Controls Combined Lecture Notes
Controls Combined Lecture Notes
Controls Combined Lecture Notes
PLANT
u(t) y(t)
G
u=θ
y=v
u
OPEN LOOP SYSTEM WITH COMPENSATOR
yd
u
DISTURBANCE EFFECT
d(t)
+
yd (t) u(t) y(t)
G−1 G
FEEDBACK
n(t) d(t)
yd (t) + +
e(t) y(t)
K A G
+ −
S
+
n(t)
3
MECHANICAL SYSTEMS
F = kx
F = bẋ
x
k
m f
b
4
kx
m f
bẋ
EXAMPLE
x2 x1
k2 k1
m2 m1 f
b
5
k1 (x1 − x2 )
m1 f
b(x˙1 − x˙2 )
k1 (x1 − x2 )
k2 x2 m2
b(x˙1 − x˙2 )
ELECTRICAL SYSTEMS
R
v = iR
C
dv
i=C
dt
L
di
v=L
dt
∑
Kirchhoff's current law: i=0
6
i1 i3
i2
+
vs − C R v
i1 = i2 + i3
vs − v
i1 =
R
d
i2 = C (v − 0) = C v̇
dt
v−0 v
i3 = =
R R
vs − v v
= C v̇ +
R R
vs − v = RC v̇ + v
vs = 2v + RC v̇
RC v̇ + 2v = vs
7
VOLTAGE DIVIDER
vin
R1 i1
vout
R2 i2
i1 = i2
vin − vout vout − 0
=
R1 R2
vout R2
=
vin R1 + R2
IMPEDANCE
ZR
ZR = R
ZC
1
ZC =
jωC
ZL
ZL = jωL
8
Z1 Z2 ZT = Z1 + Z2
Z1
1 1 1
ZT
= Z1
+ Z2
Z2
OPERATIONAL AMPLIFIER
v− −
vout
v+ +
vout = k(v+ − v− )
k=∞
v+ = v−
i2 Rout
i1 Rin
vin −
+ vout
+ +
(Virtual ground)
− −
9
i1 = i2
vin − 0 0 − vout
=
Rin Rout
vout Rout
=−
vin Rin
CONSTANT COEFFICIENT ORDINARY DIFFERENTIAL EQUATIONS:
Constant coefficient ordinary differential equations have following parts of
solutions:
Natural Forced
(Homogeneous) (Particular)
EXAMPLE:
b
m F0
bẋ=bv m F0
F0 − bẋ(t) = mẍ(t)
F0 − bv(t) = mv̇(t)
mv̇(t) + bv(t) = F0
SOLUTIONS:
11
F0 = 0
v(0) = v0
mv̇h (t) + bvh (t) = 0
Solution assumption:
vh (t) = Aest
v̇h (t) = Asest
If the assumption is correct then the assumed solution should satisfy the equa-
tion, therefore
A = vh (0) = v(0) = v0
vh (t) = v0 e− m t
b
12
TIME CONSTANT
v(t)
v0 e− m t
b
m
t
τ =t= b
FORCED RESPONSE
The transient response will be of the form Ae−st for a stable system and the
steady state response will be the part which exists after the transient decay
13
completely. The aforesaid system has the steady state response i.e.
:0
mv̇(t)
+ bvss (t) = F0
F0
vss (t) =
b
v(t) = 0, at t = 0
The complete forced response vf (t) for the system above becomes
F0
vf (t) = Ae− m t +
b
| {z } b
|{z}
TRANSIENT STEADY STATE
F0
vf (0) = 0 = A +
b
F0
A=−
b
F0 −m
b
vf (t) = 1−e t
b
vf (t)
1(t)
F0
F0 vf (t) b
F0
mv̇ + bv
t
0
−m
b F0 −m
b
vT OT (t) = vh (t) + vf (t) = v0 e t
+ 1−e t
b
14
TYPES OF INPUT:
u(t) y(t)
G
IMPULSE:
The Dirac delta function, often referred to as the unit impulse or delta func-
tion, is the function that defines the idea of a unit impulse in continuous-time.
Informally, this function is one that is infinitesimally narrow, infinitely tall, yet
integrates to one.
ˆ
+∞
δ(t)dt = 1
−∞
When input to a LTI system is an impulse the response is called the impulse
response h(t).
h(t)
δ(t)
u(t) h(t)
LTI
0 t
UNIT STEP:
0, t < 0
1(t) =
1, t ≥ 0
When input to a LTI system is unit step the response is called the step response.
15
y(t)
1(t)
1
u(t) y(t)
0
t LTI
t
d
δ(t) = 1(t)
dt
SINUSOID:
−0.01
−1
In previous example the forced response is the step response with step having
an amplitude of F0 .
F0 −m
b
vf (t) = 1−e t
b
Using the relation of impulse and step if we take the derivative of the above
16
1d −m
b
= 1−e t
b dt
!
1 −m
b b
= −e t
−
b m
!
1 b
−m
b
t
= (e
b
m
1
e− m t
b
h(t) =
m
It is a crude approach considering its limitations. Next we discuss a better
approach in time domain.
17
F0 F0
F0 m m
F0
m
´ v
1
m + v̇
−
b
m
mv̇ + bv = 0
v(t) = Aest
mAsest + bAest = 0
ms + b = 0
18
b
s=−
m
F0
v(0+ ) = v(0) =
m
F0
v(0) = =A
m
F0
A=
m
F0
e− m t
b
v(t) =
m
Now if the input is step i.e. F0 = 1, then
1
e− m t
b
h(t) = v(t) =
m
The same approach is valid for any higher order system.
...
y + a1 ÿ + a2 ẏ + a3 y = bu
Figure below is the integral block diagram. Using the same approach in the
above example.
19
bt2
1 b b bt
...
y
´ ÿ
´ ẏ
´ y
u b
+
−
a1
a2
a3
Solution assumption is
y(t) = Aest
Solving the third order polynomial will give the three values of s1 , s2 and s3 .
20
so
y(t) = A1 es1 t + A2 es2 t + A3 es3 t
Therefore we obtain following set of linear equations from the above equation
y(0) = 0 = A1 + A2 + A3
ẏ(0) = 0 = A1 s1 + A2 s2 + A3 s3
ÿ(0+ ) = ÿ(0) = b = A1 s21 + A2 s22 + A3 s23
1 1 1 A1 0
s1 s2 s3
A2 = 0
s21 s22 s23 A3 b
A y x
CONVOLUTION METHOD
Impulse response to describe all possible arbitrary inputs:CONVOLUTION IN-
TEGRAL
u(t)
∆t
u1
u2
u(t)
u0
t0 t1 t2
22
h(t − t0)u0
u0
u(t) y(t)
LT I
t0 t
h(t − t1)u1
u1
u(t) y(t)
LT I
t1 t
h(t − t2)u2
u2
u(t) y(t)
LT I
t2 t
u0 u1 u2
u(t) y(t)
LT I
t0 t1 t2 t
23
ˆt
y(t) = u(τ )h(t − τ )dτ
0
b
m F0
Impulse Response:
1 −mb
t
h(t) = m
e
F0 1(t)
F0
t
0
0, t<0
F0 1(t) =
F0 , t ≥ 0
u(t) = F0 1(t)
ˆt
y(t) = u(τ )h(t − τ )dτ
ˆ
0
t !
1
e− m (t−τ ) dτ
b
= F0
m
0
η = t − τ, dη = −dτ
η = t, τ =0
η = 0, τ =t
ˆ 0 !
1
e− m η dη
b
=− F0
m
t
! ! 0
F0 m
−m
b
η
= − − (e )
m
b
t
25
F0 −m
b
y(t) = 1−e t
b
yd e u y
− K G
y(t) = u(t) ⊛ g(t)
u(t) = e(t) ⊛ k(t) SOLVE SIMULTANEOUSLY
e(t) = yd (t) − y(t) ⊛ h(t)
26
LAPLACE TRANSFORM
Im(s)
+4i
+3i
2 + 3i
3i
2i
13
√
3
i 56.3
Re(s)
−4 −2 2 4
−i 2
−2i
−3i
−4i
s-domain, s = σ + jω
s = σ + jω
ˆ∞
F (s) = L {f (t)} =
∆
f (t)e−st dt
0
u(t) y(t)
h(t)
U (s) Y (s)
H(s)
Y (s) = U (s)H(s)
STEP FUNCTION:
0, t<0
1(t) =
F0 , t ≥ 0
ˆ∞
L {1(t)} = 1(t)e−st dt
0
! ∞
1 −st
= − (e )
s
! 0
1 1
= − (0 − 1) =
s s
1
L {1(t)} =
s
EXPONENTIAL FUNCTION:
ˆ∞
L {u(t)} = L {e−at } = e−at e−st dt
0
ˆ∞
= e−(s+a)t dt
0
! ∞
1 −(s+a)t
= − (e )
s+a
! 0
1 1
= − (0 − 1) =
s+a s+a
1
L {e−at } =
s+a
IMPULSE FUNCTION:
u(t) =
∆
1
∆
, 0≤t≤∆
δ(t) = lim
∆→0
0, otherwise
ˆ∞
f (t)δ(t)dt = f (0)
−∞
ˆ∞
L {u(t)} = L {δ(t)} = δ(t)e−st dt
0
ˆ∞
= e−s(0) δ(t)dt
0
=1×1=1
L {δ(t)} = 1
29
1
→ INTEGRATOR
s
s → DIFFERENTIATION
y ′ (t) 1 y(t)
s
y(t) y ′ (t)
s
LINEARITY:
ˆ∞
L {α1 f1 (t) + α2 f2 (t)} = {α1 f1 (t) + α2 f2 (t)}e−st dt
0
ˆ∞ ˆ∞
= α1 f1 (t)e−st dt + α2 f2 (t)e−st dt
0 0
= α1 F (s) + α2 F (s) ■
EXAMPLE
1 2
y(t) = t + t2 ←→ Y (s) = +
s2 s3
DIFFERENTIATION:
( )
d
L (f (t)) = L {f ′ (t)} = sF (s) − f (0)
dt
30
ˆ∞
L {f ′ (t)} = f ′ (t)e−st dt
0
ˆB B ˆB
udv = uv − vdu
A
A A
u = e−st , du = −se−st dt
dv = f ′ (t), v = f (t)
∞ ˆ∞
= e−st f (t) − f (t)(−se−st )dt
0
0
ˆ∞
= (0 − 1 × f (0)) + s f (t)e−st dt
0
= sF (s) − f (0) ■
EXAMPLE:
u = e−st , du = −se−st dt
1
dv = cos (6t), v = sin (6t)
6
! 0∞ ˆ∞ !
:
1 1
= 3 e−st
sin (6t) −3 sin (6t) (−se−st )dt
6 6
0 0
ˆ∞
s
= sin (6t)e−st dt
2
0
u = e−st , du = −se−st dt
1
dv = sin (6t), v = − cos (6t)
6
!∞ ˆ∞ !
s 1 s 1
= − e−st cos (6t) − − cos (6t) (−se−st )dt
2 6 2 6
0 0
!! ˆ∞
s 1 s2
= − 0− − cos (6t)e−st dt
2 6 12
0
32
ˆ∞ ˆ∞
s s2
3 cos (6t)e−st dt = − cos (6t)e−st dt
12 12
0 0
ˆ∞ ˆ∞
−st s s2
cos (6t)e dt = − cos (6t)e−st dt
36 36
0 0
! ˆ∞
s2 s
1+ cos (6t)e−st dt =
36 36
0
ˆ∞
−st s 36
cos (6t)e dt =
×
36
s2 + 36
0
ˆ∞
L {3 cos (6t) = 3 cos (6t)e−st dt
0
3s
=
s2 + 36
33
SECOND METHOD:
ˆ∞
L {3 cos (6t) = 3 cos (6t)e−st dt
0
ˆ∞ !
6ti −6ti
e +e
=3 e−st dt
2
0
ˆ∞ ˆ∞
3 3
= e(−s+6i)t dt + e(−s−6i)t dt
2 2
0 0
∞ ∞ !
3 e(−s+6i)t e(−s−6i)t
= +
2 −s + 6i −s − 6i
0 0
∞ ∞ !
3 e−(s−6i)t e−(s+6i)t
= +
2 −s + 6i −s − 6i
0 0
!
3 1 1
= (0 − 1) + (0 − 1)
2 −s + 6i −s − 6i
!
3 −1 −1
= +
2 −s + 6i −s − 6i
34
!
3 −1(−s − 6i) − 1(−s + 6i)
=
2 (−s − 6i)(−s + 6i)
!
3 s++s−
6i
6i
=
2 (−s)2 − (6i)2
!
3 2
s
= , ∵ i2 = −1
2
s2 + 36
3s
L {3 cos (6t) =
s2 + 36
35
:0
0
2 ′ *
(s + a1 s + a2 )Y (s) = bU (s) + (a1+
s)y(0) y
+ (0)
1
Y (s) = U (s)
s2 + a1 s + a2
Y (s) 1
H(s) = =
U (s) s2 + a1 s + a2
U (s) Y (s)
H(s)
Y (s) = U (s)H(s)
36
b
m f
FORCED RESPONSE:
F0
t
0
0, t<0
F0 1(t) =
F0 , t ≥ 0
37
STEP RESPONSE:
f (t) = F0 1(t)
F0
F (s) =
s
1 F0
V (s) = ×
ms + b s
a
L {1 − e−at } =
s(s + a)
b
F0 m
1
V (s) = × ×( m )×
m
b b s
s+ m
b
F0
= ( m )
b b
s s+ m
F0 ( )
1 − e− m t
b
v(t) =
b
IMPULSE RESPONSE:
f (t) = δ(t)
F (s) = 1
1
V (s) = ×1
ms + b
1 1
V (s) = × b
m s+ m
1
L {e−at } =
(s + a)
1 −bt
v(t) = h(t) = e m
m
38
h(t)
δ(t)
1
ms+b
0 t
v(t)
F0 1(t)
F0
1
t ms+b
0 t
f (t) = 0
V (0) = v0
m(sV (s) − v(0)) + bV (s) = 0
mv(0) v0
V (s) = = b
ms + b s+ m
1
L {e−at } =
s+a
( b )
v(t) = v0 e− m t
Using tables
1
→ 1(t)
s
1
→ e−at
s+a
s
→ cos(at)
s2 + a2
9s2 + 16 A Bs + C
Y (s) = = +
s(s2 + 4) s s2 + 4
= 9s2 + 16 = A(s2 + 4) + (Bs + C)(s)
= 9s2 + 16 = (A + B)s2 + Cs + 4A
A + B = 9, C = 0, 4A = 16
A = 4, B = 5, C=0
9s2 + 16 4 5s
= = +
s(s2 + 4) s s2 + 4
y(t) = 4 · 1(t) + 5 cos(2t)
40
y(t)
9
1
t
−1 3 6
EXAMPLE
τf
θ
m
T
g
sin
θ
mg cos θ l
θ
mg
GOVERNING LAW:
τ = I θ̈
where τ is the torque, θ̈ is the angular acceleration and I is moment of inertia
which measures the extent to which an object resists rotational acceleration
about a particular axis, and is the rotational analogue to mass I = mL2 .
