61
2. Z-transform and theorem Gc(z) ZOH GP(s) R(z) E(z) M(z) GHP(z) Computer system Y(z) Plant A/D D/A Gc(s) GP(s) R(s) E(s) M(s) Y(s) Controller Plant

2. Z-transform and theorem G c (z) ZOH G P (s) R(z) E(z) M(z) G HP (z) Computer system Y(z) Plant A/D D/A G c (s) G P (s) R(s) E(s) M(s)Y(s) Controller

Embed Size (px)

Citation preview

2. Z-transform and theorem

Gc(z) ZOH GP(s)R(z) E(z)M(z)

GHP(z)

Computer system

Y(z)

Plant

A/D D/A

Gc(s) GP(s)R(s) E(s) M(s) Y(s)

Controller Plant

2. Z-transform and theorem

time

f(t)

Time kT

f(kT)

A/D

Time kT

f(kT)

D/A

Time kT

f(kT)

2. Z-transform and theorem

How can we represent the sampled data mathematically?

For continuous time system, we have a mathematical tool Laplace transform. It helps us to define the transfer function of a control system, analyse system stability and design a controller. Can we have a similar mathematical tool for discrete time system?

2.1 Z-transform

For a continuous signal f(t), its sampled data can be written as,

Then we can define Z-transform of f(t) as

where z-1 represents one sampling period delay in time.

0

)()2()()0()()(n

s nTfTfTffkTftf

21

0

)2()()0(

)()]([)]([)(

zTfzTff

znTfnTfZtfZzFn

n

2.1 Z-transform

Solution:

Example 1: Find the Z-transform of unit step function. f(t)

t

kT

f(kT)

2.1 Z-transform

Apply the definition of Z-transform, we have

121

32

21

0

1

11)(

1

1

)()]([)]([)(

zzzzF

q

aaqaqaqa

zz

zkTfkTfZtfZzF k

k

2.1 Z-transform

Another method

11

3212111

21

0

1

1)(1)()(

)1()(

1

)()]([)]([)(

zzFzFzzF

zzzzzzzFz

zz

zkTfkTfZtfZzF k

k

2.1 Z-transform

Example 2: Find the Z-transform of a exponential decay.

Solution: f(t)

t

1

221

0

1

1)(

1

)]([)(

)(

zezF

zeze

zekTfZzF

ekTf

aT

aTaT

k

kakT

akT

2.1 Z-transform

Exercise 1: Find the Z-transform of a exponential decay f(t)=e-at using other method. f(t)

t

2.1 Z-transform

Example 3: Find the Z-transform of a cosine function.

Solution: As

2sin ;

2cos

sincos ;sincos

j

eet

eet

tjtetjtetjtjtjtj

tjtj

2.1 Z-transform

21

1

21

1

211

11

11

11

11

1

cos21

cos1

)cos21(2

cos22

)(1

)(2

2

1

)1)(1(

11

2

1

1

1

1

1

2

1)(

1

1][

])[][(2

1

2][cos)(

zTz

Tz

zTz

Tz

zzeze

zeze

zeze

zeze

zezezF

zeeZ

eZeZee

ZkTZzF

TjTj

TjTj

TjTj

TjTj

TjTj

aTat

tjtjtjtj

2.1 Z-transform

Exercise 2: Find the Z-transform for decayed cosine function tetf at cos)(

221

1

cos21

cos1 )(

zeTez

TezzF

aTaT

aT

2.1 Z-transform

Example 4: Find the Z-transform for

Solution:

atetf 1)(

)1)(1(

)1(

1

1

1

1)(

1

1][ ;

1

1][]1[

][]1[1)]([)(

11

1

11

11

zez

ze

zezzF

zeeZ

zstepZZ

eZZeZkTfZzF

aT

aT

aT

aTat

atat

2.1 Z-transform

Exercise 3: Find the Z-transform forattetf )(

21

1

)1( )(

aT

aT

ez

eTzzF

2.1 Z-transform

The functions can be given either in time domain as f(t) or in S-domain as F(s). They are equivalent. eg.

a) A unit step function: 1(t) or 1/s

b) A ramp function: t or 1/s2

c) f(t)=1-e-at or a/(s(s+a))

etc.

