Digital and Non-Linear Control

Digital Control and Z Transform

1
 Introduction
• Digital control offers distinct advantages over analog
control that explain its popularity.
• Accuracy: Digital signals are more accurate than their
analogue counterparts.
• Flexibility: Modification of a digital controller is possible
without complete replacement.
• Speed: Digital computers may yield superior
performance at very fast speeds
• Cost: Digital controllers are more economical than
analogue controllers.

2
Structure of a Digital Control System

3
Examples of Digital control Systems

Aircraft Turbojet Engine

4
 Difference Equation vs Differential Equation

• A difference equation expresses the change in

some variable as a result of a finite change in the
other variable.

• A differential equation expresses the change in

some variable as a result of an infinitesimal
change in the other variable.

5
 Differential Equation
• Following figure shows a mass-spring-damper-system. Where y is
position, F is applied force D is damping constant and K is spring
constant.

𝐹 𝑡 = 𝑚𝑦ሷ 𝑡 + 𝐷 𝑦ሶ 𝑡 + 𝐾𝑦(𝑡)

• Rearranging above equation in following form

1 𝐷 𝐾
𝑦ሷ 𝑡 = 𝐹 𝑡 − 𝑦ሶ 𝑡 𝑦(𝑡)
𝑚 𝑚 𝑚
6
 Differential Equation
1 𝐷 𝐾
𝑦ሷ 𝑡 = 𝐹 𝑡 − 𝑦ሶ 𝑡 − 𝑦(𝑡)
𝑚 𝑚 𝑚
• Rearranging above equation in following form

𝐹(𝑡) 𝑦ሷ 𝑦ሶ 𝑦
1
 න 𝑑𝑡 න 𝑑𝑡
𝑚

𝐷
−
𝑚

𝐾
−
𝑚

7
 Difference Equation
1 𝐷 𝐾
𝑦 𝑘 + 2 = 𝐹 𝑘 − 𝑦 𝑘 + 1 − 𝑦(𝑘)
𝑚 𝑚 𝑚

𝐹(𝑘) 1 𝑦(𝑘 + 2) 1 𝑦(𝑘 + 1) 1 𝑦(𝑘)


𝑚 𝑧 𝑧

𝐷
−
𝑚

𝐾
−
𝑚

8
 Difference Equations
• Difference equations arise in problems where the time
is assumed to have a discrete set of possible values.

𝑦 𝑘 + 𝑛 + 𝑎𝑛−1 𝑦 𝑘 + 𝑛 − 1 + ⋯ + 𝑎1 𝑦 𝑘 + 1 + 𝑎0 𝑦 𝑘
= 𝑏𝑛 𝑢 𝑘 + 𝑛 + 𝑏𝑛−1 𝑢 𝑘 + 𝑛 − 1 + ⋯ + 𝑏1 𝑢 𝑘 + 1 + 𝑏0 𝑢 𝑘

• Where coefficients 𝑎𝑛−1 , 𝑎𝑛−2 ,… and 𝑏𝑛 , 𝑏𝑛−1 ,… are

constant.
• 𝑢(𝑘) is forcing function

9
 Difference Equations
• Example 1: For the given difference equation, determine the (a)
order of the equation. Is the equation (b) linear, (c) time
invariant, or (d) homogeneous?

𝑦 𝑘 + 2 + 0.8𝑦 𝑘 + 1 + 0.07𝑦 𝑘 = 𝑢 𝑘

Solution:
a) The equation is second order.
b) All terms enter the equation linearly
c) All the terms if the equation have constant coefficients.
Therefore the equation is therefore LTI.
d) A forcing function appears in the equation, so it is
nonhomogeneous.

10
 Z-Transform
• Difference equations can be solved using z-transforms which
provide a convenient approach for solving LTI equations.

• The z-transform is an important tool in the analysis and

design of discrete-time systems.

• It simplifies the solution of discrete-time problems by

converting LTI difference equations to algebraic equations
and convolution to multiplication.

• Thus, it plays a role similar to that served by Laplace

transforms in continuous-time problems.
11
 Z-Transform
• Given the causal sequence {u0, u1, u2, …, uk}, its z-
transform is defined as

𝑈 𝑧 = 𝑢𝑜 + 𝑢1 𝑧 −1 + 𝑢2 𝑧 −2 + 𝑢𝑘 𝑧 −𝑘

𝑈 𝑧 = ෍ 𝑢𝑘 𝑧 −𝑘
𝑘=0

• The variable z−1 in the above equation can be

regarded as a time delay operator.
12
 Z-Transform
• Example 2: Obtain the z-transform of the
sequence

u  
k k 0  1, 1, 3, 2, 0, 4, 0, 0, 0,...