τf − mgl sin θ = I θ̈
τf − mgl sin θ = ml2 θ̈
sin θ
θ
− π2 π
2
−1
( )
g
ml2 θ̈ + θ = τf
l
g τf
θ̈ + θ=
l ml2
42
g Tf (s)
s2 Θ(s) − sθ(0) − θ̇(0) + Θ(s) =
( l
) ml2
g Tf (s)
s2 + Θ(s) = + sθ(0) + θ̇(0)
l ml2
Tf (s) sθ(0) + θ̇(0)
Θ(s) = ( ) + ( )
ml2 s2 + gl s2 + gl
INITIAL CONDITION RESPONSE WITH ZERO VELOCITY:
θ̇(0) = 0
θ(0) = θ0
Tf (s) = 0
sθ(0)
Θ(s) = ( g
)
s2 + l
s
= (√ )2 θ 0
g
s2 + l
s
L {cos at} =
s2 + a2
(√ )
g
θ(t) = θ0 cos t
l
Θ(s) s
= g
θ(0) s2 +
l
ZERO: s = 0
√
g
POLES: s = ± i
l
43
jω
√
g
l
σ
−2 2
√
g
−
l
6s2 + 18s + 12
H(s) =
2s3 + 10s2 + 16s + 12
(s + 1)(s + 2)
=3
(s + 3)(s + 1 − j)(s + 1 + j)
Roots of the numerator polynomial of the LTI system transfer function are
the zeros i.e.
s = −1, s = −2
s = −3, s = −1 + j, s = −1 − j
44
|H(s)|
−3 jω
−2
−1 1
0 σ
−1
45
b1 sm + b2 sm−1 + · · · + bm s + bm+1
Y (s) =
sn + a1 sn−1 + · · · + an−1 s + an
Y (s) = s is a pure derivative which in real system rolls off i.e. it’s value
goes to ∞ as ω increases rapidly.
3s2 + 4s + 5
Y (s) =
s2 + 6s + 7
3
s2 + 6s + 7 3s2 + 4s + 5
− 3s2 − 18s − 21
− 14s − 16
−14s − 16
Y (s) = 3 +
s2 + 6s + 7
In the next section, we consider Y (s) as LTI system transfer function as-
suming that an impulse was applied to the system.
46
PARTIAL FRACTIONS
GENERAL POLE ZERO FORM
Y
m
(s − zi )
Y (s) = K i=1
Yn
(s − pi )
i=1
(s − z1 )(s − z2 )(s − z3 )
Y (s) =
(s − p1 )(s − p2 )(s − p3 )
c1 c2 c3
= + +
(s − p1 ) (s − p2 ) (s − p3 )
c1 (s − p1 ) c2 (s − p1 ) c3 (s − p1 )
(s − p1 )Y (s) = + +
(s − p1 ) (s − p2 ) (s − p3 )
*0 *0
p
(s−
p1)
c2 (s −
p1)
c3 (s −
c1 1)
(s − p1 )Y (s) = + +
(s−
(s − p2 ) (s − p3 )
p1 )
s=p1
47
(s − p1 )Y (s) = c1
s=p1
∴ ci = (s − pi )Y (s)
s=pi
jω
σ
−4 −3 −2 −1
−1
48
(s + 2)(s + 4)
Y (s) =
s(s + 1)(s + 3)
c1 c2 c3
= + +
s s+1 s+3
(s + 2)(s + 4)
8
c1 = sY (s) = =
(s + 1)(s + 3) 3
s=0 s=0
(s + 2)(s + 4)
3
c2 = (s + 1)Y (s) = =−
s(s + 3) 2
s=−1 s=−1
(s + 2)(s + 4)
1
c3 = (s + 3)Y (s) = =−
s(s + 1) 6
s=−3 s=−3
8 1 3 1 1 1
= · − · − ·
3 s 2 s+1 6 s+3
8 3 1
y(t) = 1(t) − e−t − e−3t
3 2 6
49
8 1 3 1 1 1
Y (s) = · − · − ·
3 s 2 s+1 6 s+3
8 3 1
y(t) = 1(t) − e−t − e−3t
3 2 6
3 1
e−t e−3t
2 6
8
1(t)
3
8
3
t
0 t t
y(t)
2.67
1 t
0 3.81 6
50
COMPLEX ROOTS: pj = pi , p i , pj ∈ C
p1 = −σ + jω
p2 = p1 = −σ − jω
= s2 + 2σs + (σ 2 + ω 2 )
σ
= s2 + 2σs + σ 2 + 2
σ
− 2
+ ω2
= (s + σ)2 + ω 2
jω
√ σ ω
σ2
ω +
ω2
σ
−σ 2
ω
ω 2 +
√ σ
σ −ω
51
1 1
Y (s) = =
s2 + 2σs + (σ 2 + ω 2 ) (s + σ)2 + ω 2
c1 c2 = c1
= +
s + σ − jω s + σ + jω
c1 = a + jb, c2 = c1 = a − jb
a + jb a − jb
= +
s + σ − jω s + σ + jω
(a + jb)(s + σ + jω) + (a − jb)(s + σ − jω)
=
(s + σ − jω)(s + σ + jω)
+ −bω+aσ+as− jbσ −bω
= aσ+as+
jbσ +
jbs
jaω −−
jbs
jaω
(s+σ)2 +ω 2
b
2 e−σt sin bt
(s + a)2 + b2
EQUIVALENT FORM:
p a b
= 2 · e−σt · a2 + b 2 · ( √ cos ωt − √ sin ωt)
a2 + b 2 a2 + b2
b
y= √
a2 + b2
then as we know the equation of unit circle
x2 + y 2 = 1
!2 !2
a b
√ + √ =1
a2 + b2 a2 + b 2
This means the point (x, y) = ( √a2a+b2 , √a2b+b2 ) lines on unit circle.
53
1
(x, y) = (cos α, sin α)
a2 + b2 sin α
b2
1
+
2
a2
√
√
b=
α
√
x
−1 − 21 a= a2 + b2 cos α 1
− 12
−1
Hence
cos (ωt − α) = cos (ωt) cos (−α) + sin (ωt) sin (−α)
cos (−α) = cos (α), cos is an even function
sin (−α) = − sin (α), sin is an odd function
tan (−α) = − tan (α), tan is an odd function
cos (ωt − α) = cos (ωt) cos (α) − sin (ωt) sin (α)
Finally,
√
−σt
y(t) = 2 · e · a2 + b2 · cos (ωt − α)
where
( )
−1 b sin (−α) − sin (α)
α = − tan +θ ∵ = = − tan (α)
a cos (−α) cos (α)
The inverse of an odd function is odd (e.g. tan−1 (−x) = − tan−1 (x) is odd
as tan (x) is odd). Therefore − tan (α) = ab gives tan (α) = − ab then α
( ) ( )
= tan−1 − ab = − tan−1 ab . θ is required for principal value, otherwise
can be ignored for response calculation.
0◦ ,
(a, b) is in I or IV quadrants
θ = 180◦ , (a, b) is in II quadrant
−180◦ , (a, b) is in III quadrant
This is because the function θ = tan−1 (x) has principal values lies in between
(− π2 , π2 ).
55
1
x
−2π− 7π − 3π −π − 3π − π
2
π
4
π
2
π 5π 3π 2π
4 2 4 4 2
−2
π
2
x
−8 −6 −4 −2 2 4 6 8
− π2
56
EXAMPLE
1 1
Y (s) = =
s2 + s + 100.25 (s + 0.5 − j10)(s + 0.5 + j10)
c1 c1
= +
(s + 0.5 − j10) (s + 0.5 + j10)
c1 = (s + 0.5 − j10)Y (s)
s=−0.5+j10
1 1
=
=
−
0.5 + j10 +
0.5 + j10 j20
1ej0
= 0.05ej(0− 2 ) = 0.05e−j 2 = 0 − j0.05
π π
= j π2
20e
c1 = 0 − j0.05, c2 = c1 = 0 + j0.05
c1 = a + jb = 0 + j0.05
p
a = 0, b = 0.05, a2 + b2 = 0.05
57
!
b
α = tan−1 +θ
a
π
= tan−1 (∞) + 0◦ =
2
p1 = −σ + ωj = −0.5 + 10j
σ = −0.5, ω = 10
p
y(t) = 2 · e−σt · a2 + b2 · cos (ωt − α)
!
π
y(t) = 2 · e−0.5t · 0.05 · cos 10t −
2
58
0.08
0.04
y(t)
−0.04
−0.08
0 2 4 6 8 10 12
t
59
jω
σ
pi
N (s)
Y (s) =
(s + pi )3
c1 c2 c3
= + +
(s + pi ) (s + pi )2 (s + pi )3
3
c3 = (s + pi ) Y (s)
s=−pi
d
c2 = [(s + pi )3 Y (s)]s=−pi
ds
1 d2
c1 = · 2 [(s + pi )3 Y (s)]s=−pi
2 ds
60
In general
" #
1 di
ck−i = [(s + p)k Y (s)] , i = 0, 1, · · · , k − 1
i! dsi
s=−p
N (s)
Y (s) =
(s + pi )3
c1 c2 c3
= + +
(s + pi ) (s + pi )2 (s + pi )3
−pi t −pi t t2 −pi t
y(t) = c1 e + c2 te + c3 e
2!
where c1 , c2 and c3 are calculated by partial fractions.
EXAMPLE
s+2
Y (s) =
(s + 1)3
c1 c2 c3
= + +
(s + 1) (s + 1)2 (s + 1)3
k = 3, i = 0, 1, 2
" !#
1 d0 s + 2
i = 0, c3 =
(s+ 1)3
0
3
0! ds (s+
1)
s=−1
= −1 + 2 = 1
61
" !#
1
1 d s+2
i = 1, c2 =
(s+ 1)3
1! ds1
(s+
1) 3
s=−1
d
= (s + 2) = 1
ds" !#
2
1 d 3 s+2
i = 2, c1 = (s
+ 1)
2! ds2
(s+
1) 3
s=−1
2
1 d
= · (s + 2) = 0
2 ds2
−t −t t2
y(t) = 0 · e + 1 · te +1· e−t
2!
−t t2
y(t) = te + e−t
2!
t2
e−t
te−t 2!
t t
y(t)
0.6
0.4
0.2
t
0 4 8 12
62
c1 c2
= +
(s + 2 − i)2 (s + 2 − i)
c1 c2
+ +
(s + 2 + i)2 (s + 2 + i)
1
c1 = (s + 2 − i) Y (s)
2
=
(s + 2 + i)2
s=−2+i s=−2+i
1 1 1
= = = − , ∵ i2 = −1
(−2
+i+2
+ i)2 (2i)2 4
1
c1 = c1 = −
4
63
" #
d
c2 = (s + 2 − i)2 Y (s)
ds
s=−2+i
" !#
d 1
=
ds (s + 2 + i)2
s=−2+i
!
2
= −
(s + 2 + i)3
s=−2+i
2 2
=− = , ∵ i3 = −i
(−
2+i+
2+ i)3 8i
2ej0 1 1
e−j 2
π π
= j π2
= ej(0− 2 ) =
8e 4 4
1 1
e−j 2 ,
π π
c2 = c2 = ej 2
4 4
c1 c1
Y (s) = +
(s + 2 − i)2 (s + 2 + i)2
c2 c2
+ +
(s + 2 − i) (s + 2 + i)
64
1 jπ 1 jπ
4
e 4
e
Y (s) = +
(s + 2 − i)2 (s + 2 + i)2
1 −j π2 1 j π2
4
e 4
e
+ +
(s + 2 − i) (s + 2 + i)
1 !
2· · t2−1 1 π
y(t) = 4 e−2t cos (t − π) + 2 · e−2t · · cos t −
(2 − 1)! 4 2
cos (t − π) = cos (t) cos (−π) + sin (t) sin (−π) = − cos t,
!
1 1 π
y(t) = − te−2t cos (t) + e−2t cos t −
2 2 2
65
y(t)
0.025
0.02
0.015
0.01
0.005
0 t
2 4 6
−0.005
Im(s)
t t
t
Re(s)
t t
This is only applicable if the LTI system is stable and stability criteria is
ALL THE POLES OF THE SYSTEM SHOULD BE IN THE LEFT HALF PLANE OF
COMPLEX DOMAIN. i.e.
Re(pi ) < 0, ∀ pi
67
DC GAIN OF A STABLE LTI SYSTEM: This is the gain of the system when fre-
quency is zero therefore called as zero frequency gain.
METHOD 1: Using time constant form i.e.
K(1 + sτ1 )(1 + sτ1 ) · · · (1 + sτm )
H(s) =
sn (1 + sτ1′ )(1 + sτ2′ ) · · · (1 + sτn′ )
EXAMPLE:
s2 + 3s + 16
H(s) =
s3 + 7s2 + 9s + 32
(s+1.5−j3.7081)(s+1.5+j3.7081)
=
(s+6.3756)(s+0.3122−j2.2185)(s+0.3122+j2.2185)
! !
s s
(1.5 − j3.7081)(1.5 + j3.7081) +1 +1
1.5 − j3.7081 1.5 + j3.7081
= s
!
s
!
s
!
(6.3756)(0.3122 − j2.2185)(0.3122 + j2.2185) +1 +1 +1
6.3756 0.3122 − j2.2185 0.3122 + j2.2185
! !
s s
16 +1 +1
1.5 − j3.7081 1.5 + j3.7081
= ! ! !
s s s
32 +1 +1 +1
6.3756 0.3122 − j2.2185 0.3122 + j2.2185
16 1
K= =
32 2
68
EXAMPLE:
s2 + 3s + 16
H(s) =
s3 + 7s2 + 9s + 32 !
s2 + 3s + 16
K = lim H(s) = lim
s→0 s→0 s3 + 7s2 + 9s + 32
16 1
= =
32 2
DC GAIN IN COMPLEX DOMAIN:
s2 + 3s + 16
H(s) =
s3 + 7s2 + 9s + 32
(s+1.5−j3.7081)(s+1.5+j3.7081)
= (s+6.3756)(s+0.3122−j2.2185)(s+0.3122+j2.2185)
s = σ + jω
69
(σ+jω+1.5−j3.7081)(σ+jω+1.5+j3.7081)
H(σ + jω) = (σ+jω+6.3756)(σ+jω+0.3122−j2.2185)(σ+jω+0.3122+j2.2185)
(1.5 − j3.7081)(1.5 + j3.7081)
H(σ + jω) =
(6.3756)(0.3122 − j2.2185)(0.3122 + j2.2185)
0+j0
|16| 1
|H(0 + j0)| = =
|32| 2
−68◦
2.2185
−82◦
4
2.2403
jω
0
6.3756
82◦
−2.2185
68◦
−3.7081
−6.3756
−1.5
−0.3122
0
σ
70
0.8
1(t)
t
4 8 12 16
71
SINUSOIDS TUTORIAL
y
Q(x, y)
1
y = sin (θ)
ωt
θ=
x
O x = cos (θ)
x
cos (θ) = =x
1
y
sin (θ) = =y
1
Q(x, y) = Q(cos (θ), sin (θ))
θ = ωt
radians
θ is in radians. ω is angular frequency which is seconds
. 1 cycle is equal to
2π radians. If we have ω = 2π then it means we will have 1 cycle in one
72
1
√ √
2 2
2 2
Q(cos (π), sin (π)) Q(cos (0), sin (0))
4
Q(−1, 0) (ω : 2π, t : 8
) Q(1, 0) (ω : 2π, t : 0)
√ √
− 2
2
− 2
2
−1
Q(cos ( 5π
4
), sin ( 5π
4
)) Q(cos ( 7π
4
), sin ( 7π
4
))
√ √ √ √
Q(− 2
2
, − 22 ) (ω : 2π, t : 5
8
) Q( 2
2
, − 22 ) (ω : 2π, t : 7
8
)
Q(cos ( 3π
2
), sin ( 3π
2
))
Q(0, −1) (ω : 2π, t : 6
8
)
sin (2πt)
1
√
2
2
0 t
1 2 3 4 5 6 7 1
8 8 8 8 8 8 8
√
− 22
−1
73
Now if we make ω = 2(2π) = 4π and trace the red height for one second
we will have two cycles.
sin (2(2π)t)
1
√
2
2
0 t
1 2 3 4 5 6 7 1
8 8 8 8 8 8 8
√
− 22
−1
sin (3(2π)t)
1
√
2
2
0 t
1 2 3 4 5 6 7 1
8 8 8 8 8 8 8
√
− 22
−1
74
Figure below shows the concept of phase, we are advancing the signal by
1 π
time unit of 4
which gives a phase of 2
so obtain the orange curve which is
cos (2πt) i.e.