2.2 Z-transform theorems

Linearity: If f(t) and g(t) are Z-transformable and and are scalar, then the linear combination f(t)+g(t) has the Z-transform

Z[f(t)+g(t)]= F(z)+ G(z)

2.2 Z-transform theorems

Shifting Theorem:

Given that the Z-transform of f(t) is F(z), find the Z-transform for f(t-nT).

f(t)

t

f(t-nT)

tnT

2.2 Z-transform theorems

If f(t)=0 for t<0 has the Z-transform F(z), then

Proving: By Z-transform definition, we have

1

0

)()()]([

and )()]([n

k

kn

n

zkTfzFznTtfZ

zFznTtfZ

0

)(

0

)(

00

)()(

)()()]([

k

nkn

k

nnk

k

nnk

k

k

znTkTfzzznTkTf

zznTkTfznTkTfnTtfZ

2.2 Z-transform theorems

Defining m=k-n, we have

Since f(mT)=0 for m<0, we can rewrite the above as

Thus, if a function f(t) is delayed by nT, its Z-transform would be multiplied by z-n. Or, multiplication of a Z-transform by z-n has the effect of moving the function to the right by nT time. This is the so-called Shifting Theorem.

nm

mn

k

nkn zmTfzznTkTfznTtfZ )()()]([0

)(

)()()()]([0

zFzzmTfzzmTfznTtfZ n

m

mn

nm

mn

2.2 Z-transform theorems

Final value theorem:Suppose that f(t), where f(t)=0 for t<0, has the Z-transform of F(z), then the final value of f(t) can be given by

There are other theorems for Z-transform. Please read the study book or textbook for more details.

)()1(lim)(lim 1

1zFztf

zt

0

)()]([)(k

kzkTftfZzF

)()()]()([ 22112211 zFkzFktfktfkZ

)()]([ aTat zeFtfeZ

)()]([a

zFtfaZ t

)()]([ zFzkTtfZ k

])()([)]([1

0

n

k

kk zkTfzFzkTtfZ

)()1()]1()([ 1 zFztftfZ

10 1

)()(

z

zFdfZ

t

)()1(lim)( 1

1zFzf

z

)(lim)0( zFfz

Theorem Name

Definition

Linearity

Multiply by e-at

Multiply by at

Time Shift 1

Time Shift 2

Differentiation

Integration

Final Value

Initial Value

s

111

1 z

2

1

s 21

1

)1(

z

Tz

as 1

11

1 ze aT

)( ass

a

)1)(1(

)1(11

1

zez

zeaT

aT

ab

ee btat

))((

1

bsas

)1)(1(

)(111

1

zeze

zee

ab bTaT

bTaT

22 s

21

1

cos21

sin

zTz

Tz

22 s

s21

1

cos21

cos1

zTz

Tz

22)(

as 221

1

cos21

sin

zeTez

TezaTaT

aT

22)(

as

as221

1

cos21

cos1

zeTez

TezaTaT

aT

f(t) F(s) F(z)

(t) 1 1

u(t)

t

e-at

1 – e-at

sint

cost

e-atsint

e-atcost

2.3 Z-transform examples

Example 1: Assume that f(k)=0 for k<0, find the Z-transform of f(k)=9k(2k-1)-2k+3, k=0,1,2….

Solution: Obvious f(k) is a combination of three sub-function 9k(2k-1), 2k and 3. Therefore, first we can apply linearity theorem to f(k). Second, sub-function 9k(2k-1) can be considered as a product of k and 2-12k, then we can apply the theorem of multiply by ak. Finally, we can find the answer by combining these three together.

2.3 Z-transform examples

)1()21(

2

1

3

21

1

)21(

9]3[]2[)]2(9[)(

)21(

9

1)-(z/2

)2/(

2

9]2[

2

9]229[)]2(9[

21

1

12/

2/]12[]2[

)1()1(

Tz Z[t];

1z-1

1 Z[1];)]([

]3[]2[)]2(9[]32)2(9[)(

121

2

1121

11

21

1

211

1

221

1-

1-

11

zz

z

zzz

zZZkZzF

z

zzkZkZkZ

zz

zZZ

z

Tz

zz

z

a

zFtfaZ

ZZkZkZzF

kk

kkk

kk

t

kkkk

2.3 Z-transform examples

Example 2: Obtain the Z-transform of the curve x(t) shown below.

0 1 2 3 4 5 6 7 8

1

t

f(t)

2.3 Z-transform examples

Solution: From the figure, we have

K 0 1 2 3 4 5 6…

f(k) 0 0 0 1/3 2/3 1 1…

Apply the definition of Z-transform, we have

)1(313

2z

)1(3

2z

3

2

3000)()(

1

543

1

543-

21543-

6543

0

z

zzz

z

zz

zzzz

zzzz

zkfzFk

k

2.3 Z-transform examples

Example 3: Find the Z-transform of

Solution: Apply partial fraction to make F(s) as a sum of simpler terms.