𝑈 𝑧 = 1 + 1𝑧 −1 + 3𝑧 −2 + 2𝑧 −3 + ⋯

13
 Laplace Transform and Z-Transform
• Given the sampled impulse train of a signal
𝑢(𝑡) 𝑢∗ (𝑡)

𝑢(𝑡) 𝑢∗ (𝑡)

𝑈(𝑠) 𝑈 ∗ (𝑠)

Sampled impulse train:

𝑢∗ 𝑡 = 𝑢𝑜 𝛿 𝑡 + 𝑢1 𝛿 𝑡 − 𝑇 + 𝑢2 𝛿 𝑡 − 2𝑇 + ⋯ + 𝑢𝑘 𝛿 𝑡 − 𝑘𝑇
∞

𝑢∗ 𝑡 = ෍ 𝑢𝑘 𝛿 𝑡 − 𝑘𝑇
𝑘=0
14
 Relationship Between Laplace Transform and Z-
Transform
𝑢∗ 𝑡 = 𝑢𝑜 𝛿 𝑡 + 𝑢1 𝛿 𝑡 − 𝑇 + 𝑢2 𝛿 𝑡 − 2𝑇 + ⋯ + 𝑢𝑘 𝛿 𝑡 − 𝑘𝑇
• The Laplace Transform of above equation is
𝑈 ∗ 𝑠 = 𝑢𝑜 + 𝑢1 𝑒 −𝑠𝑇 + 𝑢2 𝑒 −2𝑠𝑇 + ⋯ + 𝑢𝑘 𝑒 −𝑘𝑠𝑇
∞

𝑈 ∗ 𝑠 = ෍ 𝑢𝑘 𝑒 −𝑘𝑠𝑇 (𝐴)
𝑘=0

• The Z-transform of 𝑢∗ 𝑡 is given as

𝑈 𝑧 = 𝑢𝑜 + 𝑢1 𝑧 −1 + 𝑢2 𝑧 −2 + ⋯ + 𝑢𝑘 𝑧 −𝑘
∞

𝑈 𝑧 = ෍ 𝑢𝑘 𝑧 −𝑘 (𝐵)
𝑘=0
• Comparing (A) and (B) yields
𝑧 = 𝑒 𝑠𝑇 where 𝑇 is the sample period
15
 A Note
• In general, given a transfer function in s-
domain you cannot just replace 𝑠 by s =
𝑙𝑛𝑧
(from 𝑧 = 𝑒 𝑠𝑇 ) to get its z-domain
𝑇
transfer function.
• The reason is that 𝑧 = 𝑒 𝑠𝑇 is true with the
sampled signal.

16
 Conformal Mapping between s-plane to z-plane

𝑧 = 𝑒 𝑠𝑇

• Where 𝑠 = 𝜎 + 𝑗𝜔 for real number 𝜎 and real number 𝜔.

𝑧 = 𝑒 (𝜎+𝑗𝜔)𝑇

• Then 𝑧 in polar coordinates is given by

𝑧 = 𝑒 𝜎𝑇 𝑒 𝑗𝜔𝑇

𝑧 = 𝑒 𝜎𝑇 ∠𝑧 = 𝜔𝑇

17
 Conformal Mapping between s-plane to z-plane
• We will discuss following cases to map given points on s-plane
to z-plane.
– Case-1: Real pole in s-plane (𝑠 = 𝜎)

– Case-2: Imaginary Pole in s-plane (𝑠 = 𝑗𝜔)

– Case-3: Complex Poles (𝑠 = 𝜎 + 𝑗𝜔)

𝑠 − 𝑝𝑙𝑎𝑛𝑒 𝑧 − 𝑝𝑙𝑎𝑛𝑒
18
 Conformal Mapping between s-plane to z-plane
Case-1: Real pole in s-plane (𝑠 = 𝜎) 𝑧 = 𝑒 𝜎𝑇 ∠𝑧 = 𝜔𝑇

When 𝑠 = 0

𝑧 = 𝑒 0𝑇 = 1
∠𝑧 = 0𝑇 = 0

𝑠=0

𝑠 − 𝑝𝑙𝑎𝑛𝑒 𝑧 − 𝑝𝑙𝑎𝑛𝑒
19
 Conformal Mapping between s-plane to z-plane
Case-1: Real pole in s-plane (𝑠 = 𝜎) 𝑧 = 𝑒 𝜎𝑇 ∠𝑧 = 𝜔𝑇

