Analysis and Design of Linear Control Systems, 06th
Analysis and Design of Linear Control Systems, 06th
Analysis and Design of Linear Control System Part2Spring, 2012 Instructor: Prof. Masayuki Fujita (S5-305)
Lecture Time and Place: Tuesday, 10:4512:15, S224 Schedule: 24th, 31st, July TA: Tatsuya Ibuki, Yasuaki Wasa (S5-302)
[Ex. 11.6] Roll control for a vectored thrust aircraft* Specification Error in stationary state: from u 1 to r Less than 1% P ( s ) = G u1 = Js 2 Tracking error up to around C (s) = k k = 200 10 [rad/s]: Less than 10%
P ( j )
100 101 102
20 0
101 100 101
L ( j )
102
& = mg sin c x & + u 1 cos u 2 sin m& x & = mg (cos 1) c y & + u 1 sin + u 2 cos m& y && = ru J 1 u 1 = F1 u 2 = F2 mg
(2.27)
-180
101
100
101
102
-180
101
102
Fig. 11.12(a)
[Ex. 11.6] Roll control for a vectored thrust aircraft* r e C (s) P(s) y Lead compensator s+a C (s) = k s + b (11.12) 1 k = 950 a = 2 b = 50 a b L ( j ) loop gain Loop transfer function 20[dB] L( s) = P( s)C ( s) P ( j ) improve convergence speed (by k ) improve phase margin (by
s+2 ) s + 50
[Ex. 11.6] Roll control for a vectored thrust aircraft* Gang of Four
M S = 1 . 84
M T = 1 . 58
PS
m = 40
40[deg]
Frequency response
[Ex. 11.6] Roll control for a vectored thrust aircraft* Step response
11.6 Design Example [Ex. 11.12] Lateral control of a vectored thrust aircraft
( 2.4, Exe. 8.10)
Ex. 11.6: controller for the roll dynamics Ex. 11.12: controller for the position of the aircraft
(stabilization of both the attitude and the position)
Step response
[Ex. 11.12] Lateral control of a vectored thrust aircraft 2. Design Co for the lateral position under the assumption that that we can directly control the roll angle .
l. Design Ci so that H i provides fast and accurate control of the roll angle.
[Ex. 11.12] Lateral control of a vectored thrust aircraft Performance specification (entire system) zero steady-state error in the lateral position a bandwidth of 1 rad/s a phase margin of 45
Performance specification (inner loop) the low-frequency error to be no more than 5 % a bandwidth of 10 rad/s (10 times that of the outer loop)
inner loop ( H i ): the roll dynamics and control outer loop: the lateral position dynamics and controller
[Ex. 11.12]
u1
[Ex. 11.12]
Fig. 11.17
H i (0) mg = 39.2
Fig. 11.17
approximate
Assume that the inner loop will eventually track our commanded input.
H i (0) 39.2
Fig. 11.18 (a)
Hi
C (1 mgPi ) H i ( s) = i Exercise 1 + Ci Pi
P ( s) = H i (0) Po ( s ) mg = 2 ms + cs
Lead compensator
Co ( s ) = k o
s + ao s + bo
Co ( s ) = k o
s + ao s + bo
0 45 90
1 s+b o
bo / 10 bo 10bo
[Ex. 11.12] Lateral control of a vectored thrust aircraft Combine the inner and outer loop controllers and verify that the system has the desired closed loop performance.
