Вы находитесь на странице: 1из 34

Lecture 12: Dynamic Solution in State Space

Lecture 12: Dynamic Solution in State Space


MAE3406/TRC3600 Aerospace Control/Modeling and Control

Chao Chen Outline Motivating Examples Solving Homogeneous State Equations Solving Nonhomogeneous State Equations Stabilization

Dr Chao Chen
Monash University

2010

Conclusion

Motivating Problems

Lecture 12: Dynamic Solution in State Space Chao Chen Outline Motivating Examples Solving Homogeneous State Equations Solving Nonhomogeneous State Equations Stabilization Conclusion

What have we learned?

Lecture 12: Dynamic Solution in State Space Chao Chen Outline Motivating Examples Solving Homogeneous State Equations Solving Nonhomogeneous State Equations Stabilization Conclusion

Laplace transform. Model a system in frequency domain, by means of Laplace transform. Model a system in time domain (state space).

Objectives

Lecture 12: Dynamic Solution in State Space Chao Chen Outline Motivating Examples Solving Homogeneous State Equations

Solving homogeneous state space equations. Solving nonhomogeneous state space equations. Introduction of pole-placement method.

Solving Nonhomogeneous State Equations Stabilization Conclusion

Example: Inverted Pendulum

Lecture 12: Dynamic Solution in State Space Chao Chen Outline

Motivating Examples Solving Homogeneous State Equations Solving Nonhomogeneous State Equations Stabilization

State variables: x1 = y, x2 = , x3 = y, x4 = These four equations can be put in state-space form (u = f ): x = Ax + Bu, or
x1 0 x2 0 x3 = 0 x4 0 0 0 mg/M (M + m)g/Ml 1 0 0 0 0 0 x1 1 x2 0 + u 0 x3 1/M 0 x4 1/Ml

Conclusion

Example: Inverted Pendulum

Lecture 12: Dynamic Solution in State Space Chao Chen Outline

Motivating Examples Solving Homogeneous State Equations Solving Nonhomogeneous State Equations

Assume M = 1kg, m = 1kg, g = 10m2 /s, and l = 1m. The state-space equations become into
0 x1 x2 0 x3 = 0 0 x4 0 0 10 20 1 0 0 0 0 x1 0 1 x2 0 + u 0 x3 1 0 x4 1

Stabilization Conclusion

We are interested in the behaviour of this inverted pendulum with certain inputs and initial conditions.

Example: Inverted Pendulum

Lecture 12: Dynamic Solution in State Space Chao Chen Outline

Motivating Examples Solving Homogeneous State Equations Solving Nonhomogeneous State Equations

The state-space equations become into


x1 0 x2 0 x3 = 0 x4 0 0 0 10 20 1 0 0 0 0 x1 0 1 x2 0 + u 0 x3 1 0 x4 1

Stabilization Conclusion

A trivial example is no input with zero initial conditions, i.e., x1 = x2 = x3 = x4 = 0

Example: Inverted Pendulum


l

Lecture 12: Dynamic Solution in State Space Chao Chen Outline Motivating Examples Solving Homogeneous State Equations

The state-space equations become into


x1 0 x2 0 x3 = 0 x4 0 0 0 10 20 1 0 0 0 0 x1 0 1 x2 0 + u 0 x3 1 0 x4 1

Solving Nonhomogeneous State Equations Stabilization Conclusion

How about a tiny deviation in y, i.e., x1 = 0.01, x2 = x3 = x4 = 0

How about a tiny deviation in , i.e., x2 = 0.01, x1 = x3 = x4 = 0

Homogeneous and Time-Invariant System

Lecture 12: Dynamic Solution in State Space Chao Chen Outline Motivating Examples Solving Homogeneous State Equations

x = Ax no forcing, no changing of system with respect to time.

Solving Nonhomogeneous State Equations Stabilization Conclusion

Homogeneous Differential Equation


An ordinary differential equation is given by x = ax The solution is given by x(t) = eat x(0) A vector-matrix differential equation x = Ax It can be proved that the solution is given by x(t) = eAt x(0) where eAt Ak tk k=0 k!