)()(

2

2

ass

asF

)1()1(

])1()1[(

1

1

1

1

)1(

]1

[]1

[][)]([)(

11

)()(

121

11

1121

1

2

232

21

2

2

zez

zzaTeeeaT

zezz

aTz

asZ

sZ

s

aZsFZzF

asss

a

as

k

s

k

s

k

ass

asF

aT

aTaTaT

aT

2.4 Inverse Z-transform

The inverse Z-transform: When F(z), the Z-transform of f(kT) or f(t), is given, the operation that determines the corresponding time sequence f(kT) is called as the Inverse Z-transform. We label inverse Z-transform as Z-1.

nno

mmo

nno

mmo

zazaza

zbzbzbbZzFZkTf

zazaza

zbzbzbbtfZkTfZzF

21

1

22

1111

21

1

22

11

1)]([)(

1)]([)]([)(

2.4 Inverse Z-transform

Z-transform

=

Inverse Z-transform

2.4 Inverse Z-transform

The inverse Z-transform can yield the corresponding time sequence f(kt) uniquely. However, it says nothing about f(t). There might be numerous f(t) for a given f(kT).

f(t)

t0 T 2T 3T 4T 5T 6T

2.4 Inverse Z-transform

x(kT) f(t)Zero-orderHold

Low-passFilter

2.5 Methods for Inverse Z-transform

How can we find the time sequence for a given Z-transform?

1) Z-transform table

Example 1: F(z)=1/(1-z-1), find f(kT).

F(z)=1+z-1+z-2+z-3+…

f(kT)=Z-1[F(z)]=1, for k=0, 1, 2, …

2.5 Inverse Z-transform examples

Example 2: Given ,

Find f(kT).

Solution: Apply partial-fraction-expansion to simplify F(z), then find the simpler terms from the Z-transform table.

Then we need to determine k1 and k2

)1)(1(

)1()(

11

1

zez

zezF

aT

aT

12

11

11

1

11)1)(1(

)1()(

ze

k

z

k

zez

zezF

aTaT

aT

2.5 Inverse Z-transform examples

Multiply (1-z-1) to both side and let z-1=1, we have

12

11

11

1

11)1)(1(

)1()(

ze

k

z

k

zez

zezF

aTaT

aT

11

)1(

1

)1(

1

)1(

11)1(

)1)(1(

)1()1(

1

1

1

112

1

11

1

12

111

11

11

1

z

aT

aT

aTaT

aT

aTaT

aT

ze

zek

ze

kzk

ze

ze

ze

k

z

kz

zez

zez

2.5 Inverse Z-transform examples

Similar as the above, we let multiply (1-e-aTz-1) to both side and let z-1 =eaT, we have

Finally, we have

11

)1(

1

)1(

1

)1(

11)1(

)1)(1(

)1()1(

11

1

2211

1

1

1

12

111

11

11

aTez

aTaTaT

aTaT

aT

aTaT

z

zekk

z

kze

z

ze

ze

k

z

kze

zez

zeze

0,1,2,k ,1)(

1

1

1

1

)1)(1(

)1()(

1111

1

akT

aTaT

aT

ekTf

zezzez

zezF

2.5 Inverse Z-transform examples

Exercise 4: Given the Z-transform

Determine the initial and final values of f(kT), the inverse Z-transform of F(z), in a closed form.

Hint: Partial-fraction-expansion, then use Z-transform table, and finally applying initial & final value theorems of Z-transform.

)4.03.11)(1()(

211

1

zzz

zzF

2.5 Inverse Z-transform examples

2) Direct division method

Example 1: F(z)=1/(1+z-1), find f(kT).1

1

11

1

1

1

-z

z

z

1

2

21

1

1

11

1

11

z

z

zz

-z

z

z

-

21

2

21

1

1

11

1

11zz

z

zz

-z

z

z

-

2.5 Inverse Z-transform examples

Finally, we obtain: F(z)=1-z-1+z-2-z-3+…

K = 0 1 2 3 …

F(kT)= 1 -1 1 -1

Example 2: Given ,

Find f(kT).

Solution: Dividing the numerator by the denominator, we obtain

21

1

)1(

21)(

z

zzF

21

54

543

43

432

32

321

21

21

121

741

1013

102010

710

7147

47

484

4

21

2121zz

zz

zzz

zz

zzz

zz

zzz

zz

zz

zzz

2.5 Inverse Z-transform examples

Finally, we obtain: F(z)=1+ 4z-1 + 7z-2 + 10z-3+…

K = 0 1 2 3 …

F(kT)= 1 4 7 10…

Exercise 5: ,

Find f(kT).Ans. :k 0 1 2 3 4 5…

f(kT) 0 0.3679 0.8463 1 1 1…

21

4321

3679.03679.11

05659.002221.0343.03679.0)(

zz

zzzzzF

2.5 Inverse Z-transform examples

3) Computational method using Matlab

Example: Given find f(kT).