1 π
sin (2π(t + )) = sin (2πt + ) = cos (2πt)
4 2
sin (2π(t + 41 ))
1
√
2
2
0 t
1 2 3 4 5 6 7 1
8 8 8 8 8 8 8
√
− 22
−1
rad
The following two plots are with respect to θ . Since θ = ωt, ω is in sec
and
rad
time is in sec. Therefore θ is
sec
× = rad.In this case we can observe
sec
from the following plots i.e.
!
π
sin θ + = cos (θ)
2
75
sin (θ)
1
√
2
2
0 θ = ωt
π π 3π π 5π 3π 7π 2π
4 2 4 4 2 4
√
− 22
−1
sin (θ + π2 )
1
√
2
2
0 θ = ωt
π π 3π π 5π 3π 7π 2π
4 2 4 4 2 4
√
− 22
−1
76
u(t) = est
ˆ∞
∵ H(s) = h(τ )e−sτ dτ
0
using the above result with s = ±jω , adding the scaled factors, where the
A
scaling factor is 2
.
A A
y(t) = H(s) jωt
e + H(s) e−jωt
2 2
s=jω s=−jω
H(jω) = M ejϕ
H(−jω) = M e−jϕ
M = |H(jω)|, ϕ = H(jω)
jω
c
|c
∠−c◦ |
|a|
a ∠−a ◦
ω
d b
σ
|d|
∠a◦
a |b|
∠c◦
c
|a|2 ·|b|
|H(jω)| = |c|2 ·d
, H(jω) = a + a + b − c − c − d
Figure above shows the complex domain view of |H(jω)| and H(jω) at
specific frequency ω .
79
Therefore
A A
y(t) = H(jω) ejωt + H(−jω) e−jωt
2 2
A A
= M ejϕ ejωt + M e−jϕ e−jωt
2 2 !
j(ωt+ϕ) −j(ωt+ϕ)
e +e
= MA
2
y(t) = M A cos (ωt + ϕ)
Hence it can be seen that applying the input u(t) = cos (ωt) to a LTI sys-
tem having an impulse response h(t) and transfer function H(s) gives out-
put y(t) with same frequency ω having new amplitude and phase M =
|H(jω)|, ϕ = H(jω) respectively. Figure below gives a pictorial rep-
resentation of the idea. The figure is neglecting the transient response of the
system and only explicating the steady state portion where the same frequency
of the input with amplitude and phase change can be observed.
u(t) y(t)
0
0.5 1
t LT I 0 t
0.5 1
−1 −1.5
80
Bode Diagram
−20
Magnitude (dB)
−40
−60
−80
0
−45
Phase (deg)
−90
−135
−180
10−1 100 101 102
Frequency (rad/s)
The bode plot shows that this system will attenuate −34.3537 db the 7 rad/sec
frequency and gives a phase shift of −126.34◦ . Converting db to magnitude
value and degree to radians values.
|H(j7)| = 10−
34.3537
= 0.0192
20
π
H(j7) = −126.34◦ × = −2.2066
180◦
Figure below depicts the same information in graphical view in complex do-
81
main.
ℑ{s}
0◦ 1
H(s) = = −486.4◦ 10i H(s) =
7◦
.
20
. 1
H(s)s=0+0i = √
16
4i √ = 0.05̄
18 × 18
3 3i
2i
√ 18
◦
135
4
10 .
1i
−10 −9 −8 −7 −6 −5 −4 −3 −2 −1 1 2 3 4 5 6 7 8 9 10
ℜ{s}
−135◦
−1i 6
18
√
−2i
3 −3i
−4i
−7i
−8i
−9i
GRAPHICAL VIEW
−10i
Let us calculate the total response y(t) for this system and verify the analysis.
It is verfied from the response that the steady state response has a frequency
of 7 rad/sec with amplitude of 0.0192 and phase shift of 2.5058 rad. Notice
PRELUDE: Figure below shows the general form of a stable LTI system output
response consisting of two parts transient state and steady state respectively.
y(t)
0.02 TRANSIENT SS
0.015
0.01
0.005
0 t
1 2 3 4 5 6
−0.005
−0.01
−0.015
H(s)
G1 G2 G1 G2
G1
+
2
G
+
1
G
+
G2
+ +
G1 G1
− −
G1
G1 G2 G1 G2
86
+ E Y
U G1
−
G2
Y = G1 E
E = U − G2 Y
Y = G1 U − G1 G2 Y
Y + G1 G2 Y = G1 U
Y G1
=
U 1 + G 1 G2
G3
−
U + E + Y
G1
+
G2
+
V
87
V =0
E = U − G3 Y
Y = G 1 E + G2 U
Y = G1 (U − G3 Y ) + G2 U
Y = G 1 U + G 2 U − G1 G3 Y
Y + G1 G3 Y = G1 U + G2 U
Y G1 + G2
=
U 1 + G1 G3
G1 −G3
+
V + Y
G2
Y = G 2 V − G1 G3 Y
Y (1 + G1 G3 ) = G2 V
Y G2
=
V 1 + G1 G3
PLANT INVERSION:
s+a ···
··· s+a
88
1
H(s) =
s+a
IMPULSE RESPONSE:
h(t) = e−at
1
a=1
0.9 a=5
0.8
a = 10
0.7
0.6
h(t)
0.5
0.4
0.3
0.2
0.1
0
0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5 5.5 6
t
STEP RESPONSE:
1 1
y(t) = − e−at
a a
89
1
a=1
0.9 a=5
0.8
a = 10
0.7
0.6
y(t)
0.5
0.4
0.3
0.2
0.1
0
0 1 2 3 4 5 6 7 8 9
t
PROPORTIONAL COMPENSATOR:
U + 1 Y
K
s+a
−
K
Y s+a K
= K
=
U 1 + s+a s + (a + K)
90
jω
a > 0, K > 0
jω
a > 0, K < 0
U + 1 Y
s+a
−
K
91
1
Y s+a 1
= 1 =
U 1+ s+a
s + (a + k)
SECOND ORDER SYSTEM:
A stable LTI second order system has a characteristic equation which has two
roots (poles) in the left half s-plane.
jω
ω ωd
n
θ
σ
ωn2 = σ 2 + ωd2
σ
sin (θ) =
ωn
where θ is measured counterclockwise with respect to positive jω axis.
σ
sin (θ) = =ζ
ωn
σ = ζωn
ωd2 = ωn2 − σ 2 = ωn2 − ζ 2 ωn2
√
ωd = ωn 1 − ζ 2
s2 + 2σs + σ 2 + ωd2
When finding the transfer function from second order differential equation
we typically write the result in the polynomial form
ωn2
H(s) =
s2 + 2ζωn s + ωn2
σ
The characteristic equation with the value of damping ratio ζ = ωn
= sin (θ)
classify the LTI system into four categories i.e.
π
2. Underdamped System: 0 < θ < 0 < ζ < 1 the roots of the char-
,
2
p
acteristic equation are complex i.e. s1,2 = −ζωn ± j(ωn 1 − ζ 2 ).
π
3. Critically damped System: θ = 2
, ζ = 1 the roots of the characteristic
equation are repeated and real i.e. s1,2 = −ωn .
jω
UNDAMPED
θ=0
UNDERDAMPED
ζ=0
0 < θ < π2
0<ζ<1
σ
CRITICALLY DAMPED ω ωn
θ = π2 n ωd
ζ=1 θ
σ
ζ>1
OVERDAMPED
94
IMPULSE RESPONSE:
ωn2
H(s) =
s2 + 2ζωn s + ωn2
ωn2
=
s2 + 2ζωn s + ζ 2 ωn2 − ζ 2 ωn2 + ωn2
ωn2
=
(s + ζωn )2 + ωn2 (1 − ζ 2 )
ωd2
1−ζ 2
=
(s + σ)2 + ωd2
ωd ωd
= ×
1 − ζ2 (s + σ)2 + ωd2
ωd
h(t) = p p e−σt sin ωd t
1−ζ 1−ζ
2 2
ωn
= p e−σt sin ωd t
1−ζ 2
95
p
ωn −ζωn t
h(t) = p e sin 1 − ζ 2 ωn t
1 − ζ2
Figure below plots h(t) for several values of ζ such that time has been nor-
malized to the undamped natural frequency ωn . Note the actual frequency
ωd decreases slightly as the damping ratio increases.
p
h(t) = √ωn 2 e−ζωn t sin 1− ζ2ω nt
1−ζ
1
ζ =0
0.8 ζ = 0.1
ζ = 0.2
0.6 ζ = 0.3
ζ = 0.4
0.4 ζ = 0.5
ζ = 0.6
0.2 ζ = 0.7
ζ = 0.8
h(t)
0 ζ = 0.9
−0.2
−0.4
−0.6
−0.8
−1
0 2 4 6 8 10 12
ωn t
96
STEP RESPONSE:
ωd2 + σ 2
=
s(s + σ − jωd )(s + σ + jωd )
c1 c2 c2
= + +
s s + σ − jωd s + σ + jωd
ωd2 + σ 2
c1 = sY (s) =
s=0 (s + σ − jωd )(s + σ + jωd )
s=0
2
ωd2 + σ 2 ω
d +σ
2
= = 2
=1
(σ − jωd )(σ + jωd ) σ
+ ωd2
c1 = 1
97
ωd2 + σ 2
c2 = (s + σ − jωd )Y (s) =
s=−σ+jωd s(s + σ + jωd )
s=−σ+jωd
ωd2 + σ 2
=
σ
(−σ + jωd )(− σ
+ jωd )
+ jωd +
ωd2 + σ 2
=
(−σ + jωd )(j2ωd )
ωd2 + σ 2
=
−2ωd2 − j2σωd
1 ωd2 + σ 2 −ωd + σ
= × ×
2ωd −ωd − σ −ωd + σ
2
2
1 ω
d +σ
= × 2
2
× (−ωd + σ)
2ωd ω
d +σ
1 σ 1 σ
c2 = − + j , c2 = − − j
2 2ωd 2 2ωd
" #
1
1
2
− j 2ω
σ 1
2
σ
+ j 2ω
Y (s) = − d
+ d
s s + σ − jωd s + σ + jωd
98
s
−σt 1 σ2
y(t) = 1 − 2 · e · + · cos (ωd t − θ)
4 4ωd2
s
−σt 1 σ2
=1−
2·e · 1+ · cos (ωd t − θ)
2
ωd2
σ2 ζ 2 ωn2
1+ =1+
ωd2 ωn2 (1 − ζ 2 )
ζ2 1 − ζ2 + ζ2 1
1+ = =
1 − ζ2 1 − ζ2 1 − ζ2
! !
− 2σωd σ
θ = tan−1
1 = − tan−1
ωd
2
!
ω
ζ n
= − tan−1 p
ω
n 1 − ζ2
!
ζ
θ = − tan−1 p
1 − ζ2
p
1
y(t) = 1 − e−ζωn t p cos 1 − ζ 2 ωn t + θ
1 − ζ2
99
p
y(t) = 1 − e−ζωn t √ 1 2 cos 1− ζ 2 ωn t +θ
1−ζ
2
ζ =0
1.8 ζ = 0.1
ζ = 0.2
1.6 ζ = 0.3
ζ = 0.4
1.4 ζ = 0.5
ζ = 0.6
1.2 ζ = 0.7
ζ = 0.8
y(t)
1 ζ = 0.9
0.8
0.6
0.4
0.2
0
0 2 4 6 8 10 12
ωn t
100
ω
2 e−at sin ωt
(s + a)2 + ω 2
y(t)
±1%
Mp
1
0.9
0.1
t
tr
tp
ts
RISE TIME:
!
p p
y(t) = 1 − e−ζωn t cos ωn 1 − ζ t + √
2 ζ
· sin ωn 1 − ζ 2 t
1−ζ 2
1.2
ζ = 0.5
0.8
y(t)
0.6
0.4
0.2
≈ 1.8
0
0 1 2 3 4 5 6 7 8 9 10 11 12
ωn t
If we consider the curve for ζ = 0.5 to be an average, the rise time from y
= 0.1 to y = 0.9 is approximately ωn tr = 1.8. Thus, we can say rise time
tr is equal to .
1.8
tr ≈
ωn
103
jω
ωn = 4.5
ωn = 3
ωn = 1.5
ωn = 4.5
ω n=1.5
ωn = 3
OVERSHOOT:
!
ζ
y(t) = 1 − e−ζωn t cos (ωd t) + p · sin (ωd t)
1 − ζ2
!
dy ζ
= ζωn e−ζωn t cos (ωd t) + p · sin (ωd t)
dt 1 − ζ2
!
ζω d
+ e−ζωn t ωd sin (ωd t) − p · cos (ωd t)
1 − ζ2
104
dy ζωd ζ 2 ωd
= p e−ζω
n t
p
cos (ωd tp ) + e−ζωn tp sin (ωd tp )
dt 1−
ζ2 1− ζ2
t=tp
ζωd
+ e−ζωn tp ωd sin (ωd tp ) − p
e−ζω
n t
cos (ωd tp )
p
1−
ζ 2
!
dy ζ 2
= e−ζωn tp ωd sin (ωd tp ) +1
dt 1 − ζ2
t=tp
!
ζ2 + 1 − ζ2
= e−ζωn tp ωd sin (ωd tp )
1 − ζ2
!
ωd
= e−ζωn tp sin (ωd tp ) p p
1 − ζ 1 − ζ2
2
ωn
= (sin (ωd tp )) p e−ζωn tp = 0
1− ζ2
sin (ωd tp ) = 0, ω d tp = π
π
tp =
ωd
105
Mp = y(tp ) − 1
!
−ζωn ωπ ζ
= −e d cos (π) + p · sin (π)
1 − ζ2
ω
−ζ n √π
ω
1−ζ 2
=e n
− √ ζπ
1−ζ 2
=e
− √ ζπ
The maximum percent overshoot is e 1−ζ 2
× 100%.
Mp × 100
100
90
80
70
60
50
40
30
20
16%
10
5%
ζ
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
106
jω
ζ = 0.1
ζ=
ζ
0.5
=
0.
