:
contains numerous worked examples to support
the theoretical material, and end-of-chapter
exercises
is oriented towards engineering applications
includes the necessary theory, but without going
into in-depth mathematical detail.
ISBN 0-LA-qBBYJJ-7
I
PRENTICE HALL