Solving Difference Equations-Z Transform

z Transform Solution to Difference Equations
3 y  k  2  2 y  k  1  y  k   2u  k  1  3u  k 
ɀ Transform Solution to or
(2)
3 y  k   2 y  k 1  y  k  2  2u  k 1  3u  k  2 
Difference Equations (3)
The system equation (3) can be lumped as follows;

Difference equations arise in problems where the independent
variable, usually time, is assumed to have a discrete set of possible 1
values. The linear differential equations are of the form y  k    2 y  k  1  y  k  2   2u  k  1  3u  k  2  
3
y  k  n   an 1 y  k  n  1  ...  a1 y  k  1  a0 y  k  Supposing the system has the following initial conditions y(-2) =
(1)
 bnu  k  n   bn 1u  k  n  1  ...  b1u  k  1  b0u  k  1, y(-1) = -2 with input signal u(k) = 1, k = 0, 1, 2, … and u(k) = 0
for k < 0. The response to this system can be computed as follows;
1
The forcing (input) function is u(k) and output is y(k). The y  0    2 y  1  y  2   2u  1  3u  2  
coefficients ai, bi, i = 0, 1,2,…., are constants. The difference 3
equation is said to be of order n because the difference between the 1 5
  2   2   1  2  0  3  0  
highest and the lowest time arguments of y(.) and u(.) is n. This 3 3
equation is then referred to as linear time invariant (LTI) difference 1
equation of order n. If the forcing function u(k) is equal to zero, the y 1   2 y  0   y  1  2u  0   3u  2  
3
equation is said to be homogeneous.
1 5  10
  2   2  2  
Consider the discrete-time system shown below 3 3  9
1
y  2    2 y 1  y  0   2u 1  3u  0  
3
1 10 5  26
  2    2  3 
Figure 1 Discrete-time system 3 9 3  27
The input the system is the sequence; u(k) with a sampling interval and so forth. Thus the solution of difference equations can be
T such that u(k) = u(kT), k = 0,1,2,… obtained by direct substitution which is much simpler than for
This system will generate an output sequence; y(k) such that y(k) = differential equations. However, the difference equation obtained
y(kT). Assuming this system is a Linear Time Invariant (LTI) in this way cannot be expressed in the closed-loop form and it is
system whose system inputs can be described by linear difference difficult to extract the solution general properties of the equation.
equations with constant real coefficients such as
Moses Kavi Page 1
Department of Electrical and Communications Engineering
Papua New Guinea University of technology
To overcome this difficulty we introduce the z transform solution Consider f(k-1), it is f(k) shifted to the right or delayed by one
to difference equation. sample period (T) as shown in Figure 2(b). Its z transform is,
Referring to equation (1), and to describe such difference equation  

  f  k  1    f  k  1 z  k  z 1  f  k  1 z
 k 1
in the z plane we take z transform of each term in the equation. (5)
k 0 k 0
First, review shifting/Translational Theorem (see Section 2.4). _

Let k  k  1 , then (5) becomes
Time Delay and Time Advance: Review

_ _  
_ _
Consider the time sequence f(k) shown in Figure 2(a). it is not   f  k  1   z 1  f  k  z  k  z 1  f  1 z   f  k  zk 
necessarily zero for k ≤ 0. Its z transform is, _
k 1
   _
k 0
  
  z 1  f  1 z  F  z   (6)
F  z     f  k     f  k  z k
(4)
k 0
This has a simple physical interpretation. F(z) consists of f(k) with
k ≥ 0. If f(k) is delayed by one sampling period, f(-1) into k = 0 and
must be included in the z transform of f(k-1). Thus we add f(-1)z to
F(z) and then delay it ( multiplying by z-1 ) to give Z[f(k-1)]. Using
this same argument, we have
  f  k  2    z 2  f  2  z 2  f  1 z  F  z   (7)
  f  k  3   z 3  f  3 z 3  f  2  z 2  f  1 z1  F  z   (8)
and so forth
Now we consider f(k+1). This is a shifting of f(k) to the left

(advancing by one sampling period T) as shown in Figure 2(c).
Figure 2 (a) Sequence. (b) Time delay. (c) Time advance Because f(0) is moved to k = -1, it will not be included in the z
Equation (4) is a familiar equation, and it is defined only for f(k) transform of f(k + 1), so it must be excluded from F(z). Thus, we
with k ≥ 0, and f(-1), f(-2),.. do not appear in F(z). subtract f(0) from F(z) and then advance it (multiply by z), to give
Z[f(k+1)] or
Moses Kavi Page 2

  f  k  1   z  F  z   f  0  (9) Now, suppose y(-2) = 1, y(-1) = -2, and u(k) is a unit-step
sequence, then U(z) = z/(z-1) and
Similarly, we have
Y z 
5  2 z 1

 
2 z 1  3z 2 z
1 2 1 2
  f  k  2    z 2  F  z   f  0   f 1 z 1  (10) 3  2z  z 3  2z  z z 1
5z  z
2
z  2 z  3
  f  k  3   z 3  F  z   f  0   f 1 z 1  f  2  z 2  (11)  2 
3 z  2 z  1  3 z  1 z  1 z  1
Solving Linear Difference Equations Using z Transform 


z 5z 2  4 z  2 
 3z  1 z  1 z  1
Consider again the equation given in (3). This equation is of
second order; therefore, the response y(k) depends on the input Also note that the linear difference equation in (1) is characterized
condition u(k) and two initial conditions. by the u(k) and y(k) which are the system input and output
respectively. In obtaining the z transform of such equation, the z
Assume, initial conditions are y(-1) and y(-2) and input signal is transform of each term in the difference equation is taken
u(k) = 0 for k ≤ 0. Application of the z transform (5-8) gives individually. The z transform of the terms in the difference
equation are given in Table 2-3.
3Y  z   2 z 1 Y  z   y  1 z   z 2 Y  z   y  1 z  y  2  z 2 
(12) Review Examples 2-18 and 2-19.
 2 z 1U  z   3z 2U  z 
which can be grouped as
3  2z 1

 z 2 Y  z    2 y  1  y  1 z 1  y  2  
 
 2 z 1  3z 2 U  z 
Thus we have
Y z 
 2 y  1  y  1 z 1  y  2  


2 z 1  3z 2 
U z
3  2 z 1  z 2 3  2 z 1  z 2
 2 y  1 z 2  y  1 z  y  2  z 2   2 z  3z  U z
 1
 2   (13)
3  2z 1 3z  2 z  1
Zero-Input Response Zero-State Response
Moses Kavi Page 3


Solving Difference Equations-Z Transform

Uploaded by

Copyright:

Available Formats

Solving Difference Equations-Z Transform

Uploaded by

Document Information

Original Description:

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Solving Difference Equations-Z Transform

Uploaded by

Copyright:

Available Formats

z Transform Solution to Difference Equations

The system equation (3) can be lumped as follows;

Referring to equation (1), and to describe such difference equation  

First, review shifting/Translational Theorem (see Section 2.4). _

  f  k  3   z 3  f  3 z 3  f  2  z 2  f  1 z1  F  z   (8)

Now we consider f(k+1). This is a shifting of f(k) to the left

Moses Kavi Page 2

Solving Linear Difference Equations Using z Transform 

Moses Kavi Page 3

You might also like

Pfad - The Proxy pFad of © 2024 Garber Painting. All rights reserved.