ζ=
7
0.9
SETTLING TIME:
The settling time as that value of ts when the decaying exponential reaches
1%.
e−ζωn ts = 0.01
ln
e −ζωn ts = ln 0.01
−ζωn ts = −4.6
107
4.6 4.6
ts = =
ζω σ
jω
|σ| = 1.75
|σ| = 0.5
|σ| = 3
|σ| = 3
ωn = 3
ζ=
ζ
0.9
=
|σ| = 1.75 0.
7 ζ=
ω n=1.5
0.5
ω
n
σ
jω
jω
ωd
σ
σ
108
G
O
O
D
B
A
D
σ
109
110
jω
σ
ω ωd
n θ
σ
0.06
0.05
0.04
0.03
0.02
0.01
0
0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5 5.5 6
STEP RESPONSE OF STABLE SECOND ORDER LTI SYSTEM WITHOUT ANY RHP ZERO
111
0.15
0.1
0.05
−0.05
−0.1
−0.15
0 1 2 3 4 5 6 7
STEP RESPONSE OF STABLE SECOND ORDER LTI SYSTEM WITH A RHP ZERO
EXAMPLE:
τ
θ
112
τ (t) = J θ̈(t)
T (s) = Js2 Θ(s)
h(t)
δ(t)
1
0 Js2 t
jω
Θc + T Θ
1
sKv + Kp
J s2
−
113
sKv +Kp
Θ J s2
= sKv +Kp
Θc 1 + Js 2
sKv + Kp
=
Js2 + sKv + Kp
( ) Kp
Kv J Kp J
= s+ Kv Kp
J Kp KV s2 + J
s + J
( ) Kp
Kv Kp J
= s+ Kv Kp
Kp KV s2 + J
s + J
Kv Kp
s2 + 2σs + ωd2 + σ 2 = s2 + s+
J J
Kp
ωn2 = ωd2 + σ 2 =
√ J
Kp
ωn =
J
Kv
2σ =
J
Kv
σ = ζωn =
2J
114
Kp
Kv = 0, s2 +
J
Kv Kp
Kp = fixed, s2 + s+
J J
Kv
2J √
Kp
√
√ Kp Kv2 J
K −
J 4J 2
J
p θ
σ
√
Kp
−
J
STABILITY:
The LTI system stability criterion is
ALL THE POLES OF THE SYSTEM SHOULD BE IN THE LEFT HALF PLANE OF
COMPLEX DOMAIN. i.e.
Re(pi ) < 0, ∀ pi
sn 1 a2 a4 ···
sn−1 a1 a3 a5 ···
sn−2 α β ···
sn−3 γ ···
1 a2
det
a1 a3
α=−
a1
116
1 a4
det
a1 a5
β=−
a1
a1 a3
det
α β
γ=−
α
After completing the pattern for the general characteristic equation, extract
the first column i.e.
1
a
1
α
γ
..
.
The number of sign changes in this column is equal to the number of poles in
right half plane.
There is a necessary condition before applying this method, if all the co-
efficients of the characteristic equation are present with positive sign then
there may or may not be right half plane poles. But if there are some zero
coefficients or with negative values then it is sure that all the poles are not
in the left half plane.
117
EXAMPLE:
∆(s) = s4 + s3 + 2s2 + 3s + 4
s4 1 2 4 0
s3 1 3 0 0
s2 α β
s γ
1 2
det
1 3
α=− = −1
1
1 4
det
1 0
β=− =4
1
118
1 3
det
−1 4
γ=− =7
−1
1
1
−1
7
There are two sign changes + to − from second row to third row and then
− to + from third to fourth row. So there are two right half plane poles. The
roots of the characteristic equation are
s1 = −1 + j0.83
s2,3 = 0.5 ± j1.94
The pattern is to check the stability of the second order system then check the
steady state value using final value theorem for the step response. The time
domain specifications are interpreted to Laplace domain pole locations for
desired step response. The design of the system is dictated by the dominant
poles which are close to the imaginary axis as they have major role in the
transient response. Zeros are considered for any specific lows in the desired
response.
119
EXAMPLE:
U (s) Y (s)
1
s(s+2)
Pole-Zero Map
1 0.84 0.75 0.66 0.55 0.43 0.310.17
0.90 sys
0.90
0.8
0.4
0.98
0.30
Imaginary Axis
0.2
0 0 rad/s
−0.2
0.30
0.98
−0.4
−0.6 0.60
0.95
−0.8
0.90
0.90 0.84 0.75 0.66 0.55 0.43 0.310.17
−1
−2 −1.8 −1.6 −1.4 −1.2 −1 −0.8 −0.6 −0.4 −0.2 0 0.2
Real Axis
120
Step Response
sys
1.2
0.8
y(t)
0.6
0.4
0.2
0
0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2 2.2 2.4 2.6 2.8 3
t
π2ζ 2
= 5.3019
1 − ζ2
π 2 ζ 2 = 5.3019 − 5.3019ζ 2
ζ 2 (π 2 + 5.3019) = 5.3019
15.1715ζ 2 = 5.3019
ζ 2 = 0.3495
ζ = 0.5912 = 0.6, ∵ζ>0
U + 1 Y
K s(s+2)
−
K
Y s(s+2)
= K
U 1 + s(s+2)
K
=
s2 + 2s + K
∆(s) = s2 + 2s + K = s2 + 2ζωn s + ωn2
ωn2 = K
2ζωn = 2
ζ = 0.6
ωn = 1.6667 ≈ 1.7 ̸= 1.8
K = 2.89
122
Y 2.89
=
U s2 + 2s + 2.89
Pole-Zero Map
2
sys
1.60
1.5 0.55 0.43 0.31 0.17 1.40
0.66 1.20
1 0.75 1.00
0.80
0.84
0.60
0.5 0.90
Imaginary Axis
0.95 0.40
0.98 0.20
0 0 rad/s
0.98 0.20
0.95 0.40
−0.5 0.90
0.60
0.84
0.80
−1 0.75 1.00
0.66 1.20
0.55 0.43 0.31 0.17 1.40
−1.5
1.60
−2
−2 −1.8 −1.6 −1.4 −1.2 −1 −0.8 −0.6 −0.4 −0.2 0
Real Axis
123
Step Response
sys
0.8
0.6
y(t)
0.4
0.2
0
0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5
t
Mp = 0.102, tr = 1.07s
EXAMPLE:
U (s) Y (s)
s+1
s(s−1)(s+6)
Pole-Zero Map
2
sys
1.5
0.5 0.50
Imaginary Axis
0 0
−0.5 0.50
−1.5
−2
−7 −6 −5 −4 −3 −2 −1 0 1 2
Real Axis
The target is to make the system stable using proportional compensator with
unity feedback.
U + s+1 Y
K s(s−1)(s+6)
−
125
K(s+1)
Y s(s−1)(s+6)
= K(s+1)
U 1 + s(s−1)(s+6)
K(s + 1)
=
s(s − 1)(s + 6) + K(s + 1)
s3 1 K−6 0 0
s2 5 K 0 0
s1 α β
s0 γ
126
1 K − 6
det
5 K 4K − 30
α=− =
5 5
β=0
5 K
det
5
4K−30
0
γ=− 4K−30
=K
5
1
5
4K−30
5
K
K>0
4K − 30
>0
5
K > 7.5
K=8
Y 8(s + 1)
=
U s3 + 5s2 + 2s + 8
Pole-Zero Map
1.5
0.95 0.90 0.84 0.75 0.66 0.55 0.430.310.17
sys
1.20
1
0.98
0.60
0.5
Imaginary Axis
0 0 rad/s
−0.5
0.60
0.98
−1
1.20
SYSTEM TYPE:
OPEN LOOP SYSTEM:
Defining G(s) as
K(Ta s + 1)(Tb s + 1) · · · (Tm s + 1)
G(s) =
sN (T1 s + 1)(T2 s + 1) · · · (Tp s + 1)
U Y
K PLANT
128
U (s) Y (s)
G(s)
G(s) is open loop transfer function of a unity feedback system. It involves the
term sN in the denominator, representing a pole of multiplicity N at the ori-
gin. The present classification scheme is based on the number of integrations
indicated by the open-loop transfer function. A system is called type 0, type
1, type 2, · · · , if N = 0, N = 1, N = 2, · · · , respectively. Note that this
classification is different from that of the order of a system. As the type num-
ber is increased, accuracy is improved; however, increasing the type number
aggravates the stability problem. A compromise between steady-state accu-
racy and relative stability is always necessary. If G(s) is written so that each
term in the numerator and denominator, except the term sN , approaches
unity as s approaches zero, then the open loop gain K is directly related to
the steady-state error.
+ E Y
U G(s)
−
129
Y =E×G
E =U −Y
E =U −E×G
E(1 + G) = U
E 1
=
U 1+G
U + E
G(s)
E 1
=
U 1+G
1
E(s) = U (s)
1 + G(s)
1
lim e(t) = ess (t) = lim sE(s) = lim s U (s)
t→∞ s→0 s→0 1 + G(s)
130
Next the analysis depends on the input U (s) for the steady state error of the
unity feedback system corresponding to the open loop transfer function G(s).
STATIC POSITION ERROR CONSTANT: Considering the steady state error of
the unity feedback system for unit step input i.e.
u(t) = 1(t)
1
U (s) =
s
1 1
ess (t) = lim
s
s→0 1 + G(s)
s
1
=
1 + G(0)
The static position error constant Kp is defined as
Type 1:
K(Ta s + 1)(Tb s + 1) · · · (Tm s + 1)
G(s) =
s1 (T1 s + 1)(T2 s + 1) · · · (Tp s + 1)
Kp = G(0) = ∞
ess (t) = 0
Type 2:
K(Ta s + 1)(Tb s + 1) · · · (Tm s + 1)
G(s) =
s2 (T1 s + 1)(T2 s + 1) · · · (Tp s + 1)
Kp = G(0) = ∞
ess (t) = 0
EXAMPLE:
1
G(s) = , Type 0
(s + 2)(s + 3)
1
Kp = K = G(0) =
6
1 6
ess (t) = =
1+K 7
6 1
yss (t) = 1 − = = 0.143, u(t) = 1(t)
7 7
Note: The steady state error and output response value is for the unity feed-
back system.
132
Step Response
0.2
syscl
0.18
0.16
0.14
0.12
y(t)
0.1
0.08
0.06
0.04
0.02
0
0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2
t
1
G(s) = , Type 1
s(s + 2)(s + 3)
Kp = K = G(0) = ∞
ess (t) = 0
yss (t) = 1, u(t) = 1(t)
133
Step Response
1
0.8
0.6
y(t)
0.4
0.2
0
0 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30
t
u(t) = t, t≥0
= 0, t<0
1
U (s) = 2
s
1 1
ess (t) = lim
s
s→0 1 + G(s) s2
1
= lim
s→0 sG(s)
134
Kv = lim sG(s)
s→0
1
ess (t) =
Kv
Now let us find the value of Kv and steady state error for types 0, 1 and 2
systems.
Type 0:
K(Ta s + 1)(Tb s + 1) · · · (Tm s + 1)
G(s) =
s0 (T1 s + 1)(T2 s + 1) · · · (Tp s + 1)
Kv = lim sG(s) = 0
s→0
ess (t) = ∞
Type 1:
K(Ta s + 1)(Tb s + 1) · · · (Tm s + 1)
G(s) =
s1 (T1 s + 1)(T2 s + 1) · · · (Tp s + 1)
Kv = lim sG(s) = K
s→0
1
ess (t) =
K
Type 2:
K(Ta s + 1)(Tb s + 1) · · · (Tm s + 1)
G(s) =
s2 (T1 s + 1)(T2 s + 1) · · · (Tp s + 1)
Kv = lim sG(s) = ∞
s→0
ess (t) = 0
135
EXAMPLE:
1
G(s) = , Type 1
s(s + 2)(s + 3)
1
Kv = K = lim sG(s) =
s→0 6
1
ess (t) = = 6, u(t) = t
K
80
60
y(t)
40
20
0
0 10 20 30 40 50 60 70 80 90 100
t
ror of the unity feedback system for unit parabolic input i.e.
t2
u(t) = , t≥0
2
= 0, t<0
1
U (s) = 3
s
1 1
ess (t) = lim
s 2
s→0 1 + G(s) s3
7
1
= lim
s→0 s2 G(s)
The static acceleration error constant Ka is defined as
Ka = lim s2 G(s)
s→0
1
ess (t) =
Ka
Now let us find the value of Ka and steady state error for types 0, 1 and 2
systems.
Type 0:
K(Ta s + 1)(Tb s + 1) · · · (Tm s + 1)
G(s) =
s0 (T1 s + 1)(T2 s + 1) · · · (Tp s + 1)
Ka = lim s2 G(s) = 0
s→0
ess (t) = ∞
137
Type 1:
K(Ta s + 1)(Tb s + 1) · · · (Tm s + 1)
G(s) =
s1 (T1 s + 1)(T2 s + 1) · · · (Tp s + 1)
Ka = lim s2 G(s) = 0
s→0
ess (t) = ∞
Type 2:
K(Ta s + 1)(Tb s + 1) · · · (Tm s + 1)
G(s) =
s2 (T1 s + 1)(T2 s + 1) · · · (Tp s + 1)
Ka = lim s2 G(s) = K
s→0
1
ess (t) =
K
EXAMPLE:
(s + 1)(s + 3)
G(s) = , Type 2
s2 (s + 2)(s + 3)
3 1
Ka = K = lim s2 G(s) = =
s→0 6 2
2
1 t
ess (t) = = 2, u(t) =
K 2
138
180
160
140
120
y(t)
100
80
60
40
20
0
0 2 4 6 8 10 12 14 16 18 20
t
90
80
70
y(t)
60
50
40
30
20
PROPORTIONAL-INTEGRAL-DERIVATIVE CONTROLLER
+
Xc + E U X
K G
−
Increased oscillations
No energy absorbing mechanism
Correction after the error is injected in the system.
Assume K(s) = Kp + sKd i.e. proportional and derivative controller
Absorbs energy.
Preempt the error.
f
m
d
mẍ = f + d
f = −Kp x − Kd ẋ
mẍ = −Kp x − Kd ẋ + d
ms2 X(s) = −Kp X(s) − Kd sX(s) + D(s)
ms2 X(s) = −X(s)(sKd + Kp ) + D(s)
X(s) D(s)
X(s) = − 2
(sKd + Kp ) +
ms[ ms2 ]
1
= − X(s)(sKd + Kp ) + D(s)
ms2
142
+
X
1
sKd + Kp
ms2
−1
+
Xc = 0 X
1
sKd + Kp
ms2
+
−
143
1 Kd s + Kp
= × Kd Kp
m s2 + m
s + m
Kd K
m
s + mp
= K
s2 + Km
d
s + mp
Kd Kp
∆(s) = s2 + s+
m m
The closed loop system is second order, comparing the characteristic polyno-
mial with the general form of second order system we obtain
Kd Kp
s2 + 2ζωn s + ωn2 = s2 + s+
m m
Kp
ωn2 =
m
√
Kp
ωn =
m
Kd
2ζωn =
m
Kd
ζ=
2mωn
144
√
Kp
√ m
K
m p θ
σ
√
Kp
−
m
Using the values of Kp and Kd , the poles can be placed anywhere in the
left half plane.
Next is the error analysis with respect to desired output and disturbance in
order to see the role of Integral part of the controller.