When 𝑠 = −∞

𝑧 = 𝑒 −∞𝑇 = 0
∠𝑧 = 0

0
−∞

𝑠 − 𝑝𝑙𝑎𝑛𝑒 𝑧 − 𝑝𝑙𝑎𝑛𝑒
20
 Conformal Mapping between s-plane to z-plane
Case-1: Real pole in s-plane (𝑠 = 𝜎) 𝑧 = 𝑒 𝜎𝑇 ∠𝑧 = 𝜔𝑇

Consider 𝑠 = −𝑎

𝑧 = 𝑒 −𝑎𝑇
∠𝑧 = 0

0 1
−𝑎

𝑠 − 𝑝𝑙𝑎𝑛𝑒 𝑧 − 𝑝𝑙𝑎𝑛𝑒
21
 Conformal Mapping between s-plane to z-plane
Case-2: Imaginary pole in s-plane (𝑠 = ±𝑗𝜔) 𝑧 = 𝑒 𝜎𝑇 ∠𝑧 = 𝜔𝑇

Consider 𝑠 = 𝑗𝜔

𝑧 = 𝑒 0𝑇 = 1
∠𝑧 = 𝜔𝑇

𝑠 = 𝑗𝜔
𝜔𝑇
−1 1

−1

𝑠 − 𝑝𝑙𝑎𝑛𝑒 𝑧 − 𝑝𝑙𝑎𝑛𝑒
22
 Conformal Mapping between s-plane to z-plane
Case-2: Imaginary pole in s-plane (𝑠 = ±𝑗𝜔) 𝑧 = 𝑒 𝜎𝑇 ∠𝑧 = 𝜔𝑇

When 𝑠 = −𝑗𝜔

𝑧 = 𝑒 0𝑇 = 1
∠𝑧 = −𝜔𝑇

−1 −𝜔𝑇 1

𝑠 = −𝑗𝜔
−1

𝑠 − 𝑝𝑙𝑎𝑛𝑒 𝑧 − 𝑝𝑙𝑎𝑛𝑒
23
 Conformal Mapping between s-plane to z-plane
Case-3: Complex pole in s-plane (𝑠 = 𝜎 ± 𝑗𝜔) 𝑧 = 𝑒 𝜎𝑇 ∠𝑧 = 𝜔𝑇

𝑧 = 𝑒 𝜎𝑇
∠𝑧 = ±𝜔𝑇

−1 1

−1

𝑠 − 𝑝𝑙𝑎𝑛𝑒 𝑧 − 𝑝𝑙𝑎𝑛𝑒
24
Mapping regions of the s-plane onto
the z-plane

25
 z-Transforms of Standard Discrete-Time Signals
• The following identities are used repeatedly to derive several
important results.

𝑛 𝑛+1
1 − 𝑎
෍ 𝑎𝑘 =
1−𝑎
𝑘=0

∞
1
෍ 𝑎𝑘 = 𝑎 <1
1−𝑎
𝑘=0

26
 z-Transforms of Standard Discrete-Time Signals
• Unit Impulse

1, 𝑘=0
𝛿 𝑘 =ቊ
0, 𝑘≠0

• Z-transform of the signal

𝛿 𝑧 =1

27
 z-Transforms of Standard Discrete-Time Signals
• Sampled Step
1, 𝑘≥0
𝑢 𝑘 =ቊ
0, 𝑘<0
• or

𝑢 𝑘 = 1 𝑘 = 1, 1, 1,1, … 𝑘≥0

• Z-transform of the signal

𝑛

𝑈 𝑧 = 1 + 𝑧 −1 + 𝑧 −2 + 𝑧 −3 + ⋯ + 𝑧 −𝑘 = ෍ 𝑧 −𝑘
𝑘=0

1 𝑧
𝑈 𝑧 = = 𝑧 −1 < 1
1−𝑧 −1 𝑧−1 28
 z-Transforms of Standard Discrete-Time Signals
• Sampled Ramp
𝑘, 𝑘≥0 𝑟 𝑘
𝑟 𝑘 =ቊ
0, 𝑘<0
……

