Professional Documents
Culture Documents
2012
www.PDHcenter.com
www.PDHonline.org
Part II:
Laplace Transform and Time-Domain Analysis
Page 1 of 25
www.PDHcenter.com
IV.
www.PDHonline.org
Laplace Transform:
The Laplace transform is one of the mathematical tools used for the solution of ordinary
linear differential equations. The Laplace transform method has the following two attractive
features:
1. The homogeneous equation and the particular integral are solved in one operation.
2. The Laplace transform converts the differential equation into an algebraic equation in
s. It is possible to manipulate the algebraic equation by simple algebraic rules to
obtain the solution in the s-domain. The final solution is obtained by taking the
inverse Laplace transform.
Definition of the Laplace Transform:
Given the function f (t ) that satisfies the condition:
f (t ) e t dt < ,
(4-1)
F(s ) = f (t )e st dt ,
(4-2)
F(s ) = L [f (t )].
(4-3)
or
The variable s is referred to as the Laplace operator, which is a complex variable, s = + j ,
where is the real part and is the imaginary part, as shown in Figure 4-1.
Im[s]
Re[s]
Page 2 of 25
www.PDHcenter.com
www.PDHonline.org
Example 4-1: Let f (t ) be a unit step function that is defined to have a constant value of unity for
t 0 and a zero for t < 0, namely, f (t ) = u s (t ) . What is its Laplace transform F(s)?
Solution:
1
F(s ) = f (t )e st dt = u s (t )e st dt = e st
0
0
s
1
1
= [0 1] =
s
s
F(s ) = f (t )e st dt = e- at e st dt = e- (s + a )t dt =
Page 3 of 25
1
e (s + a ) t
(s + a )
1
s+a
www.PDHcenter.com
www.PDHonline.org
2.
(4-4)
3.
Differentiation df (t )
= sF(s ) lim f (t ) = sF(s ) f (0 )
L
t
dt
(4-6)
d n f (t )
L n = s n F(s ) s n 1f (0 ) s n 2f (1) (0) f (n 1) (0)
dt
(4-7)
f (k ) (0 ) =
where
4.
(4-5)
d k f (t )
dt k
t =0
Integrationt
F(s )
L f ( ) d =
0
L
0
t1
5.
t2
t n 1
0
f ( )d dt
(4-8)
tn
F(s )
dt 2 dt n 2 dt n 1 = n
s
6.
(4-10)
7.
(4-11)
8.
(4-12)
s 0
Complex shifting-
L e m at f (t ) = F(s a )
9.
(4-9)
(4-13)
Page 4 of 25
(4-14)
www.PDHcenter.com
www.PDHonline.org
1 c + j
F(s )est ds ,
c
j
2j
(4-15)
where c is a real constant that is greater than the real parts of all the singularities of F(s).
Equation (4-15) represents a line integral that is to be evaluated in the s-plane. However, for
most engineering purposes the inverse Laplace transform operation can be done simply by
referring to the Laplace transform table, such as the one given in Table 4-1. Before using Table
4-1, one may need to do partial-fraction expansion first, then, use the equations in table 4-1.
Partial-fraction expansion when all the poles(*) of the transfer function are simple and real:
Example 4-3: Find the partial-fraction expansion of the following transfer function, and its
inverse Laplace.
G (s ) =
s+3
(s + 1)(s + 2)
Solution:
G (s ) =
s+3
A
B
=
+
(s + 1)(s + 2) s + 1 s + 2
where,
A = (s + 1)G (s ) s = 1 =
s+3
1+ 3
=
=2
s + 2 s = 1 1 + 2
and
B = (s + 2)G (s ) s = 2 =
s+3
2+3
=
= 1
s + 1 s = 2 2 + 1
Therefore,
G (s ) =
1
s+3
2
=
+
(s + 1)(s + 2 ) s + 1 s + 2
The inverse Laplace transform of the given transfer function can be obtained by using
some equation in Table 4-1, namely,
1
2
1 2
1 1
t
2t
+
g(t ) = L 1
+L
= 2e 1e
=L
s +1
s+ 2
s +1 s + 2
t0
Note (*): A pole is a value of s that makes a function, such a G(s), infinite, by making the denominator of the
function to zero.