150 100
50
bo = 10
Ensure adequate phase lead
90 45 0
(s + ao )
50
90
180
ao / 10 ao 10ao
270 360
104
103
10 2
101
100
101
10 2
103
Co ( s ) = 0.98
ko = 0.98 Exercise
s + 0.3 s + 10
[Ex. 11.12] Lateral control of a vectored thrust aircraft Not have integral action Gang of Four
M T = 1 . 18
PS
CS
M S = 1 . 11
Feedforward
Feedback
r
Fig. 11.20 Gang of Four for vectored thrust aircraft system
F ( s)
C (s)
u
y
P( s)
Fu ( s )
Fm ( s )
ym
u ff
C (s)
u fb
P(s)
Fu ( s )
Fm ( s )
ym
u ff
C (s)
u fb
P(s)
Fu : feedforward reference
controller
1
G yr ( s ) = F m + PF u F m 1 + PC
(11.4)
Fm ( s ) P (s)
C (s)
P(s)
G yr ( s ) =
P ( CF m + F u ) 1 + PC PF u F m = Fm + 1 + PC
(11.4)
Ideal Feedforward
Fu ( s )
Fm ( s )
Ideal Feedforward
Fu =
Feedback
(11.5)
Fu = 1 1 + PC
Feedback
Fm P
ym
(11.5)
desired response
= F m + (PF u F m )S
small S =
small S =
1 1 + PC
Fu ( s )
Fm ( s)
ym
u fd
Fd ( s)
d
P = P2 P 1
C (s)
u fb
P1 ( s)
u ff
Fu ( s )
Fm ( s )
ym
u fd
Fd ( s )
d
P2 ( s )
P2 ( s )
C (s)
u fb
P1 ( s )
u ff
1
Fig. 11.3 2DOF controller for improved response to reference signals and measured disturbances
Fig. 11.3
d y
ideal feedforward
desired output when Fu : feedforward reference controller applied as input to the process Fd : feedforward disturbance controller u fb: feedback control input C : feedback controller d : load disturbance : process P = P2 P P 1 r : reference signal 1, P 2
G yd ( s ) =
P2 (1 + F d P1 ) 1 + PC (11.6) = P2 (1 + F d P1 ) S
Fd =
(11.7)
feedback
P = P2 P 1
small S =
1 1 + PC
Fu ( s )
Fm ( s )
ym
u ff
P = P2 P 1
P( s) =
(s + 1) s2
C (s)
u fb
P1 ( s )
P2 ( s )
Fig. 11.3
1
Step Response of
desired response
Fig. 2.16
a2 Fm ( s ) = ( s + a) 2
2 aa== 01 .5
(s + 1) P( s) = s2
* = a/b
Fu =
r
Fig. 11.4(a) Overhead view
Fm ( s ) P ( s)
C (s)
ym
P(s)
y
Step Response
Fu ( s )
Fm ( s )
1
desired response
a = 0.5
process
1 P( s) = ( s + 1) 3
Fm ( s ) =
1 (0.5s + 1) 3
PI feedback controller
feedforward controller
0 . 6 s + 0 .5 C (s) = s
Fig. 11.4(b) Position and steering
Fu =
Fm ( s + 1) 3 = P (0.5s + 1) 3
must be proper
Step response
Time Responses (5.3 )
Tp
Mp
y
G yr ( j )
Ts
0
10 % ~ 90 %
= 2% and = 5%
are the most widely used
Fig. 11.5 (a) Step response Fig. 11.5 (b) Frequency response
Tr
Fm ( s ) P ( s)
C (s)
ym
P(s)
Second-order System*
Fm ( s ) =
2 K n 2 s + 2 n s + n 2
F Fu = m P
y = Fmr
Fu ( s )
Fm ( s )
How to build Fm ?
First-order System K K Fm ( s ) = 1 Ts + 1 Rise time 0.8 Tr = (ln 9)T 2.2T 0.6 0.4 Settling time 0.2 3 if = 5 % T Ts 4T if = 2% 0 The step response has no overshoot
Step Response
K T
= n 1 2
Step Response
K
Tp
Mp
Ts
0
1 2
3 4 5
Ts
1 2
M p = Ke /
Tr
10% ~ 90%
Butterworth Filter*
Butterworth Filter*
n : order of filter
2n
Low-pass filter
Fm ( j ) = F0
Denominator polynomial
n =1 n=2
2
1+ c
s + c
s + 1.4cs + c
2
F0 = 1
n =1 n=2
n=3 c = 1
Re
n=3 n=4
n=5
s3 + 2.0cs2 + 2.0c s + c
2 3
c = 1 n = 5
n=3 n=4
Low-pass filter
s
High-pass filter
1 s
Summary