𝑘
• Z-transform of the signal 0 1 2 3

𝑧
𝑈 𝑧 = 2
𝑧−1

29
 z-Transforms of Standard Discrete-Time Signals
• Sampled Exponential Signal
𝑎𝑘 , 𝑘≥0
𝑢 𝑘 =ቊ
0, 𝑘<0

• Then
𝑛

𝑈 𝑧 = 1 + 𝑎𝑧 −1 + 𝑎2 𝑧 −2 + 𝑎3 𝑧 −3 + ⋯ + 𝑎𝑘 𝑧 −𝑘 = ෍ (𝑎𝑧)−𝑘
𝑘=0

1 𝑧
𝑈 𝑧 = = 𝑎𝑧 −1 < 1
1−𝑎𝑧 −1 𝑧−𝑎
30
 Example 3
• Find the z-transform of following causal sequences.

1. 𝑓 𝑘 = 2 × 1 𝑘 + 4 × 𝛿 𝑘 , 𝑘 = 0,1,2, …

4, 𝑘 = 2,3, …
2. 𝑓 𝑘 = ቊ
0, 𝑜𝑡ℎ𝑒𝑟𝑤𝑖𝑠𝑒

31
 Example 3
• Find the z-transform of following causal sequences.
𝑓 𝑘 =2×1 𝑘 +4×𝛿 𝑘 , 𝑘 = 0,1,2, …

Solution: Using Linearity Property

𝐹 𝑧 = 𝒵{2 × 1 𝑘 + 4 × 𝛿 𝑘 }

𝐹 𝑧 =2×𝒵 1 𝑘 + 4 × 𝒵{𝛿 𝑘 }
𝑧
𝐹 𝑧 =2× +4
𝑧−1
6𝑧 − 4
𝐹 𝑧 =
𝑧−1 32
 Example 3
• Find the z-transform of following causal sequences.

4, 𝑘 = 2,3, …
𝑓 𝑘 =ቊ
0, 𝑜𝑡ℎ𝑒𝑟𝑤𝑖𝑠𝑒
Solution: The given sequence is a sampled step starting at k-2 rather than
k=0 (i.e. it is delayed by two sampling periods). Using the delay property,
we have

𝐹 𝑧 = 𝒵{4 × 1 𝑘 − 2 }

𝐹 𝑧 = 4𝑧 −2 𝒵{1 𝑘 }
−2
𝑧 4
𝐹 𝑧 = 4𝑧 =
𝑧 − 1 𝑧(𝑧 − 1) 33
 Inverse Z-transform
• Partial Fraction Expansion: This method is very
similar to that used in inverting Laplace
transforms. However, because most z-functions
have the term z in their numerator, it is often
convenient to expand F(z)/z rather than F(z).

34
 Inverse Z-transform
• Example 4: Obtain the inverse z-transform of
the function
𝑧+1
𝐹 𝑧 = 2
𝑧 + 0.3𝑧 + 0.02
• Solution
• Partial Fractions
𝐹 𝑧 𝑧+1
=
𝑧 𝑧(𝑧 2 + 0.3𝑧 + 0.02)

𝐹 𝑧 𝑧+1
=
𝑧 𝑧(𝑧 2 + 0.1𝑧 + 0.2𝑧 + 0.02) 35
 Inverse Z-transform
𝐹 𝑧 𝑧+1
=
𝑧 𝑧(𝑧 + 0.1)(𝑧 + 0.2)

𝐹 𝑧 𝐴 𝐵 𝐶
= + +
𝑧 𝑧 𝑧 + 0.1 𝑧 + 0.2

F ( z) 1 1
A z  F (0)    50
z z 0 0.1 0.2 0.02

F ( z) 1 z 1  0.1  1
B  ( z  0.1)  ( z  0.1)   90
z z  0.1 z ( z  0.1)( z  0.2) z  0.1 (0.1)(0.1  0.2)

F ( z) 1 z 1  0.2  1
C  ( z  0.2)  ( z  0.2)   40
z z  0.2 z ( z  0.1)( z  0.2) z  0.2 (0.2)(0.2  0.1)
36
 Inverse Z-transform
𝐹 𝑧 50 90 40
= − +
𝑧 𝑧 𝑧 + 0.1 𝑧 + 0.2
90𝑧 40𝑧
𝐹 𝑧 = 50 − +
𝑧 + 0.1 𝑧 + 0.2

• Taking inverse z-transform (using z-transform table)

𝑘 𝑘
𝑓 𝑘 = 50𝛿 𝑘 − 90 −0.1 + 40 −0.2

37