Page 5 of 25
www.PDHcenter.com
www.PDHonline.org
Partial-fraction expansion when some poles of the transfer function are of multiple order:
Example 4-4: Find the partial-fraction expansion of the following transfer function, and its
inverse Laplace.
G (s ) =
1
s (s + 1)
2
Solution:
G (s ) =
1
A A
B
= 1 + 22 +
s (s + 1) s
s
s +1
2
where,
A 2 = s 2G (s )
A1 =
s=0
d 2
s G (s )
ds
1
=
=1
s + 1 s = 0
s=0
1
d 1
=
= 1
ds s + 1 s = 0 (s + 1)2 s = 0
and
B = (s + 1)G (s ) s = 1 =
1
s2
=1
s = 1
Therefore,
G (s ) =
1
1 1
1
=
+ 2+
s (s + 1) s s
s +1
2
The inverse Laplace of the given transfer function can be obtained by using equations in
Table 4-1, namely,
1
1 1
1 1
1 1
1 1
t
+ 2+
g(t ) = L 1
= 1 + t + e
=L +L 2 +L
s +1
s +1
s s
s
s
t0
Example 4-5: Find the partial-fraction expansion of the following transfer function, and its
inverse Laplace.
G (s ) =
s(s + 1) (s + 2)
3
Solution:
G (s ) =
s(s + 1) (s + 2 )
3
A1
A2
A3
B
C
+
+
+
+
2
3
s + 1 (s + 1) (s + 1) s + 2 s
where
Page 6 of 25
www.PDHcenter.com
A 3 = (s + 1) G (s )
3
www.PDHonline.org
1
=
= 1
s = 1
s(s + 2 ) s = 1
A2 =
d
(s + 1)3 G(s )
ds
A1 =
1 d2
(s + 1)3 G(s )
2
2! ds
s = 1
(2s + 2)
d 1
= 2
=0
ds s(s + 2 ) s = 1 s (s + 2)2 s = 1
s = 1
1 d2 1
1 d 2(s + 1)
=
2
2! ds s(s + 2) s = 1 2 ds s 2 (s + 2)2 s = 1
1
2(s + 1)
2(s + 1)
= 2
+ 2
+ 3
= 1
2
3
2
(
)
(
)
(
)
+
+
+
s
s
2
s
s
2
s
s
2
s = 1
B = (s + 2)G (s ) s = 2 =
1
1
=
= 0.5
3
s(s + 1) s = 2 2( 2 + 1)
and
C = sG (s ) s = 0 =
(s + 1) (s + 2) s = 0 (0 + 1) (0 + 2)
3
= 0.5
Therefore,
G (s ) =
s(s + 1) (s + 2 )
3
0
0.5 0.5
1
1
+
+
+
+
2
3
s + 1 (s + 1) (s + 1) s + 2
s
The inverse Laplace of the given transfer function can be obtained by using Table 4-1,
namely,
1
0.5 0.5
0.5
1
1
1
1
1 0.5
= L 1
+ L 1
g(t ) = L 1
+
+
+
+ L
+L
3
3
s
s + 1
s+ 2
s
s + 1 (s + 1) s + 2
(s + 1)
1
= 0.5 e t + 0.5e 2t
t0
t 31e t
(3 1)!