Lecture 12: Dynamic Solution in State Space Chao Chen Outline Motivating Examples Solving Homogeneous State Equations Solving Nonhomogeneous State Equations Stabilization Conclusion

10

State Transition Matrix


x(t) = e x(0) The state transition matrix (STM) is dened as (t, t0 ) e If t0 = 0, then (t, t0 ) = (t) eAt Properties of the STM (t)
1. 2. 3. 4. 5. (t, t) = I (No change in time = no state change) 1 (t) = (t) (Reverse time = reverse state change) (t1 + t2 ) = (t1 ) (t2 ) [ (t)]n = (nt) (t2 t1 ) (t1 t0 ) = (t2 t0 ) = (t1 t0 ) (t2 t1 ) (Reversible)
A(tt0 ) At

Lecture 12: Dynamic Solution in State Space Chao Chen Outline Motivating Examples Solving Homogeneous State Equations Solving Nonhomogeneous State Equations Stabilization Conclusion

11

Computing State Transition Matrix


(t) e
At

Lecture 12: Dynamic Solution in State Space Chao Chen Outline Motivating Examples Solving Homogeneous State Equations Solving Nonhomogeneous State Equations Stabilization Conclusion

Ak t k k=0 k!

Cayley-Hamilton Theorem: Consider n n matrix A and its characteristic equation given by | I A| = n + a1 n1 + . . . + an1 + an = 0 Then, An + a1 An1 + . . . + an1 A + an I = 0 Theorem eAt = 0 I + 1 A + . . . + n1 An1 ei t = 0 + 1 i + . . . + n1 in1 where i are the eigenvalues of the matrix A.

12

Computing State Transition Matrix


x(t) = e x(0) = (t)x(0) Step 1: Find all eigenvalues of A. Step 2: write down a set of equations with eigenvalues,
e 1 t e
2 t

Lecture 12: Dynamic Solution in State Space Chao Chen

At

Outline Motivating Examples Solving Homogeneous State Equations Solving Nonhomogeneous State Equations Stabilization Conclusion

= = = =

n1 0 + 1 1 + . . . + n1 1 n1 0 + 1 2 + . . . + n1 2

. . . e n t

. . .
n1 0 + 1 n + . . . + n1 n

Step 3: solve these equations for i . Step 4: compute the state transition matrix by eAt = 0 I + 1 A + . . . + n1 An1

13

Example

Lecture 12: Dynamic Solution in State Space Chao Chen Outline


l

Motivating Examples Solving Homogeneous State Equations Solving Nonhomogeneous State Equations Stabilization Conclusion

x1 0 x2 0 x3 = 0 x4 0

0 0 10 20

1 0 0 0

0 x1 1 x2 , 0 x3 0 x4

x1 (0) 0 x2 (0) 0.01 x3 (0) = 0 x4 (0) 0

x(t) = eAt x(0) Step 1: nd all eigenvalues, 1 = 0, 2 = 0, 3 = 4.47, 4 = 4.47

14

Example
Step 2: write down a set of equations with eigenvalues,
e1 t e2 t e3 t e4 t with 1 = 0, 2 = 0, = = = =
3 0 + 1 1 + . . . + 3 1 3 0 + 1 2 + . . . + 3 2 3 0 + 1 3 + . . . + 3 3 3 0 + 1 4 + . . . + 3 4

Lecture 12: Dynamic Solution in State Space Chao Chen Outline Motivating Examples Solving Homogeneous State Equations Solving Nonhomogeneous State Equations Stabilization Conclusion

3 = 4.47,

4 = 4.47.

Here we encounter a special case with two identical eigenvalues. In order to deal with it, we need to build another equation, which can be done by
de t d0 + 1 + . . . + n1 n1 = d d te t = 1 + 22 + . . . + (n 1)n1 n2

This is a general rule for identical eigenvalues. Times of differentiation is equal to the repeat number.

15

Example
Therefore, we setup a set of equations as follows,
e 1 t te2 t e e with 1 = 0,
3 t 4 t

Lecture 12: Dynamic Solution in State Space Chao Chen Outline Motivating Examples Solving Homogeneous State Equations Solving Nonhomogeneous State Equations Stabilization Conclusion

= = = =

3 0 + 1 1 + . . . + 3 1

1 + 22 2 + 33 2
3 0 + 1 3 + . . . + 3 3 3 0 + 1 4 + . . . + 3 4

2 = 0,

3 = 4.47,

4 = 4.47.