Solution:

num=[1 2 0]; den=[1 –2 1]

Say we want the value of f(kT) for k=0 to 30

u=[1 zeros(1,30)]; F=filter(num, den, u)

1 4 7 10 13 16 19 22 25 28 31…

21

1

)1(

21)(

z

zzF

12

2

)1(

2

)1(

21)(

2

2

2

2

21

1

zz

zz

z

zz

z

zzF

2.5 Inverse Z-transform examples

Exercise 6: Given the Z-transform

Use 1) the partial-fraction-expansion method and 2) the Matlab to find the inverse Z-transform of F(z).

Answer: x(k)=-8.3333(0.5)k+8.333(0.8)k-2k(0.8)k-1

x(k)=0;0.5;0.05;–0.615;–1.2035;-1.6257;-1.8778…

211

11

)8.01)(5.01(

)5.0()(

zz

zzzF

Reading

Study book

• Module 2: The Z-transform and theorems

Textbook

• Chapter 2 : The Z-transform (pp23-50)

TutorialExercise: The frequency spectrum of a continuous-

time signal is shown below.

1) What is the minimum sampling frequency for this signal to be sampled without aliasing.

2) If the above process were to be sampled at 10 Krad/s, sketch the resulting spectrum from –20 Krad/s to 20 Krad/s.

-8 -4 4 8

Krad/s

F()

TutorialSolution: 1) From the spectrum, we can see that the

bandwidth of the continuous signal is 8 Krad/s. The Sampling Theorem says that the sampling frequency must be at least twice the highest frequency component of the signal. Therefore, the minimum sampling frequency for this signal is 2*8=16 Krad/s.

-8 -4 4 8

Krad/s

F()

Tutorial2) Spectrum of the sampled signal is formed by

shifting up and down the spectrum of the original signal along the frequency axis at i times of sampling frequency. As s=10 Krad/s, for i =0, we have the figure in bold line. For i=1, we have the figure in bold-dot line.

4 8

Krad/s

F()

122 6 14 181610-4-8

TutorialFor I=-1, 2,… we have

4 8

Krad/s

122 6 14 181610

-18

F()

4 8

Krad/s

122 6 14 181610-2-4-6-8-14

Tutorial

Exercise 1: Find the Z-transform of a exponential decay f(t)=e-aT using other method.

f(t)

t

Tutorial

11

33221

22111

221

0

1

1)(1)()(

)1()(

1

)()]([)]([)(

zezFzFzezF

zezeze

zezezezFze

zeze

kTfkTfZtfZzF

aTaT

aTaTaT

aTaTaTaT

aTaT

k

Tutorial

Exercise 2: Find the Z-transform for a decayed cosine function

Solution 1:

tetf at cos)(

221

1

21

1

21

1

cos21

cos1

cos21

cos1]cos[

)()]([ );()]([

)(cos21

cos1cos

zeTze

Tze

zTz

TzteZ

zeFtfeFzFtfZ

zFzTz

TztZ

aTaT

aT

zez

at

aTat

aT

Tutorial

Solution 2:

221

1

11

1

cos21

cos1

1

1

1

1

2

1)(

1

1][

])[][(2

1

2]cos[)(

zeTze

Tze

zezezF

zeeZ

eZeZ

eeZteZzF

aTaT

aT

TjaTTjaT

aTat

tjattjat

tjattjatat

Tutorial

Exercise 3: Find the Z-transform for

Solution:

attetf )(

21

1

21

1

21

1

)1(

)1(][

)()]([ );()]([

)()1(

ze

zTe

z

TzteZ

zeFtfeFzFtfZ

zFz

TztZ

aT

aT

zez

at

aTat

aT

Tutorial

Exercise 4: Given the Z-transform

Determine the initial and final values of f(kT), the inverse Z-transform of F(z), in a closed form.