145
+
Xc = 0 E X
1
sKd + Kp
ms2
+
−
1
= sKd +Kp
1+ ms2
ms2
=
ms2 + Kd s + Kp
From the G(s)
sKd + Kp
G(s) =
ms2
It can be observed that it is a system type 2, so the closed loop system can
track step, ramp without any error and parabola with a constant error in
absence of disturbance D .
146
E =0−X
E = −X
1
X= [E(sKd + Kp ) + D]
ms2
1
−E = [E(sKd + Kp ) + D]
ms2
−ms2 E = E(sKd + Kp ) + D
−E(ms2 + Kd s + Kp ) = D
E 1
=−
D ms2 + Kd s + Kp
It can be observed from the transfer function, it is a type 0 with reference
to disturbance input, so for step disturbance there will be a constant error in
steady state output value of X , so the steady state output of the closed loop
system x(t) for step disturbance d(t) = 1(t) with no reference input xc (t)
= 0 will be
xss (t) = −ess (t)
The steady state error can be decreased by increasing the value of Kp be-
cause for step disturbance with amplitude D
D
ess (t) = E =
Kp
For example if we want the absolute error for step disturbance is less than 0.1,
it requires the value of Kp greater than 10. But increasing Kp will induce
oscillations.
147
In order to eradicate this error we need to increase the type by using integral
part of the controller.
Ki
s
D
+ +
Xc = 0 U X
1
sKd + Kp
E ms2
+
−
Ki Kd s2 + Kp s + Ki
K(s) = + Kd s + Kp =
s s
+
Xc = 0 E X
Kd s2 + Kp s + Ki 1
s ms2
+
−
148
E =0−X
E = −X
[ ( ) ]
2
1 Kd s + Kp s + Ki
X= E +D
ms2 s
[ ( ) ]
2
1 Kd s + Kp s + Ki
−E = E +D
ms2 s
( )
Kd s2 + Kp s + Ki
−ms2 E = E +D
s
( )
Kd s2 + Kp s + Ki
−E ms2 + =D
s
E s
=−
D ms3 + Kd s2 + Kp s + Ki
The type has increased from 0 to 1 so the steady state error will be 0 for this
case. Therefore the goal of rejecting the step disturbance is achieved.
Normally in designing process Kp is used to achieve the steady state value
or be in the proximity of the value with a realistic value then integral part is
used to remove any error. Integral gain is one hundredth of the proportional
gain in order to tackle the issues of stability.
149
ANALYSIS:
U + E K Y
s+2
−
1
G(s) =
s+2
K(s) = K
K
Y GK s+2
= = K
U 1 + GK 1 + s+2
K
=
s + (2 + K)
∆(s) = 1 + GK = s + (2 + K) = 0
s = −5, K = 3
s = −10, K = 8
151
jω
σ
−5 −2
Any input (u(t), d(t), etc) for a particular closed loop LTI system
∆(s) = 1 + G(s)K(s)
will remain same. Closed loop poles are the roots of ∆(s). Therefore the
location of closed loop poles always satisfies the following equation for
specific values of s = σ + jω
G(s)K(s) = −1
The path generated by changing the parameter value K for the closed loop
152
system is the root locus. So we are generating root locus from open loop
transfer function for the closed loop transfer function.
U + Y
K G(s)
−
−0.01
−1
153
jω
ω
σ + jω σ
G(s), s = jω, s = σ + jω
KM = 1, ∵ |ejϕ | = 1
1
K=
M
ϕ = G(s) = ±180◦ , ∵ K =0
s=σ+jω
jω
s
M
ϕ
σ
155
1
G(s) =
s+2
1 ◦ ◦
KG(s) = K ej0 ̸= 1ej(±180 )
2
s=0+j0
K
KG(s) =
2 + j2
s=0+j2
1 ◦ ◦
= K √ e−j45 ̸= 1ej(±180 )
8
K K
KG(s) = =
−3 + j3 + 2 −1 + j3
s=−3+j3
◦ ◦
= K0.32e−j108.4 ̸= 1ej(±180 )
156
K K
KG(s) = =
−5 + j0 + 2 −3 + j0
s=−5+j0
1 ◦ ◦ 1 1
= K e−j180 = 1ej(±180 ) , when K = = 1 =3
3 M 3
Im(s)
a + jb
c + jd
b2
+
M
a2
√
b
1
√
2
ϕ2
d
c2
+ ϕ1
d2
Re(s)
c a
s-domain, s = σ + jω
s1 = a + jb = M1 ejϕ1
s2 = c + jd = M2 ejϕ2
p p
M1 = a2 + b2 , M2 = c2 + d2
! !
b d
ϕ1 = tan−1 , ϕ2 = tan−1
a c
157
s1 M1 ejϕ1 M1
= = ej(ϕ1 −ϕ2 )
s2 M2 ejϕ2 M2
s1 × s2 = (M1 ejϕ1 )(M2 ejϕ2 ) = (M1 M2 )ej(ϕ1 +ϕ2 )
1 1 jϕ1 −1 1 −jϕ1
= = (M 1 e ) = e
s1 M1 ejϕ1 M1
1. Guess s
3. | · | → abs(·)
4. ϕ → angle(·)
5. if ϕ = ±180◦ then K = 1
|·|
158
ZERO:
jω
ω2
+
a2
ω
√
−1 ω
tan
a
σ
a
−a
G(s) = s + a
G(s) = G(jω) = a + jω
s=jω
p
|G(jω)| = a2 + ω 2
!
ω
G(jω) = tan−1
a
159
jω
ω2
+
)2
σ
ω
−
(a
p
ϕ
σ
a−σ
−a
G(s) = s + a
G(s) = G(−σ + jω) = a − σ + jω
s=−σ+jω
q
|G(−σ + jω)| = (a − σ)2 + ω 2
!
ω
ϕ = G(−σ + jω) = tan−1
a−σ
160
POLE:
jω
ω2
+
)2
σ
ω
−
(a
p
ϕ
σ
a−σ
−a
1
G(s) =
s+a
1
G(s) = G(−σ + jω) =
a − σ + jω
s=−σ+jω
1
|G(−σ + jω)| = p
(a − σ)2 + ω 2
!
ω
ϕ = G(−σ + jω) = − tan−1
a−σ
161
jω
lp
lz
ϕp ϕz
σ
Gz (s) = lz ejϕz
s=−σ+jω
Gp (s) = lp ejϕp
s=−σ+jω
lz
G(s) = ej(ϕz −ϕp )
lp
s=−σ+jω
162
jω
l4
l3 l2 l1
ϕ4 ϕ2 ϕ1
ϕ3
σ
l1 l2
G(s) = ej(ϕ1 +ϕ2 )−(ϕ3 +ϕ4 )
l3 l4
s=−σ+jω
l1 l2
|G(−σ + jω)| =
l3 l4
G(−σ + jω) = ϕ1 + ϕ2 − ϕ3 − ϕ4
163
jω
l
−a + jb ω−b
ϕ
a−σ
l
ω+b
−a − jb ϕ
a−σ
1
G(s) = e−j(ϕ+ϕ)
ll
s=−σ+jω
1
|G(−σ + jω)| =
ll
1
G(−σ + jω) =
ej(ϕ+ϕ)
164
∆(s) = 1 + K0 K(s)G(s)H(s) = 0
K0 K(s)G(s)H(s) = −1
N (s)
K(s)G(s)H(s) =
D(s)
N (s)
K0 = −1, Evan's Form
D(s)
U Y
K0 K(s) G(s)
H(s)
Root locus starts at open loop poles and ends at open loop zeros
jω
N (s)
∆(s) = 1 + K0 =0
D(s)
D(s) + K0 N (s) = 0
K0 = 0 → D(s) = 0, Open loop poles
K0 = ∞ → N (s) = 0, Open loop zeros
166
Points on the real axis to the left of an odd number of poles and zeros are
on the locus
jω
The phases of Conjugate pair poles do not contribute in the phase calcula-
tion of a point on the real axis
167
jω
−ϕ
The asymptotes divide a circle into equal sectors along with symmetric
around real axis.
s+6
G(s) =
(s + 2)(s + 4)(s + 8)
asymptotes
locus
10
open loop poles
open loop zero
5
Imaginary Axis
−5
−10
1
G(s) =
(s + 1)4
1 asymptotes
locus
0.8 open loop poles
0.6
0.4
Imaginary Axis
0.2
−0.2
−0.4
−0.6
−0.8
−1
−2 −1.8 −1.6 −1.4 −1.2 −1 −0.8 −0.6 −0.4 −0.2 0
Real Axis gain = [0, 3.86207]
170
jω
ϕd
ϕp ϕz
σ
ϕc
ϕz − ϕd − ϕp − ϕc = ±180◦
ϕc = 90◦
EXAMPLE:
1
G(s) =
(s + 1)3
−3ϕd = −180◦
ϕd = 60◦
0.5
Imaginary Axis
−0.5
−1
−1.5
−2
−3 −2.5 −2 −1.5 −1 −0.5 0
Real Axis gain = [0, 9.31034]
172
EXAMPLE:
Draw an approximate root locus for the following open loop transfer function
using the aforementioned rules.
s2 + 16
G(s) =
s3 (s + 3)
Pole-Zero Map
5
sys
2
Imaginary Axis
−1
−2
−3
−4
−5
−4 −3.5 −3 −2.5 −2 −1.5 −1 −0.5 0 0.5 1
Real Axis
173
Real Axis
Pole-Zero Map
5
sys
2
Imaginary Axis
−1
−2
−3
−4
−5
−4 −3.5 −3 −2.5 −2 −1.5 −1 −0.5 0 0.5 1
Real Axis
174
Number of Asymptotes = 4 − 2 = 2
−3 − 0
α= = −1.5
2
180◦
ϕ1 = = 90◦
2
540◦
ϕ2 = = 270◦ = −90◦
2
Pole-Zero Map
5
sys
2
Imaginary Axis
−1
−2
−3
−4
−5
−4 −3.5 −3 −2.5 −2 −1.5 −1 −0.5 0 0.5 1
Real Axis
175
Pole-Zero Map
5
sys
ϕa1
4
2
Imaginary Axis
1
ϕd1 ϕd2,3,4
0
−1
−2
−3
ϕa2
−4
−5
−4 −3.5 −3 −2.5 −2 −1.5 −1 −0.5 0 0.5 1
Real Axis
Angle of Departures
!!
!!
4
4
◦
−ϕd1 + 180+
tan
−1
− 180◦+
tan
−1
− (3 × 180◦ )
3 3
◦
= −180
ϕd1 = −360◦ = 0◦
−(3 × ϕd2 ) −0◦ −
90◦
+
90◦
= −180◦ + 360◦ (l − 1), l=1
ϕd2 = 60◦
176
−(3 × ϕd3 ) −0◦ −
90◦
+
90◦
= −180◦ + 360◦ (l − 1), l=2
ϕd3 = −60◦
−(3 × ϕd4 ) −0◦ −
90◦
+
90◦
= −180◦ + 360◦ (l − 1), l=3
ϕd4 = −180◦
Angle of Arrivals
!
4
ϕa1 + 90◦ − (3 × 90◦ ) − tan−1 = −180◦
3
ϕa1 = 53.1◦
ϕa2 = −53.1◦
177
asymptotes
locus
open loop poles
10
open loop zeros
5
Imaginary Axis
−5
−10
The roots of the above equation gives the potential breakaway or break-in
points.
D(s)
K=−
N (s)
dK D ′ (s)N (s) − D(s)N ′ (s)
=−
ds N 2 (s)
dK
=0
ds
= D(s)N ′ (s) − D ′ (s)N (s) = 0
POINTS WHERE THE ROOT LOCI MAY CROSS THE IMAGINARY AXIS: Let s
= jω in the closed loop characteristic equation, equate both the real part
and the imaginary part to zero, and then solve for possible ω and K .
179
APPLIED EXAMPLE:
f b k
AM P LOGIC
M ẍ = −bẋ − kx + f
M ẍ + bẋ + kx = f
M s2 X(s) + bsX(s) + kX(s) = F (s)
X 1
=
F M s2 + bs + k
1
M
= b k
s2 + M
s + M
b = SMALL, k = Large
b k
s2 + s+ = s2 + 2ζωn s + ωn2
M M s
k
ωn = , ζ→0
M
jω
45◦
Xc = 0 + E F X
K(s) G(s)
−
181
X GK
=
Xc 1 + GK
1 + GK = 0
GK = −1
|GK| = 1
GK = ±180◦
K(s) 1
K(s)G(s) = · b k
M s2 + M s+ M
f = Kp e(t) + Kd ė(t)
!
Kp
= Kd s + E(s)
Kd
!
F Kp
= K(s) = Kd s +
E Kd
Kp
Kd s+ Kd
K(s)G(s) = · b k
M s2 + M
s + M
182
locus
open loop poles
open loop zero
Imaginary Axis
Real Axis
jω
l1
l3 45◦
l2
183
|KG| = 1
l3
|K(s)G(s)| =
l1 l2
Kd l1 l2
=
M l3
MATHEMATICAL MODELLING ERROR
x2
x k2
f b k
AM P LOGIC
184
mẍ2 = k2 (x − x2 )
M ẍ = k2 (x − x2 ) − bẋ − kx + f
ms2 X2 = k2 (X − X2 )
M s2 X = k2 (X − X2 ) − bsX − kX + F
ms2 X2 + k2 X2 − k2 X = 0
−k2 X + (ms2 + k2 )X2 = 0
M s2 X = k2 X − k2 X2 − bsX − kX + F
M s2 X − k2 X + bsX + kX + k2 X2 = F
(M s2 + bs + k − k2 )X + k2 X2 = F
" #
−k2 2
ms + k2 X 0
=
M s2 + bs + k − k2 k2 X2 F
A y x
185
0 ms + k2
2
F k2
X(s) =
−k ms2
+ k 2
2
M s + bs + k − k2
2
k2
X ms2 + k2
=
F M ms4 + bms3 + (M k2 + km − k2 m)s2 + bk2 s + kk2
!
Kp
Kd s + (ms2 + k2 )
Kd
K(s)G(s) =
M ms4 + bms3 + (M k2 + km − k2 m)s2 + bk2 s + kk2
locus
open loop poles
open loop zeros
Imaginary Axis
Real Axis
186
x2
f b k
AM P LOGIC
Root Locus
asymptotes
locus
open loop poles
zero
Imaginary Axis
Real Axis
187
LEAD/LAG COMPENSATION
PROPORTIONAL:
DERIVATIVE:
K(s) = sKd
jω
INTEGRAL:
Ki
K(s) =
s
188
jω
PROPORTIONAL-DERIVATIVE:
K(s) = Kp + sKd
!
Kp
= Kd s +
Kd
jω
σ
189
CONSTRAINT: Generally, the sensor directly gives the derivative for practical
purposes otherwise calculating derivative numerically from sensor values is
not feasible especially for higher frequencies.
Pure derivative is difficult to implement for higher frequencies in real time
and it rolls off.
Every signal has noise which is generally of high frequency so derivative
is tracking a noise rather the original signal.
d
sin (ωt) → sin (ωt) = ω cos (ωt)
dt
LOW PASS FILTER:
!
a a ω
ϕ=∠ =∠ = − tan−1
s + a a + jω a
s=jω
a
A sin (ωt) A M sin (ωt + ϕ )
s+a
a a
a
M = = = √ 2
s + a jω + a a + ω2
s=jω
jω
a
M ω
ϕ
σ
−a
190
LEAD:
Proportional-Derivative with low pass filter on the derivative.
