Tutorial 8

Q1 Choose so= −1 + j,
R(s) E(s) K C(s)
+ Gc(s)
−
s
2
∠G (so ) = −135o − 135o = −270o ζ=0.7
so
j
Design Gc(s) = Kc(s+zc) such that.
C (s ) K 1 K β
= 2 → C (s) = × 2
R(s ) s + K s s +K so is the root of the characteristics equation:
−1

1 + K c (s + zc )G (s ) = 0
For K = 1:

c(t ) = 1− cos(t ) This is a sinusoidal response with ζ=0

if ∠Gc (so ) + ∠G (so ) = −180o
Desired s1,2 = −1 ± j, with ζ = 0.7 to give a stable response
% M p = 100 × exp (−σπ ω) = 100 × exp (−π ) = 4.32 ∠Gc (so ) = −180o − ∠G (so ) = −180o + 270o = 90o
ts = 4 1 2

s + zc
This means that zc = 1 Let the Lead compensator: be: Gc (s ) = K c
s + pc
ζ=0.7 o
Using the magnitude condition so then ∠zc − ∠pc = 90 Choose zc = 0.5 < 1
j
o −1 ⎛ 1 ⎞ o
∠zc = 180 − tan ⎜ ⎟ = 116.565
⎝ 0.5 ⎠
−zc = −1 so
j o o o
K c ( s + 1) 1 ∠p
∠pc = 116.565 − 90 = 26.565
= Kc = 1 ⇒ Kc = 2
2
2× 2
s s = −1+ j
−p
pc 1 −zc
−1 (
tan 26.565 =
o
) 1
pc − 1
⇒ pc = 3

PD Compensator Gc(s) = 2(s+1) 2

K c ( s + 0.5) 0.5 + 1
= Kc = 1 ⇒ Kc = 4
s ( s + 3)
2
In the design of a Lead compensator for the same problem, its zc s = −1+ j
2 5
must be less than 1.
s + 0.5
Gc (s ) = 4
3 s+3 4
 Alternatively, using the formula given in the text book
1.4 31% 34%
33%
β = 45, ϕ = 90, σ = 1, ω = 1 1.2
21% 4 3 2
1
1
β φ
zc = ω tan ⎡⎢ − ⎥⎤ + σ = 0.586

Amplitude
⎣2 2⎦
0.8

0.6

A
β φ
pc = ω tan ⎡⎢ + ⎤⎥ + σ = 3.414 s + 0.586 0.4
⇒ Gc (s ) = 4.83
⎣2 2⎦ s + 3.414 02
0.2 3.5 4.4
3.8 4.85
0
0 1 2 3 4 5 6
s 2 ( s + 3.414) Time ((sec))
Kc = = 4.83
(s + 0.586) s = −1+ j 1 PD Controller 2 Lead design: zc = 0.5

3 Text book design 4 Lead: zc =0.75

=0 75
5 6

System is not a standard 2nd order system with zero & additional The overshoots may be too large for the desired closed loop
ppole in the case of Lead compensator;
p ; hence % OS is higher
g than pole; we can use root locus technique to select another closed
for ζ = 0.7 loop pole by varying Kc. For example: Gc = Kc(s+1); choose
equal roots with Kc = 4;
zc pc Kc %Mp R
Remark
k jω
1.4
21%
1.2 13.5% Kc=2
Kc=2 Kc=4
4
05
0.5 3 4 33 L d
Lead 1

Amplitude
0.8
0.585 3.414 4.83 34 Textbook Kc=4
−1 σ 0.6

0.4
0.75 5 8 31 Lead
0.2
2.7 3.5
0
1 - 2 21 PD controller 0 1 2 3 4 5 6
Time (sec)

