Vous êtes sur la page 1sur 11

Outline

Time domain solution for zero-input


response.
Properties of matrix exponential.
Complete solution using integrating factor.
Laplace transform solution of state
equations.
Transfer function.

Solution of State Equations


M. S. Fadali
Professor EE
University of Nevada, Reno
1

LTI Systems

Time Domain Approach

State-space model

Zero-input (homogeneous) state equation


Differentiate and substitute

Solve differential equation.


Homogeneous equation: time domain series
solution + integrating factor for the
complete solution.
Laplace transform solution.
3

Example: Diagonal State Matrix

Taylor Series (t = 0)

Properties of e At

Numerical Example

Identity at

Determine the matrix exponential for


the matrix

Proof

Product
Proof
0

Solution

Properties of

(Cont.)

Derivative of

Inverse for any is


Proof

Proof:
Transpose
Proof:

10

Proof:

Commutative Multiplication

Assume

11

then

12

Proof:

Complete Solution
Multiply by an integrating factor
Integrate both sides from

Subtract

to then

from both sides




13

14

Impulse Response

Output Response

Response due to

Zero-state (output) response (zero ICs)

Impulse response matrix


15

cols of

(resp.)
16

Laplace Transform

Solve for X(s)

State equation
Use Laplace transform of derivative
_
_

Resolvent Matrix:
We need to inverse Laplace transform the
resolvent matrix to obtain the solution.

17

Matrix Exponential
(state-transition matrix)
1
1

18

Complete Solution
Inverse Laplace transform (use convolution)

1
1

Solution for non-zero initial time

Substitute in the output equation

By analogy with scalar exponential


_
19

20

Example: Matrix Exponential

Partial Fractions

is in companion form

Partial Fractions

21

Partial Fractions: Matrix Coeffts.

22

Example: Zero-input Response

Inverse Laplace transform


23

24

Exercise

Example: Zero-state Response

Find the matrix exponential and the response to


1,

You can show that:

3,

4 3
12 9
8 6

2
6 ,
4

4
8
4

2 1
4 2 ,
2 1
1
5,
4
4

25

Linearity

1 1
4 4
4 4
26

Transfer Function

Zero-input linearity

Use formula if given

only.

Transform of solution with zero initial conditions


Zero-state linearity

Additivity
27

28

Example: Zero-state Response

Example: Transfer Function


0
8

1
6
1 1

0
1

Obtain output from transfer function

Can obtain transfer function by inspection (special case)

1 1
0.5
2

2 0.5
4
1
2
1.5
4

1
4

0.5
2
4

0
1

1
6

8
29

30

MATLAB

Resolvent Matrix

>> A=[0,1;-8,-6];B=[0;1];C=[1,1];
>> poly(A) % characteristic polynomial
ans =
1 6 8
>> p = ss(A, B, C, 0);
>> g= tf(p) % can also use g=zpk(p)
s+1
------------s^2 + 6 s + 8

>> n=size(A);
>> resolvent=tf(ss(A, eye(n), eye(n), 0))
Justification:
For

31

32

Impulse Response Matrix

MATLAB Simulation

>> p = ss(A, B, C, 0); step(p) % (output) step response


>> step(ss(A,B,eye(n),0)) %Make C=eye(n) to plot x

Example:

Step Response

Amplitude

To: Out(1)

0.1

0.05

To: Out(2)

0.1

0.05

0.5

1.5
2
Time (seconds)

2.5

3.5

33

34

Example

Modes
n 1
distinct i
s Z i
i
i 1
1
sI n A nd ni
1

Z n distinct, multiplicity ni

i 1 j 1 s i j ij d

Repeated eigenvalue

n
it
distinct i
Z i e
i 1
At
e n n 1
d
i
t j it

Z
e nd distinct, multiplicity ni
i j 1

j!
i 1 j 0

System Modes:

or

(some books

)
35

36

Modal Vectors

Example: Modal Vectors

Eigenvectors of :

is an eigenvector of

with eigenvalues

State stays on eigenvector is it is the initial state.

Similarly, for the second modal vector as


response involves the second mode only.

37

For a real matrix , if


is an eigenpair,
then is an eigenpair.
denotes the complex conjugate
eigenpair:
Complex conjugate:

38

MATLAB Example

Complex Conjugate Eigenpairs

the

>> a=[0 1;-10,-2]


a=
0 1
-10 -2
>> [v,d]=eig(a)
v=
-0.0953 - 0.2860i -0.0953 + 0.2860i
0.9535
0.9535
d=
-1.0000 + 3.0000i
0
0
-1.0000 - 3.0000i


39

40

Complex Modal Vectors


If

is in

References
1. P. Belanger, Control Engineering, Saunders, Fort
Worth, 1995.
2. N. S. Nise, Control Systems Engineering, Wiley,
Hoboken, NJ, 2011.
3. R. L. Williams & D. A. Lawrence, Linear StateSpace Control Systems, J. Wiley, Hoboken, NJ, 2011.

plane then so is

41

42

Vous aimerez peut-être aussi