Practically implementable
!
a
K(s) = Kp + Kd s
s+a
Kp (s + a) + Kd sa
=
s+a
aKp
s+ Kp +aKd
= (Kp + aKd ) ·
s+a
aKp
b= , K = Kp + aKd
Kp + aKd
s+b
=K , where |b| < |a| or b > a
s+a
191
jω
σ
−a −b
PROPORTIONAL-INTEGRAL:
Ki
K(s) = Kp +
s
sKp + Ki
=
s
!
Ki
Kp s +
Kp
=
s
192
jω
CONSTRAINT:
Integral keep on generating the control input ignoring actuator saturation.
Integral windup and unwound issues.
Stability concerns
v(t)
EXAMPLE:
1
G(s) =
s2 + s + 9.25
193
Pole-Zero Map
4
1
Imaginary Axis
−1
−2
−3
−4
−3.5 −3 −2.5 −2 −1.5 −1 −0.5 0 0.5 1
Real Axis
1 1
= lim sG(s)U (s) = lim
s
s→0 s→0 s2 + s + 9.25
s
1
= = 0.1081
9.25
194
Step Response
0.18
0.16
0.14
0.12
0.1
y(t)
8 · 10−2
6 · 10−2
4 · 10−2
2 · 10−2
0
0 1 2 3 4 5 6 7 8 9 10
t
The target is to remove steady state error for step input. Deploying integral
controller with unity feedback to achieve the task.
K(s + 3)
K(s) =
s
Using the gain K = 1, the open loop transfer function is given by
s+3
K(s)G(s) =
s3 + s2 + 9.25s
The relation is the confirmation of system type 1. Figure below shows the pole
zero plot of the open loop system
195
Pole-Zero Map
4
1
Imaginary Axis
−1
−2
−3
−4
−3.5 −3 −2.5 −2 −1.5 −1 −0.5 0 0.5 1
Real Axis
U + E K(s+3) 1 Y
s s2 +s+9.25
−
Step Response
0.8
0.6
y(t)
0.4
0.2
0
0 2 4 6 8 10 12 14 16 18 20
t
K(s + 3)
=
s3 + s2 + 9.25s + sK + 3K
∆(s) = s3 + s2 + 9.25s + sK + 3K
10
5
Imaginary Axis
−5
−10
Hence PI controller has the stability issue, since K > 4.625 makes the system
unstable.
198
LAG:
Practically implementable
s+b
K(s) = K , where |b| > |a| or b < a
s+a
jω
σ
−b −a
U + E s+b Y
K s+a G(s)
−
199
E(s) 1
=
R(s) 1 + K(s)G(s)
1
= !
s+b
1+ K G(s)
s+a
1
R(s) =
s
ess (t) = lim e(t) = lim sE(s)
t→∞ s→0
1 1
= lim
s !
s→0 s+b s
1+ K G(s)
s+a
s+a
= lim
s→0 s + a + K(s + b)G(s)
a
=
a + KG(0)b
The parameter b can be increased to reduce the value of steady state error.
200
Real Axis
The requirement is to pass the root locus from a point in order to achieve
specific time specifications. So system poles phase angles from the desired
point and the phase angles of the lead zero and pole should sum to ±180◦
to make the root locus pass through the desired point with some gain value
K . Figure below gives the interpretation of the requirement.
201
jω
ϕ2 = 110◦ ϕ1 = 135◦
ϕp ϕz
σ
−ϕ1 − ϕ2 − ϕp + ϕz = −180◦
−135◦ − 110◦ − ϕp + ϕz = −180◦
ϕz − ϕp = 65◦
202
ϕz
−ϕ jω
p
ϕz
−ϕ
p
ϕp ϕz
σ
203
jω
ϕp
z
−
ϕ
ϕz
ϕp
−ϕ
ϕz −
ϕp ϕz
σ
204
DOMINANT POLES:
Root Locus
Imaginary Axis
Real Axis
Referring above figure, the fast response requirement compel to place the
poles on the green triangle spots, so lead (orange zero and pole) compensator
is deployed after location calculation using phase relation. The root locus
confirms that it passes through the target spot. The higher gain K value
ensure the place with the dominant real pole is still not far from the imaginary
axis. The transients of the desired spot poles will die rapidly but the decay of
the real dominant pole will be really slow. Thus the response will have major
effect by the dominant pole.
A Pole-zero cancellation for the poles in right half plane can be catas-
trophic.
205
Pole-zero cancellation for the poles in left half plane may be used ac-
cording to specification requirements.
NOTCH FILTER:
f b k
AM P LOGIC
Root Locus
Imaginary Axis
Real Axis
The placement of lead pole and zero will provide the desired specification
with practical implementation. Considering the model where sensor is on the
spring resulting two unstable poles with varying K . Analyzing the behaviour
of the system in presence of the lead compensator.
207
x2
f b k
AM P LOGIC
Root Locus
Imaginary Axis
Real Axis
The root locus clarifies that the system will become unstable with increasing
208
K in presence of lead controller. The next step is to add notch filter in the
present controller and observe the root locus.
Pole-Zero Map
ωz
Imaginary Axis
0 rad/s
Real Axis
EXAMPLE:
1 1
G(s) = =
s2 + s − 2 (s + 2)(s − 1)
Root Locus
3
2.83
0.
71
2
1
Imaginary Axis
−1
−2
7 1
0.
2.83
−3
−3 −2.5 −2 −1.5 −1 −0.5 0 0.5 1 1.5
Real Axis gain = [0, 22.5]
In order to pass the root locus from the desired point, finding the compensation
angles i.e.
angle restriction of 56.3◦ . The location of the lead pole comes out to be s =
−3.1424
Pole-Zero Map
3
1 29.7◦ 26.6◦
146.3◦
Imaginary Axis
−1
−2
−3
−4 −3.5 −3 −2.5 −2 −1.5 −1 −0.5 0 0.5 1 1.5 2
Real Axis
K = 7.4278
211
Now finding the location of closed loop poles at the aforesaid K value. The
closed loop transfer function with unity feedback
K(s)G(s)
Tcl (s) =
1 + K(s)G(s)
K(s + 1)
=
s3 + 4.1424s2 + (K + 1.1424)s + K − 6.2848
U + E Y
K(s) G(s)
−
212
Root Locus
2
Imaginary Axis
K = 7.4278
K = 7.4278
0
−2 K = 7.4278
−4
−6
−8
−3 −2.5 −2 −1.5 −1 −0.5 0 0.5 1
Real Axis gain = [0, 63.9274]
The root locus confirms the locus passing through the desired point but the
dominant pole is in the proximity of imaginary axis. The dominant pole can
be distant from imaginary axis by moving the lead pole in the right direction.
Generally, K > 2pl works for the design of controller. The following pl
location gives the corresponding open loop transfer function and gain values.
s + 1.8976 1
K(s)G(s) = K · ·
s + 4.69 (s + 2)(s − 1)
K(s)G(s) = |K| × 0.0777 = 1
s=−2+j2
K = 12.07
After finding the closed transfer function and substituting the above K value
213
in the characteristic equation will make the dominant root at s = −1.68. The
dominant pole can further be distant from imaginary axis by moving the lead
pole in right direction keeping the constraint of 56.3◦ between the desired
point and lead pole-zero.
214
TUTORIAL:
100 100.3 100.48 100.6 100.7 100.78 100.85 100.9 100.95 101 101.3 101.48
1 2 3 4 5 6 7 8 9 10 20 30
90
80
70
60
50
40
30
20
10
0 0
10 101 102
215
Y = m · log10 (x) + b
(x0 , y0 ) = (2, 0)
(x1 , y1 ) = (50, 90)
y1 − y0
m=
log10 (x1 ) − log10 (x2 )
90 − 0
=
log10 (50) − log10 (2)
m = 64.3804
b = −19.3804
= 45
216
MATLAB SCRIPT:
x = logspace(log10(2),log10(50));
m = 90/(log10(50)-log10(2));
b = 0-m*log10(2);
y = m*log10(x)+b;
semilogx(x,y)
grid on
217
2
0.50
Imaginary Axis
0 0 rad/s
0.50
−2
−4
−6
−5 −4.5 −4 −3.5 −3 −2.5 −2 −1.5 −1 −0.5 0 0.5 1
Real Axis
0.50
10
5 K = 26
4
Imaginary Axis
0 0 rad/s
4
−5 K = 26
−10
0.50
−7 −6 −5 −4 −3 −2 −1 0 1
Real Axis gain = [0, 176.061]
The step response of the closed loop system shown in figure is the antithesis
of the tracking perception. The steady state response has a value of 0.464
contrary to 1.
Step Response
1
0.8
0.6
y(t)
0.4
0.2
0
0 0.5 1 1.5 2 2.5 3
t
219
The reason of this anomaly is ignoring the steady state part of the response.
Generally, the root locus has the capability to alter the transient response by
the placement of poles with parameter K . The locus of the closed loop poles is
generated from the open loop poles and zeros. The analysis is encompassing
the characteristic equation of the system which remain same for any transfer
function of a specific system. Therefore the root locus analysis and design
is not focusing on the input of the system. Frequency response analysis and
design is used to cover the steady state part of the response. The root locus
and frequency response are complementary to each other and used for the
complete analysis-design.
+
U Y
K G
+
−
Y KG Y G Y KG
= = =
U 1 + KG D 1 + KG V 1 + KG
220
|K(jω)G(jω)|
TRACKING
DISTURBANCE REJECTION
BIG
NOISE REJECTION
PLANT UNCERTAINTY
STABILITY
SMALL
a
G(s) =
s+a
a
G(jω) =
a + jω
a a
|G(jω)| = √ =
a2 + ω 2 l
!
ω
G(jω = ϕ = − tan−1
a
221
ω=0 −→ |G(jω)| = 1, ϕ = 0◦
ω=∞ −→ |G(jω)| = 0, ϕ = −90◦
1
ω=a −→ |G(jω)| = √ , ϕ = −45◦
2
jω
ω
l
ϕ
σ
−a
a b
G(s) = ·
s+a s+b
a b
G(jω) = ·
a + jω b + jω
= aM1 e−jϕ1 · bM2 e−jϕ2
= a · b · M1 · M2 · e−j(ϕ1 +ϕ2 )
K(jω)G(jω) |K(jω)G(jω)|
ω
222
223
−20
−40
−60
−80
−100
−120
−10
−20
K(jω)G(jω)
−30
−40
−50
−60
−70
−80
−90
−100
10−1 100 101 102 103 104 105
ω
224
−20
−40
K(jω)G(jω)
−60
−80
−100
−120
−140
−160
−180
−200
10−1 100 101 102 103 104 105
ω
(s + a)(s + b)
G(s) =
(s + c)(s + d)
M1 ejϕ1 · M2 ejϕ2
=
M3 ejϕ3 · M4 ejϕ4
log |G(s)| = log M1 + log M2 − log M3 − log M4
G(s) = ϕ1 + ϕ2 − ϕ3 − ϕ4
225
G(s) = K, K>0
|G(jω)| = |K| = K
20 log10 |G(jω)| = 20 log10 K
G(jω) = 0◦
dB
0 ω
G(jω
0◦ ω
226
1
sn
sn
n=1 n=1
n=3 n=3
G(s) = sn
G(jω) = (jω)n , G(jω) = n(90◦ )
G(jω) = jω, n = 1, G(jω) = 90◦
G(jω) = −ω 2 , n = 2, ∵ j 2 = −1, G(jω) = 180◦
G(jω) = −jω 3 , n = 3 ∵ j 3 = −j, G(jω) = 270◦
G(jω) = ω 4 , n = 4 ∵ j 4 = 1, G(jω) = 360◦
1
G(s) =
sn
1
G(jω) = , G(jω) = n(−90◦ )
(jω)n
1
G(jω) = , n = 1, G(jω) = −90◦
jω
1
G(jω) = − 2 , n = 2, ∵ j 2 = −1, G(jω) = −180◦
ω
1
G(jω) = − 3 , n = 3 ∵ j 3 = −j, G(jω) = −270◦
jω
227
1
G(jω) = , n = 4 ∵ j 4 = 1, G(jω) = −360◦
ω4
G(s) = sn
G(jω) = (jω)n
20 log10 |G(jω)| = n · 20 log10 ω
G(jω) = n(90◦ )
Bode Diagram
60
Magnitude (dB)
40
20
0
270
225
Phase (deg)
180
135
90
100 100.1 100.2 100.3 100.4 100.5 100.6 100.7 100.8 100.9 101
Frequency (rad/s)
228
1
G(s) =
sn
1
G(jω) =
(jω)n
20 log10 |G(jω)| = −n · 20 log10 ω
G(jω) = n(−90◦ )
Bode Diagram
0
Magnitude (dB)
−20
−40
−60
−90
−135
Phase (deg)
−180
−225
−270
100 100.1 100.2 100.3 100.4 100.5 100.6 100.7 100.8 100.9 101
Frequency (rad/s)
229
s+a
G(s) =
a !
s
a
+1
a
=
a
G(s) = τ s + 1
G(jω) = 1 + jωτ
p
|G(jω)| = 1 + ω 2 τ 2
20 log10 |G(jω)| = 20 log10 1 = 0dB, ωτ << 1
√
20 log10 |G(jω)| = 20 log10 2 = 3dB, ωτ = 1
20 log10 |G(jω)| = 20 log10 ωτ, ωτ >> 1
dB
60
40
20
20 dB/decade
0 ωτ
100 101 102 103
230
40
35
|G(jω)|(dB)
30
25
20
15
10
!
ωτ
G(jω) = tan−1
1
= 0◦ , ωτ << 1
= 45◦ , ωτ = 1
= 90◦ , ωτ >> 1
ϕ = tan−1 (ωτ )
u = ωτ, x = ln ωτ
ex = eln ωτ = ωτ = u
ϕ = tan−1 (u)
d d
(ϕ) = (tan−1 (u))
d(ln ωτ ) dx
ωτ =1
d 1 d
(tan−1 (u)) = · (u)
dx 1 + u2 dx
d d
(ωτ ) = (u) = ex
d(ln ωτ ) dx
ln ωτ
=e = e0 = 1
ωτ =1
d 1
(tan−1 (u)) = ·1
dx 1 + u2
1 1
= =
1 + u2 2
u=ωτ =1
232
ωτ = 4.81 ≈ 5
90◦
−11◦
1
2
11◦
0◦ ωτ = 100 = 1
1 1
ωτ = 4.81 ≈ 5
ALTERNATIVE PROOF:
ϕ = tan−1 (ωτ )
!
ω 1
= tan−1 , τ =
ω0 ω0
d 1 1
(ϕ) = 2 ·
dω 1 + ωω2 ω0
0
ω02 1 ω0
= · =
ω02 + ω 2 ω
ω + ω02
0
ω
d 0 1
(ϕ) = =
dω 2ω02 2ω0
ω=ω0
233
!
π ω
ϕSL (ω) = + m log10
4 ω0
!!
d d ω
(ϕSL (ω)) = m log10
dω dω ω0
!
d ln( ωω0 )
= m·
dω ln 10
!!
m d ω
= · ln
ln 10 dω ω0
!
m 1 1
= · ω ·
ln 10 ω
ω
0
0
m
=
ω ln 10
d m
(ϕSL (ω)) =
dω ω0 ln 10
ω=ω0
m 1
=
ω
0 ln 10 2ω0
ln 10
m=
2 ! !