Step 3: solving these equations yields


0 = 1 1 = t 2 = 0.05 + 0.025e4.47t + 0.025e4.47t 3 = 0.05t 0.005e4.47t + 0.005e4.47t

Step 4: compute the state transition matrix by (t) = eAt = 0 I + 1 A + . . . + n1 An1

16

Example

Lecture 12: Dynamic Solution in State Space Chao Chen Outline

Step 5: write down the solutions as x(t) = (t)x(0) where x(0) = 0 0.01 0 0 . The solutions are
x1 x2 x3 x4 = = = = 0.005 0.0025e4.47t 0.0025e4.47t 0.005e4.47t + 0.005e4.47t 0.011e4.47t 0.011e4.47t 0.022e4.47t + 0.022e4.47t
T

Motivating Examples Solving Homogeneous State Equations Solving Nonhomogeneous State Equations Stabilization Conclusion

17

Example

Lecture 12: Dynamic Solution in State Space Chao Chen Outline


l

Motivating Examples Solving Homogeneous State Equations Solving Nonhomogeneous State Equations Stabilization Conclusion

18

Example
l

Lecture 12: Dynamic Solution in State Space Chao Chen Outline Motivating Examples Solving Homogeneous State Equations Solving Nonhomogeneous State Equations Stabilization Conclusion

Question: how can we stabilize this system? Using input force f . How? Lets choose some inputs for test.
x1 0 x2 0 x3 = 0 x4 0 with x(0) = 0 0.01 0 0 0 10 20 0 .
T

1 0 0 0

0 x1 0 1 x2 0 + u 0 x3 1 0 x4 1

19

Inhomogeneous and Time-Invariant System


x = Ax + Bu

Lecture 12: Dynamic Solution in State Space Chao Chen Outline

Rewrite the state equations as x Ax = Bu Multiply both sides by eAt : eAt ( Ax) = eAt Bu(t) x Considering a function of eAt x d At e x = eAt Bu(t) dt Time integration from t0 to t: eAt x(t) eAt0 x(t0 ) =
t t0

Motivating Examples Solving Homogeneous State Equations Solving Nonhomogeneous State Equations Stabilization Conclusion

eA Bu()d

20

Inhomogeneous and Time-Invariant System


x = Ax + Bu Rearranging: x(t) = eA(tt0 ) x(t0 ) + So x(t) = (t, t0 )x(t0 ) + In case of t0 = 0, we have
t t t0 t t0

Lecture 12: Dynamic Solution in State Space Chao Chen Outline Motivating Examples Solving Homogeneous State Equations

eA(t) Bu()d

Solving Nonhomogeneous State Equations Stabilization Conclusion

eA(t) Bu()d

x(t) = (t)x(0) +
0

eA(t) Bu()d

21

Example

Lecture 12: Dynamic Solution in State Space Chao Chen Outline


l

Motivating Examples Solving Homogeneous State Equations Solving Nonhomogeneous State Equations

x1 0 x2 0 = x3 0 x4 0 with x(0) = 0 0.01 0

0 0 10 20 0 .
T

1 0 0 0

0 x1 0 1 x2 0 + u 0 x3 1 0 x4 1

Stabilization Conclusion

Given an input u = f = 2N, nd out the state response of the system.

22

Example

Lecture 12: Dynamic Solution in State Space Chao Chen Outline

The solution is
t

Motivating Examples

x(t) = (t)x(0) +
0

eA(t) Bu()d

Solving Homogeneous State Equations Solving Nonhomogeneous State Equations

where
0 0 A= 0 0 0 0 10 20 1 0 0 0 0 1 , 0 0 0 0 B = , 1 1 0 0.01 x(0) = 0 , 0 u=2

Stabilization Conclusion

Further (t) has been computed.

23

Example
The solutions are given by
x1 (t) = 0.005 0.0025e4.47t 0.0025e4.47t +0.025 x2 (t) = 20t2 e2(2.23t) 2e2(2.23t) + 1 + e4(2.23t) e2(2.23t) 2e2(2.23t) + 1 + e4(2.23t) 0.005e4.47t + 0.005e4.47t 0.05 e2(2.23t)

Lecture 12: Dynamic Solution in State Space Chao Chen Outline Motivating Examples Solving Homogeneous State Equations Solving Nonhomogeneous State Equations Stabilization Conclusion

24

Example
How about another input f = u = 0.95 sin(t)N. The solution is
t

Lecture 12: Dynamic Solution in State Space Chao Chen Outline Motivating Examples

x(t) = (t)x(0) +
0

eA(t) Bu()d

Solving Homogeneous State Equations Solving Nonhomogeneous State Equations Stabilization Conclusion

25

Example
l

Lecture 12: Dynamic Solution in State Space Chao Chen Outline Motivating Examples Solving Homogeneous State Equations Solving Nonhomogeneous State Equations Stabilization Conclusion

x1 0 x2 0 x3 = 0 x4 0 with x(0) = 0 0.01 0

0 0 10 20 0 .
T

1 0 0 0

0 x1 0 1 x2 0 + u 0 x3 1 0 x4 1

Given an input of u = f = 2N, the balance is maintained for about 0.3s Given an input of u = f = 0.95 sin(t)N, the balance is maintained for about 1.2s. Is there an input to stabilize the system constantly?