Solution: Apply the initial value theorem and the final value theorem respectively, we have

)4.03.11)(1()(

211

1

zzz

zzF

Tutorial

))8.0(4)5.0(31(27

1)(

8.01

48.1

5.01

11.1

1

37.0

8.015.011

)5.01)(8.01)(1()4.03.11)(1()(

7.2

1

)4.03.11)(1(

)1()]()1[()(

0)4.03.11)(1(

)()0(

11113

12

11

111

1

211

1

211

111

1

211

1

limlim

limlim

kk

zz

zz

kf

zzzz

k

z

k

z

k

zzz

z

zzz

zzF

zzz

zzzFzf

zzz

zzFf

Tutorial

Exercise 5: Given

Find f(kT) using direct-division method.

Solution:

21

4321

3679.03679.11

05659.002221.0343.03679.0)(

zz

zzzzzF

1

432

321

432121

3679.0

0565.01576.00.8463

1354.05033.03679.0

0565.002221.0343.03679.03679.03679.11

z

zzz

zzz

zzzzzz

Tutorial

Continuous

4321

4321

54

543

43

432

432

321

432121

8463.03679.0)(

8463.03679.0

3679.0

3679.03679.1

3679.0

3114.01576.18463.0

0565.01576.00.8463

1354.05033.03679.0

0565.002221.0343.03679.03679.03679.11

zzzzkf

zzzz

zz

zzz

zz

zzz

zzz

zzz

zzzzzz

Tutorial

Exercise 6: Given the Z-transform

Use 1) the partial-fraction-expansion method and 2) the Matlab to find the inverse Z-transform of F(z).

Solution1: To make the expanded terms more recognizable in the Z-transform table, we usually expand F(z)/z into partial fractions.

211

11

)8.01)(5.01(

)5.0()(

zz

zzzF

2)5.0(

15.0

5.0

)8.0(

5.0

)8.0(

5.0

15.0

8.0)8.0()5.0()8.0(

)8.0)(5.0(

15.0)8.0(

333.8)8.0(

15.00.5zlet

8.0

)5.0(

)8.0(

)5.0(

)8.0(

15.0

8.0)8.0()5.0()5.0(

)8.0)(5.0(

15.0)5.0(

8.0)8.0()5.0()8.0)(5.0(

15.0)(

)8.0)(5.0(

)15.0(

)8.01)(5.01(

)5.0()(

8.0

23

2

21

32

2122

2

5.021

32

212

32

212

32

212

2211

11

z

z

z

zk

z

zkk

z

zk

z

z

z

k

z

k

z

kz

zz

zz

z

zk

z

zk

z

zkk

z

z

z

k

z

k

z

kz

zz

zz

z

k

z

k

z

k

zz

z

z

zF

zz

zz

zz

zzzF

kkk

zz

kf

zz

z

zzF

zzzzz

z

z

zF

z

zz

z

zk

zz

zkzk

z

zk

z

z

derivativez

zkk

z

zk

z

z

z

k

z

k

z

kz

zz

zz

)8.0(333.8)8.0(2)5.0(333.8)(

8.01

333.8

)8.01(

2

)5.01(

333.8)(

8.0

333.8

)8.0(

2

)5.0(

333.8

)8.0)(5.0(

15.0)(

333.83.0

6.03.0*5.0

)5.0(

)15.0()5.0(5.0

5.0

15.0

8.0let ;)5.0(

)8.0()5.0(0

5.0

)8.0(

5.0

15.0

5.0

)8.0(

5.0

)8.0(

5.0

15.0

8.0)8.0()5.0()8.0(

)8.0)(5.0(

15.0)8.0(

1

121

1

1

22

28.0

2

8.0

'

3

33

'21

'

32

21

32

2122

2

Tutorial

Partial fraction for inverse Z-transform

If F(z)/z involve s a multiple pole, eg. P1, thenipz

iin

n

n

mmmm

nnnn

mmmm

z

zFpza

pz

a

pz

a

pz

a

pzpzpz

bzbzbzb

azazaz

bzbzbzb

z

zF

)()( ;

)())((

)(

2

2

1

1

21

11

10

11

0

11

10

11

)()(;

)()( ;

)()(

)()()()(

)(

212

211

3

3

1

22

1

1

32

1

11

10

pzpzpzii

n

n

n

mmmm

z

zFpz

dz

dc

z

zFpzc

z

zFpza

pz

a

pz

a

pz

c

pz

c

pzpzpz

bzbzbzb

z

zF

i

Tutorial

Solution 2: Expand F(z) into a polynomial form

Num=[0 0.5 –1 0];

Den=[1 –2.1 1.44 –0.32];

U==[1 zeros(1,40)];

F=filter(Num, den,U)

0 0.5 0.05 -0.615 -1.2035…

321

21

211

11

32.044.11.21

5.0

)8.01)(5.01(

)5.0()(

zzz

zz

zz

zzzF