π ln 10 ω
ϕSL (ω) = + · log10
4 2 ω0
!
!
π
ln 10
ln( ωω0 )
= + ·
4 2 ln10
!
π 1 ω
= + · ln
4 2 ω0
234
!
π 1 ω
ϕSL (ω) = + · ln
4 2 ω0
ω −→ ϕSL (ω) = 0
!
π 1 ω
0= + · ln
4 2 ω0
!
1 ω π
· ln =−
2 ω0 4
!
ω π
ln =−
ω0 2
ω
ln
= e− 2
π
e ω0
ω
= e− 2
π
ω0
ω0
ω = ω0 · e− 2 ≈
π
5
π
ω −→ ϕSL (ω) =
2!
π π 1 ω
= + · ln
2 4 2 ω0
!
1 ω π
· ln =
2 ω0 4
!
ω π
ln =
ω0 2
ln ω π
e ω0 = e 2
ω π
= e2
ω0
π
ω = ω0 · e 2 ≈ 5 · ω0
235
90
80
70
60
G(jω)
50
40
30
20
10
0
10−2 10−1 100 101 102
ωτ
a
G(s) =
s+a
a
= !
s
a
+1
a
1
G(s) =
τs + 1
1
G(jω) =
1 + jωτ
1
|G(jω)| = √
1 + ω2τ 2
20 log10 |G(jω)| = 20 log10 1 = 0dB, ωτ << 1
√
20 log10 |G(jω)| = −20 log10 2 = −3dB, ωτ = 1
20 log10 |G(jω)| = −20 log10 ωτ, ωτ >> 1
236
−5
−10
|G(jω)|(dB)
−15
−20
−25
−30
−35
−40
!
ωτ
G(jω) = − tan−1
1
= 0◦ , ωτ << 1
= −45◦ , ωτ = 1
= −90◦ , ωτ >> 1
237
−10
−20
−30
G(jω)
−40
−50
−60
−70
−80
−90
10−2 10−1 100 101 102
ωτ
238
LEAD
s + 10 100
G(s) = ·
10 s + 100
60
40
|G(jω)|(dB)
20
−20
−40
100
80
60
40
G(jω)
20
−20
−40
−60
−80
−100
10−1 100 101 102 103 104
ω
239
LAG
s + 100 10
G(s) = ·
100 s + 10
40
30
20
|G(jω)|(dB)
10
−10
−20
−30
−40
−50
−60
10−1 100 101 102 103 104
ω
100
80
60
40
G(jω)
20
−20
−40
−60
−80
−100
10−1 100 101 102 103 104
ω
240
Bode Diagram
20
10
Magnitude (dB)
−10
−20
90
45
Phase (deg)
−45
−90
10−1 100 101 102 103 104
Frequency (rad/s)
241
s2 + 2ζωn s + ωn2
G(s) =
ωn2
!2 !
s s
= + 2ζ +1
ωn ωn
!2 !
ω ω
G(jω) = 1 − + j2ζ
ωn ωn
v !2 !2 !!2
u
u ω ω
|G(jω)| = t 1 − + 2ζ
ωn ωn
= 0◦ , ω << ωn
= 90◦ , ω = ωn
= 180◦ , ω >> ωn
242
60
Magnitude (dB)
40
20
−20
180
135
Phase (deg)
90
45
0
10−2 10−1 0
10 101 102
ω
Frequency ωn
(rad/s)
2ζ ωωn
ϕ(ω) = tan−1 2
1− ω
ωn
At High Frequencies
ω
ln
π ωn
π= +
2 ζ
π
ζ
ω = ωn e 2
ωn2
G(s) =
s2 + 2ζωn s + ωn2
20
0
Magnitude (dB)
−20
−40
−60
−45
Phase (deg)
−90
−135
−180
10−2 10−1 0
10 101 102
ω
Frequency ωn
(rad/s)
245
s+a
G(s) =
s(s2 + 2ζωn s + ωn2 )
Bode Diagram
20
0
Magnitude (dB)
−20
−40
−60
−80
−45
−90
Phase (deg)
−135
−180
−225
10−1 100 101 102
Frequency (rad/s)
jω + a
G(s) = G(jω) =
(jω)((jω)2 + 2ζωn (jω) + ωn2 )
|G(jω)|ω=1 rad/sec = −8.3924 dB = 0.3805
G(jω) = −78.4763◦
246
R(s) + Y(s)
K G(s)
−
Y (s) KG(s)
Tcl (s) = =
R(s) 1 + KG(s)
Unstable Closed Loop System Response Increases without any Bound.
∆(s) = 1 + KG(s) = 0
1 + KG(s) = 0
KG(s) = −1
KG(jω) = −1
|KG(jω)| = 1 = 0 dB
KG(jω) = ±180◦
Bode Diagram
20
0
Magnitude (dB)
−20
−40
−60
−80
−45
−90
Phase (deg)
−135
−180
−225
10−1 100 101 102
Frequency (rad/s)
13.3
GM = 13.3 dB = 10 20 = 4.6238
248
Bode Diagram
40
20
0
Magnitude (dB)
−20
−40
−60
−80
−45
−90
Phase (deg)
−135
−180
−225
10−1 100 101 102
Frequency (rad/s)
249
Root Locus
10
5
Imaginary Axis
−5
−10
Root Locus Confirms the Poles of the Closed Loop system are on imaginary
axis when K = 4.62.
EXAMPLE
1
KG(s) =
s−1
When K = 1.
250
Pole-Zero Map
1
0.8
0.6
0.4
Imaginary Axis
0.2
−0.2
−0.4
−0.6
−0.8
−1
0.9 0.92 0.94 0.96 0.98 1 1.02 1.04 1.06 1.08 1.1
Real Axis
251
Bode Diagram
−10
Magnitude (dB)
−20
−30
−40
−90
Phase (deg)
−135
−180
10−2 10−1 100 101 102
Frequency (rad/s)
252
Root Locus
0.1
8 · 10−2
6 · 10−2
4 · 10−2
Imaginary Axis
2 · 10−2
0
K=1
−2 · 10−2
−4 · 10−2
−6 · 10−2
−8 · 10−2
−0.1
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
Real Axis gain = [0, 1]
R(s) + 1 Y(s)
K=1
s−1
−
1
Tcl (s) =
s
253
1
KG(s) =
s−1
When K = 2.
Bode Diagram
10
0
Magnitude (dB)
−10
−20
−30
−40
−90
Phase (deg)
−135
−180
10−2 10−1 100 101 102
Frequency (rad/s)
254
Root Locus
0.1
8 · 10−2
6 · 10−2
4 · 10−2
Imaginary Axis
2 · 10−2
0
K=2 K=1
−2 · 10−2
−4 · 10−2
−6 · 10−2
−8 · 10−2
−0.1
−1 −0.8 −0.6 −0.4 −0.2 0 0.2 0.4 0.6 0.8 1
Real Axis gain = [0, 2]
R(s) + 1 Y(s)
K=2
s−1
−
2
Tcl (s) =
s+1
255
NYQUIST PLOTS
POLAR PLOT
COMBINING BODE PLOT
4
G(s) =
s(s + 5)
BODE PLOT
Bode Diagram
20
0
Magnitude (dB)
−20
−40
−60
−80
−90
Phase (deg)
−135
−180
10−1 100 101 102
Frequency (rad/s)
256
NYQUIST PLOT
Nyquist Diagram
1
Imaginary Axis
−1
−2
−4
−8
−1 −0.5 0
257
Nyquist Plot −→ Mapping from ’s’ plane (space) to G(s) plane (space)
D-Contour
jω
P σ
RH
s-plane (space)
G(jω) = G(−jω)
Nyquist Diagram
4
1
Imaginary Axis
−1
−2
−3
−4
−1 0 1 2 3 4 5
Real Axis
Conformal Mappings
Analytical Function G(s) −→ If locus of s space moves through an angle
’α’ then the locus of G(s) space must also move through an angle ’α’.
259
α
s = s1
G(s1)
α
σ Re{G(s)}
G(s)
1
MAPPING OF s
jω jω
1
G(s) =
s
σ 1 σ
G(jω) =
jω
−1 PASSING
Nyquist plot passed with −1 point on right or the left (Direction in increasing
ω)
Im{G(s)}
RIGHT
Re{G(s)}
−1
LEFT
s-space G(s)-space
jω Im{G(s)}
θ s1 θ
ω G(s ) = −1
1
σ G(s1) Re{G(s)}
ω −1
s-space G(s)-space
jω Im{G(s)}
s1 θ
G(s1) = −1
ω
σ G(s1) Re{G(s)}
−1 θ
Conformal mappings indicate that for closed-loop stability the Nyquist plot
must pass with the −1 point on the left (in direction of increasing ω ).
ORIGIN ENCIRCLEMENT
s-space G(s)-space
jω Im{G(s)}
G(s) = s + a
σ Re{G(s)}
−a 0 0 a
G(jω) = jω + a
No Encirclement of Origin
264
s-space G(s)-space
jω Im{G(s)}
1
G(s) = s+a
1
a
σ Re{G(s)}
−a 0 0
1
G(jω) = jω+a
No Encirclement of Origin
s-space G(s)-space
jω ∠G(jω)
90
G(s) = s + a
σ
−a 0 0
G(jω) = jω + a
−90
s-space G(s)-space
jω ∠G(jω)
180
G(s) = s − a
180 − α
α
σ
0 a 0
G(jω) = jω − a
−180
265
s-space G(s)-space
jω Im{G(s)}
G(s) = s − a
180 − α
α
σ Re{G(s)}
0 a 0
G(jω) = jω − a
s-space G(s)-space
jω Im{G(s)}
1
G(s) = s−a
180 − α
α − a1
σ Re{G(s)}
0 a 0
1
G(jω) = jω−a
(s − α)
G(s) =
(s + γ)(s − β)
266
G(s)-space
Im{G(s)}
Re{G(s)}
0
KG(s)
Tcl (s) =
1 + KG(s)
∆(s) = 1 + KG(s) = 0
KG(s) = −1
n
KG(s) =
Po
n n + Po Pc
1 + KG(s) = 1 = = =
Po Po Po
267
NQ = Nc − No
NQ + No = Nc
268
EXAMPLES
EXAMPLE 1
2
KG(s) =
s−1
Nyquist Diagram
0.8
0.6
0.4
Imaginary Axis
0.2
−0.2
−0.4
−0.6
−0.8
−1
−2 −1.8 −1.6 −1.4 −1.2 −1 −0.8 −0.6 −0.4 −0.2 0
Real Axis
NQ = 1, Anticlockwise Encirclement
No = 1
NQ = No , Closed Loop Stable
269
Bode Diagram
0
Magnitude [dB]
−10
−20
−30
−100
Phase [deg]
−120
−140
−160
−180
10−2 10−1 100 101 102
Frequency [rad/s]
Root Locus
0.1
8 · 10−2
6 · 10−2
4 · 10−2
Imaginary Axis
2 · 10−2
0
K=2 K=1
−2 · 10−2
−4 · 10−2
−6 · 10−2
−8 · 10−2
−0.1
−1 −0.8 −0.6 −0.4 −0.2 0 0.2 0.4 0.6 0.8 1
Real Axis gain = [0, 2]
EXAMPLE 2 ( )
1
s+ 3
KG(s) = 2 ×
(s + 1)(s + 2)
270
Nyquist Diagram
0.4
0.3
0.2
0.1
Imaginary Axis
−0.1
−0.2
−0.3
−0.4
−1 −0.8 −0.6 −0.4 −0.2 0 0.2 0.4 0.6
Real Axis
NQ = 0
No = 0
Nc = NQ + No = 0, Closed Loop Stable
Bode Diagram
−10
Magnitude [dB]
−20
−30
0
Phase [deg]
−50
Root Locus
0.1
0.08
0.06
0.04
0.02
Imaginary Axis
0
K=2 K=2
−0.02
−0.04
−0.06
−0.08
−0.1
−10 −9 −8 −7 −6 −5 −4 −3 −2 −1 0
Real Axis gain = [0, 7.11905]
EXAMPLE 3
K
KG(s) =
(s + 1)3
K=1
Nyquist Diagram
0.8
0.6
0.4
0.2
Imaginary Axis
−0.2
−0.4
−0.6
−0.8
−1 −0.8 −0.6 −0.4 −0.2 0 0.2 0.4 0.6 0.8 1
Real Axis
272
NQ = 0
No = 0
Nc = NQ + No = 0, Closed Loop Stable
K=15
Nyquist Diagram
12
10
4
Imaginary Axis
−2
−4
−6
−8
−10
−12
−4 −2 0 2 4 6 8 10 12 14
Real Axis
NQ = 2, Clockwise Encirclement
No = 0
Nc = NQ + No = 2, Closed Loop Not Stable
273
Root Locus
3
2
K = 15
1
Imaginary Axis
0
K = 15
−1
K = 15
−2
−3
−4.5 −4 −3.5 −3 −2.5 −2 −1.5 −1 −0.5 0 0.5 1 1.5
Real Axis gain = [0, 13.9655]
K
KG(s) =
(s + 1)3
K
KG(jω) =
(jω + 1)3
KG(jω) = −3 tan−1 (ω) = −180◦
ω = tan (60◦ ) = 1.732
K
|KG(jω)| = √
( 1 + ω 2 )3
K
|KG(jω)|ω=1.732 = √ =1
( 1 + (1.732)2 )3
K=8
274
K=1
Bode Diagram
−10
−20
Magnitude (dB)
−30
−40
−50
−60
−70
0
−45
−90
Phase (deg)
−135
−180
−225
−270
10−1 100 101
Frequency (rad/s)
K=8
Bode Diagram
20
10
0
Magnitude (dB)
−10
−20
−30
−40
−50
0
−45
−90
Phase (deg)
−135
−180
−225
−270
10−1 100 101
Frequency (rad/s)
275
Nyquist Diagram
2
Imaginary Axis
−2
−4
−6
−2 −1 0 1 2 3 4 5 6 7 8
Real Axis
EXAMPLE 4
s+3
KG(s) = 0.5 ×
s(s − 1)
Nyquist Diagram
2
Imaginary Axis
−1
−2
−3
−4
NQ = 1, Clockwise Encirclement
No = 1
Nc = NQ + No = 2, Closed Loop Not Stable
277
Root Locus
3.5
2.5
1.5
K = 0.5
1
Imaginary Axis
0.5
−0.5
−1
K = 0.5
−1.5
−2
−2.5
−3
−3.5
−28 −26 −24 −22 −20 −18 −16 −14 −12 −10 −8 −6 −4 −2 0 2
Real Axis gain = [0, 31.8843]
0.5
Imaginary Axis
−0.5
−1
−0.1667 × K = −1
K=6
−|a| × K = −1
1
GM = K =
|a|
1) G(jωp ) = −180◦
2) |G(jωp )|
1 1
3) a = −|G(jωp )| −→ GM = K = =
|G(jωp )| |a|
1
|G(jωp )| =
GM
GM should be at least 3
GM < 3 −→ Unsatisfactory behaviour
279
EXAMPLE
1
G(s) =
s(s + 1)(s + 2)
1
G(jω) =
(jω)(jω + 1)(jω + 2)
!