26

State Feedback
x = Ax + Bu

Lecture 12: Dynamic Solution in State Space Chao Chen Outline

We can choose the control signal to be u = Kx The original system is now converted to a homogeneous system (a CL system): x = (A BK)x = A2 x The characteristic equation of the CL system is given by |sI A2 | = 0 If all the CL poles are in the left-half s plane, the system is stable. We can choose K properly to enforce the CL poles in the left. How?

Motivating Examples Solving Homogeneous State Equations Solving Nonhomogeneous State Equations Stabilization Conclusion

27

Example
l

Lecture 12: Dynamic Solution in State Space Chao Chen Outline Motivating Examples Solving Homogeneous State Equations

0 x1 x2 0 x3 = 0 0 x4

0 0 10 20

1 0 0 0

0 x1 0 1 x2 0 + u, 0 x3 1 0 x4 1

0 0.01 x(0) = 0 0

Solving Nonhomogeneous State Equations Stabilization Conclusion

Assume u = Kx = (K1 x1 + K2 x2 + K3 x3 + K4 x4 ) The CL system is hence given by


x1 0 x2 0 x3 = K1 x4 K1 0 0 10 K2 20 + K2 1 0 K3 K3 0 x1 1 x2 K4 x3 x4 K4

28

Example
The CL characteristic equation is then given by
|sI A2 | = s4 + (K3 K4 )s3 + (K1 K2 20)s2 10K3 s 10K1 = 0

Lecture 12: Dynamic Solution in State Space Chao Chen Outline Motivating Examples Solving Homogeneous State Equations Solving Nonhomogeneous State Equations Stabilization Conclusion

We can create a CL characteristic equation with all poles on the left-half s plane, i.e.,
(s + 3)(s + 4)(s + 5)(s + 6) = s + 18s + 119s + 342s + 360 = 0
4 3 2

Comparing these two characteristic equations yields a set of equations on Ki ,


K4 + K3 18 = 0 139 K2 + K1 = 0 10K3 342 = 0 10K1 360 = 0

Solving the above equations yields


K1 = 36, K2 = 175, K3 = 171/5, K4 = 261/5

29

Example

Lecture 12: Dynamic Solution in State Space Chao Chen Outline

The state-feedback system is given by


x1 0 x2 0 = x3 36 x4 36 0 0 165 155 1 0 171/5 171/5 0 x1 1 x2 , 261/5 x3 261/5 x4 0 0.01 x(0) = 0 0

Motivating Examples Solving Homogeneous State Equations Solving Nonhomogeneous State Equations Stabilization

Solving it yields
x1 (t) = 1.38e4t t 0.415e3t 0.13e5t + 0.55e6t x2 (t) = 3.68e4t t + 1.63e3t 1.15e5t 0.46e6t

Conclusion

30

Example

Lecture 12: Dynamic Solution in State Space Chao Chen Outline


l

Motivating Examples Solving Homogeneous State Equations Solving Nonhomogeneous State Equations Stabilization Conclusion

31

Motivating Problems

Lecture 12: Dynamic Solution in State Space Chao Chen Outline Motivating Examples Solving Homogeneous State Equations Solving Nonhomogeneous State Equations Stabilization Conclusion

32

Summary

Lecture 12: Dynamic Solution in State Space Chao Chen Outline Motivating Examples Solving Homogeneous State Equations

Solving homogeneous state space equations. Solving nonhomogeneous state space equations. Introduction of pole-placement method.

Solving Nonhomogeneous State Equations Stabilization Conclusion

33

What is Next?

Lecture 12: Dynamic Solution in State Space Chao Chen Outline Motivating Examples

We will not discuss the state-space technology further in this unit. More advanced topics and systematic studies in state space will be in MEC4418, Control Systems. We will study the technologies in classical control engineering in the rest of this unit.

Solving Homogeneous State Equations Solving Nonhomogeneous State Equations Stabilization Conclusion

34

Вам также может понравиться