Tables of various combination of zc and pc

7 8
 Lead Compensator - observations PV Controller
R( )
R(s) K C(s)
From the set previous graphs, it can + Kp
be observed that the closer the Lead − s
2

so zero
ero is to the Origin,
Origin the less likely
likel
j nd Gc(s)
2 order dominance can be achieved. Kd
The best result was obtained for the 1+ s
Kp
−pc −1 −zc PD with zc directly below the desired
closed loop poles K pK ⎛ ⎞
The characteristics equation is given by: 1 + ⎜1 + K d s ⎟ = 0
2 ⎜ K ⎟
s ⎝ p ⎠
⎛ Kp ⎞
⎟ = 1 + G (s )K d (s + zc ) = 0
so A good guide to follow: do not place K
ζ=0.7
j the Lead zero closer to the Origin 1+ 2
Kd ⎜ s +
than the real part of the desired
s ⎝ Kd ⎠
closed loop complex poles, Compare with PD design
−zc = −1 otherwise a real pole terminating at
the Lead zero will dominate instead. K d = K c = 2 & zc = 1 ⇒ K p = 2
9 10

C K pK C 2 Q2
= K=1 = 2 R(s) E(s) 820 C(s)
R s2 + K + K pK R s + 2s + 2 + Gc(s)
s(s + 10 )(s + 20 )
K d Ks −

14
1.4
Peak amplitude: 1.04
1.2 Overshoot (%): 4.32 Time (sec): 4.29 Assume that Gc(s) = 1
At time (sec): 3.18 Amplitude: 1.02
1 Solve the characteristics equation:
0.8
s(s + 10 )(s + 20 ) + 820 = 0
Amplitude

06
0.6
Desired complex dominant poles are:
0.4
s1,2 = −3.6 ± j4.8, ζ = 0.6
02
0.2

For unit ramp, static 820

0
K v = Lim s = 4.1
0 1 2 3
Time (sec)
4 5 6 velocity error constant: s →0 s(s + 10 )(s + 20 )
11 12
 360 360 360 At so = −3.6 + j4.8
For R( s ) = 2
, error, ess = = = 87.8o
s Kv 4.1 so
× j4.8
Desired error,, ess = 8.78° ⇒ K v = 41
o ×
−3.6 −0.1 −0.01
Lag compensator to increase gain but retain the above dominant poles.
s + zc ⎡ −1 ⎛ ⎞⎤ − ⎡180o − tan −1 ⎛ 4.8 ⎞⎤
∠Gc (so ) = 180 − tan ⎜
4.8
Gc = K c ; pc < zc ⎢
o
⎟ ⎜ ⎟
s + pc ⎣ ⎝ 3.6 − 0.1 ⎠⎥⎦ ⎢⎣ ⎝ 3.6 − 0.01 ⎠⎥⎦
o
= −0.7
K z K z
New K v = 41 = 4.1 × c c ⇒ c c = 10
pc pc Ph
Phase llag off Gc is
i quite
i small;
ll resulting
l i root locus
l off 1 +G
GcG=0
G 0
should be quite close to the root locus of 1 +G =0
Choose Kc =1, zc = 0.1 and pc = 0.01 to get the new Kv value

13 14

The Lag compensator will contribute an additional pole: Root locus of 1 + Gc G = 0

solving 1 +G
Gc G
G=0
0 Choose another pair of zc and pc
s(s + 10 )(s + 20 )(s + 0.01) + 820(s + 0.1) = 0 Choose Kc =1, zc = 0.5 and pc = 0.05
s1 = −22.8 ∠Gc (so ) = −3.63o
Roots are:
s2,3 = −3.55 ± j 4.75
s4 = −0.1023 s (s + 10 )(s + 20 )(s + 0.05) + 820(s + 0.5) = 0

s1 = −22.77
The pole at s = −22.8 does not affect the 2 complex poles Roots are: s2,3 = −3.36 ± j 4.54
p poles
The complex p are quite
q close to the original
g complex
p poles
p s4 = −0.57
The pole closes to the Origin will stretch the time to reach final The pole closes to the origin will stretch the time to reach final
value to about 40 sec compared with the original settling time of 1.1 value to about 7 sec compared with the original settling time of 1.1
11
sec sec
15 16
Error of with Gc1 & Gc2 for closed loop response to ramp of 360°/sec
100
00
90
80 G

70
60
Errorr

50
40 Gc1
30 Gc2
20
10
0
0 5 10 15 20 25 30 35 40 45 50
Time, sec 17