ωp
G(jωp ) = −90◦ − tan−1 (ωp ) − tan−1 = −180◦
2
ω
ω + 2p
−1 p
−90◦ − tan ωp2
= −180◦
1− 2
ω
ω + 2p
−1 p
tan ωp2
= 90◦
1− 2
ωp
ωp + 2
ωp2
= tan (90◦ ) = ∞
1− 2
ωp2
1− =0
2
√
ωp = 2
a = −|G(jωp )|
1
=−
(jωp )(jωp + 1)(jωp + 2)
1
= −q √ √ √
( 2)2 (1 + ( 2)2 )(4 + ( 2)2 )
1
=−
6
1
GM = =6
|a|
280
Nyquist Diagram
2
Imaginary Axis
−2
−4
−6
−8
−4 −3.5 −3 −2.5 −2 −1.5 −1 −0.5 0
Real Axis
PHASE MARGIN
Nyquist Diagram
0.5
Imaginary Axis
−0.5
ϕ
−1
P M = ϕ = 180◦ + p
Nyquist Diagram
0.5
Imaginary Axis
−0.5
ϕ p
p
−1
1) |G(jωG )| = 1
2) G(jωG )
EXAMPLE
1
G(s) =
s(s + 0.2)
1
G(jω) =
(jω)(jω + 0.2)
1
|G(jωG )| = =1
(jωG )(jωG + 0.2)
1
= q =1
2 2
ωG (0.22 + ωG )
q
2 2
= ωG (0.22 + ωG )=1
2
= ωG (0.22 + ωG
2
)=1
4
= ωG + 0.22 ωG
2
−1=0
2
ωG = 0.96
√
ωG = 0.96
283
√ !
0.96
G(jωG ) = −90◦ − tan−1
0.2
= −168.5◦
P M = ϕ = 180◦ − 168.5◦
= 11.5◦
10
−2
−4
−6
−8
−10
−15 −14 −13 −12 −11 −10 −9 −8 −7 −6 −5 −4 −3 −2 −1 0
284
DESIGN STEPS
1. Fix Crossover Frequency
LAG
LEAD
0dB
−180◦
285
R(s) + 1 Y(s)
K
(s + 1)2
−
K
Tol (s) =
(s + 1)2
Tol (s)
Tcl (s) =
1 + Tol (s)
K
(s+1)2
= K
1 + (s+1) 2
K
=
s2 + 2s + K + 1
∆(s) = s2 + 2s + K + 1 = 0
s2 + 2ζωn s + ωn2 = s2 + 2s + K + 1
ωn2 = K + 1 = 3 + 1
ωn = 2
K=3
286
2ζωn = 2
ζ = 0.5
0.50
1.5
0.5
Imaginary Axis
0 0 rad/s
−0.5
−1
−1.5
0.50
−1.1 −1 −0.9 −0.8 −0.7 −0.6 −0.5 −0.4 −0.3 −0.2 −0.1 0
Real Axis
ERROR ANALYSIS:
E(s) 1
TE (s) = = K
R(s) 1+ (s+1)2
(s + 1)2
=
s2 + 2s + K + 1
1
R(s) = , K = 3
s
ess (∞) = lim sE(s) = lim sTE (s)R(s)
s→0 s→0
(s + 1)2 1
s×
= lim ×
s→0 s2 + 2s + 4 s
= 0.25 ≰ 0.1
287
Step Response
0.9
0.8
0.7
0.6
0.5
Output
0.4
0.3
0.2
0.1
0
0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5
Time [s]
E(s) 1
TE (s) = = K
R(s) 1+ (s+1)2
(s + 1)2
=
s2 + 2s + K + 1
1
R(s) = , K = 9
s
ess (∞) = lim sE(s) = lim sTE (s)R(s)
s→0 s→0
(s + 1)2 1
s×
= lim ×
s→0 s2 + 2s + 10 s
= 0.1 ≤ 0.1
288
Increasing the value of K to 9 in order to reduce the steady state error will
decrease the damping.
3
K(s)G(s) =
(s + 1)2
Bode Diagram
0
Magnitude (dB)
−20
−40
−60
−45
Phase (deg)
−90
−135
−180
10−2 10−1 100 101 102
Frequency (rad/s)
Since the DC gain need to be amplified so Lag controller will be a best option
with gain of 3.
289
s+b
K(s) = 3 ×
s+a
b
=3
a
s + 0.3
K(s) = 3 ×
s + 0.1
Bode Diagram
20
18
16
Magnitude (dB)
14
12
10
8
0
−5
−10
Phase (deg)
−15
−20
−25
−30
10−3 10−2 10−1 100 101
Frequency (rad/s)
290
3(s + 0.3)
K(s)G(s) =
(s + 0.1)(s + 1)2
Bode Diagram
20
0
Magnitude (dB)
−20
−40
−60
−80
0
−45
Phase (deg)
−90
−135
−180
10−3 10−2 10−1 100 101 102
Frequency (rad/s)
291
Root Locus
2
0.50
1 1.00
Imaginary Axis
0 0 rad/s
−1 1.00
0.50
−2
−1 −0.9 −0.8 −0.7 −0.6 −0.5 −0.4 −0.3 −0.2 −0.1 0 0.1
Real Axis
Step Response
0.8
0.6
Output
0.4
0.2
0
0 2 4 6 8 10 12 14 16 18 20
Time [s]
292
GAIN MARGIN
EXAMPLE
0.4(s + 5)
Tol (s) = KK(s)G(s) = K
s(s + 2)(s + 1)
Bode Plot
60
K=1
40 K=0.5
K=4
20
Magnitude (dB)
-20
-40
-60
-80
-100
-90
-135
Phase (deg)
-180
-225
10 -2 10 -1 10 0 10 1 10 2
Frequency (rad/s)
294
Given: GM = x dB
Desired: GMd = y dB
−20 log10 (K) = y − x
K = 10−(
y−x
20
)
PHASE MARGIN
|KK(jωG )G(jωG )| = 1
ϕ = P M = 180◦ + K(jωG )G(jωG )
EXAMPLE
R(s) + 1 Y(s)
K
s(s + 1)(s + 5)
−
1
Tol (s) = KK(s)G(s) = K
s(s + 1)(s + 5)
Bode Plot
50
K=1
K=2
K=10
0
Magnitude (dB)
-50
-100
-150
-90
-135
Phase (deg)
-180
-225
-270
10 -2 10 -1 10 0 10 1 10 2
Frequency (rad/s)
296
EXAMPLE
R(s) + 1 Y(s)
K
(s + 4)4
−
1
Tol (s) = KG(s) = K
(s + 1)4
Bode Diagram
Gm = 12 dB (at 1 rad/s) , Pm = -180 deg (at 0 rad/s)
0
K=1
-20
Magnitude (dB)
-40
-60
-80
-100
0
-45
-90
Phase (deg)
-135
-180
-225
-270
-315
-360
10 -1 10 0 10 1
Frequency (rad/s)
298
P Md = 60◦
P Md − 180◦ = G(jωG )
60◦ − 180◦ = G(jωG )
−120◦ = G(jωG )
G(s) = −4 s + 1
G(jωG ) = −4 jωG + 1
−4 jωG + 1 = −120◦
jωG + 1 = 30◦
tan−1 ωG = 30◦
ωG = tan (30◦ )
1
ωG = √
3
1 1
K= =
|G(jωG |
G(j √
1
3)
s 4
1
= +1
3
16
= = 1.78
9
1
KG(s) = 1.78
(s + 1)4
299
Bode Diagram
Gm = 7.03 dB (at 1 rad/s) , Pm = 59.9 deg (at 0.578 rad/s)
20
K=1.78
0
Magnitude (dB)
-20
-40
-60
-80
0
-45
-90
Phase (deg)
-135
-180
-225
-270
-315
-360
10 -1 10 0 10 1
Frequency (rad/s)
EXAMPLE
1−s
Tol (s) = K
(s + 1)2
Bode Diagram
Gm = 6.02 dB (at 1.73 rad/s) , Pm = -180 deg (at 0 rad/s)
0
-5
Magnitude (dB)
-10
-15
-20
360
K=1
315
270
Phase (deg)
225
180
135
90
10 -1 10 0 10 1
Frequency (rad/s)
(b) Estimate K such that the closed loop system phase margin is 45◦ . What
is the corresponding gain margin.
301
P Md = 45◦
G(jωg ) = P Md − 180◦
= −135◦
ωG = 1 rad/s, From Bode Plot
|G(jωG )| = |G(j1)| = −3.01 dB, From Bode Plot
√
K = 0 dB − (−3.01 dB) = 3.01 dB = 2
√ 1−s
Tol (s) = 2
(s + 1)2
Bode Diagram
Gm = 3.01 dB (at 1.73 rad/s) , Pm = 45 deg (at 1 rad/s)
5
0
Magnitude (dB)
-5
-10
-15
-20
360
315
270
Phase (deg)
225
180
135
90
10 -1 10 0 10 1
Frequency (rad/s)
(c) The closed loop stability through Nyquist stability criterion must be con-
firmed.
302
303
s+b b
Tlg (s) = K , where α=
s+a a
s + αa
=K , 1 ≤ α ≤ 10
s+a
Bode Diagram
13
12
11
Magnitude (dB)
10
6
0
−5
Phase (deg)
−10
−15
−20 √
a a α αa
Frequency (rad/s)
! !
ω ω
Tlg (jω) = tan−1 − tan−1
αa a
! ! ! !
a a αa αa
tan−1 − tan−1 = tan−1 − tan−1
αa a αa a
!
1
= tan−1 − 45◦
α
The largest (negative) phase occurs at the geometric mean of the corner
frequencies i.e
√
ωm = a α
The phase at ωm
! !
ωm ωm
Tlg (jωm ) = tan−1 − tan−1
αa a
√ ! √ !
a α a α
= tan−1 − tan−1
αa a
305
!
1 √
= tan−1 √ − tan−1 ( α)
α
√
√1
α −
−1 α
= tan √
1 + √1α α
√
√ −
1
α
−1 α
= tan
2
α 2 3 4 6 8 10
Phase −19◦ −30◦ −37◦ −46◦ −51◦ −55◦
EXAMPLE
s + 0.5
Tlg (s) = 0.2 , α = 5, K = 0.2, a = 0.1
s + 0.1
Bode Diagram
0
-2
-4 System: sys
Magnitude (dB)
-8
-10
-12
-14
0
System: sys
Phase (deg)
System: sys
Frequency (rad/s): 0.1
Phase (deg): -33.7 System: sys
Frequency (rad/s): 0.222
Phase (deg): -41.8
-60
10 -3 10 -2 10 -1 10 0 10 1
Frequency (rad/s)
EXAMPLE
s+5 3
Tol (s) = KK(s)G(s) = 2 × ×
s+1 (s + 1)(s + 4)
307
System: sys l ag
Frequency (rad/s): 0.0265 System: sys l ag
Magnitude (dB): 20 Bode Diagram
Frequency (rad/s): 2.2
20
Magnitude (dB): 13.1
System: untitled1
10 Frequency (rad/s): 0.0265
Magnitude (dB): 17.5 System: sys p lant
0
System: untitled1 Frequency (rad/s): 2.2
System: sys p lant Frequency (rad/s): 2.2 Magnitude (dB): -11.3
-10
Frequency (rad/s): 0.0265 Magnitude (dB): 1.77
Magnitude (dB): -2.5
Magnitude (dB)
-20
-50
System: sys p lant
-60 Frequency (rad/s): 31.6
Magnitude (dB): -50.5
-70
-80
0
-45
System: sys l ag
Frequency (rad/s): 2.21
Phase (deg)
-90
System: sys p lant
Frequency (rad/s): 2.2
Phase (deg): -94.3
-135
System: untitled1
Frequency (rad/s): 2.2
Phase (deg): -136
-180
10 -2 10 -1 10 0 10 1 10 2
Frequency (rad/s)
308
s+a b
Tld (s) = K , where α=
s+b a
s+a
=K , 1 ≤ α ≤ 10
s + αa
Bode Diagram
16
14
12
Magnitude (dB)
10
2
40
30
Phase (deg)
20
10
0 √
a a α αa
Frequency (rad/s)
! !
ω ω
Tld (jω) = tan−1 − tan−1
a αa
Phase is always positive ∵ α > 1
Phase at corner frequencies ω = a and ω = αa must be the same i.e.
! ! ! !
a a αa αa
tan−1 − tan−1 = tan−1 − tan−1
a αa a αa
!
1
= 45◦ − tan−1
α
The largest phase occurs at the geometric mean of the corner frequencies i.e
√
ωm = a α
The phase at ωm
! !
ωm ωm
Tlg (jωm ) = tan−1 − tan−1
a αa
√ ! √ !
a α a α
= tan−1 − tan−1
a αa
310
!
√ 1
= tan−1 ( α) − tan−1 √
α
√
α− √1
−1 α
= tan √
1+ √1 α
α
√
α− √1
α
−1
= tan
2
α 2 3 4 6 8 10
Phase 19◦ 30◦ 37◦ 46◦ 51◦ 55◦
EXAMPLE
s + 0.5
Tld (s) = 2 , α = 6, K = 2, a = 0.5
s+3
0.5 1
ωlow GAIN −→ 2 × = −9.5dB
=
3 3
ωhigh GAIN −→ 2 = 6.02dB
√
ωm −→ 0.5 × 3 = 1.22
√
6 − √16
−1 ◦
Tld (jωm ) −→ tan = 45.6
2
311
Bode Diagram
6
System: sys
Frequency (rad/s): 50.7
4 Magnitude (dB): 6.01
System: sys
0 Frequency (rad/s): 1.22
Magnitude (dB)
-4
-6
System: sys
-8 Frequency (rad/s): 0.0359
Magnitude (dB): -9.52
-10
60
System: sys
Frequency (rad/s): 1.2
Phase (deg): 45.5
System: sys
Frequency (rad/s): 0.501
Phase (deg): 35.6
Phase (deg)
System: sys
30 Frequency (rad/s): 2.99
Phase (deg): 35.6
0
10 -2 10 -1 10 0 10 1 10 2
Frequency (rad/s)
EXAMPLE
s + 0.4 0.2
Tol (s) = KK(s)G(s) = 2 × ×
s+3 (s + 0.1)(s + 0.2)(s + 0.6)
312
Bode Diagram
50 System: G
Frequency (rad/s): 0.00318
Magnitude (dB): 24.4 System: L
System: L Frequency (rad/s): 40.1
Frequency (rad/s): 1.1 Magnitude (dB): 6
System: untitled1 Magnitude (dB): -2.71
0
Frequency (rad/s): 0.00318
Magnitude (dB): 13
System: L
System: G
Magnitude (dB)
System: untitled1
Frequency (rad/s): 40
Magnitude (dB): -104
-100
System: G
Frequency (rad/s): 40
Magnitude (dB): -110
-150
90
45
System: L
Frequency (rad/s): 1.1
0 Phase (deg): 49.9
-45
Phase (deg)
-90
System: untitled1
-135 Frequency (rad/s): 1.1
Phase (deg): -176
-180
-225
System: G
-270 Frequency (rad/s): 1.1
Phase (deg): -226
10 -3 10 -2 10 -1 10 0 10 1 10 2
Frequency (rad/s)