= 0.5 e t + 0.5e 2t 0.5t 2e t
t0
2n
s (s 2 + 2n s + 2n )
Solution:
Page 7 of 25
www.PDHcenter.com
G (s ) =
www.PDHonline.org
2n
2n
A1
A2
B
=
=
+
+
2
2
s (s + 2n s + n ) s(s + j)(s + + j) s + j s + + j s
and
= n ; = n 1 2
where
A1 = (s + j )G (s )
s = + j
A 2 = (s + + j )G (s )
with
s = j
= tan 1
2n
=
s(s + + j )
=
2n
s(s + j)
s = + j
2n
j +
=
= n e 2
( + j)(2 j) 2
=
s = j
2n
j +
= n e 2
( j)( 2 j) 2
and
B = sG (s )
s=0
2n
(s 2 + 2n s + 2n )
=1
s=0
Therefore,
- j + 2
j +
n e
e 2 1
=
+
G (s ) =
+
s (s 2 + 2n s + 2n ) 2 s + j s + + j s
2
n
The inverse Laplace of the given transfer function can be obtained by using Table 4-1,
namely,
j + t
j +
n j + 2 ( + j )t
n t j + 2 t
2 ( j )t
+e
+ e 2
g (t ) = 1 +
e
e
e e
e
=1+
2
2
j t 2
j t
2
+e
n t e
n t e j(t ) e j(t )
=1+
=
+
e
1
e
= 1 + n e t sin (t )
2j
=1+
1
1
e n t sin n 1 2 t
t0
where
1 2
= tan 1
Page 8 of 25
www.PDHcenter.com
www.PDHonline.org
(t)
(Impulse at t = 0)
1
s
us(t)
(Unit Step at t = 0)
1
s2
t
(or t [u(t)], Ramp at t = 0)
n!
s n +1
tn
(n = positive integer)
1
s+a
e at
(s + a )(s + b )
e at e bt
ba
n
s + 2n
sin n t
s
s + 2n
cos n t
1
(s + a )2
te at
n!
(s + a )n +1
t n e at
n
(s + a )2 + 2n
e at sin n t
s+a
(s + a )2 + 2n
e-at cos n t
Page 9 of 25
www.PDHcenter.com
1
(1 + sT )n
www.PDHonline.org
1
n 1
T
t
e
T n (n 1)!
2n
s 2 + 2n s + 2n
1
1+
2
n
s (s + 2n s + 2n )
1
1
e n t sinn 1 2 t
e n t sin n 1 2 t -
where = tan
2n
1
2
n
s
s + 2n s + 2n
1 2
e n t sin n 1 2 t +
where = tan 1
1
s (1 + sT )
1- e
1
2
s (1 + sT )
t 2T + (t + 2T )e
t
T
1 + as
s (1 + sT )
t + (a T )1 e T
1
t sin n t
2n
(s
(s
t
T
t + T T
1
e
T
1
2
2
s (1 + sT )
2 2
n
1 2
) (s
2 2
n1
(s
s2
2
2 2
n
2 2
n2
1
(cos n1t - cos n2 t )
2n1
2
n2
1
(sin n t + n t cos n t )
2 n
Page 10 of 25
www.PDHcenter.com
V.
www.PDHonline.org
Time-Domain Analysis:
The time response of a control system is usually divided into two parts: the transient response
and the steady-state response. Let c(t ) denote a time response and be expressed as follows:
c(t ) = c t (t ) + css (t ) ,
(5-1)
R t0
0 t<0
(5-2)
(5-3)
where us(t) is the unit step function. The step function as a function of time is shown in Figure
5-1.
r(t)
Rt t 0
0 t<0
(5-4)
Page 11 of 25
www.PDHcenter.com
www.PDHonline.org
As one can see, the slope of the signal is a constant, R. The ramp function is shown in Figure 52.
r(t)
Slope = R
R 2
t t0
2
0 t<0
(5-5)
Page 12 of 25
www.PDHcenter.com
E(s)
R(s) +
G(s)
www.PDHonline.org
C(s)
H(s)
Fig. 5-4. A closed-loop system with a negative feedback.
C(s ) =
G (s )
R (s ) .
1 + G (s )H(s )
(5-6)
(5-7)
G (s )H(s )
1
E(s ) = 1
R (s ) =
R (s ) .
1 + G (s )H(s )
1 + G (s )H(s )
(5-8)
then,
The steady-state error in time domain can be obtained by applying the final-value theorem,
namely,
e ss (t ) = lim e(t ) = lim sE(s ) = lim s
t
s 0
s0
1
R (s ) .
1 + G (s )H(s )
(5-9)
R
R
1
1
.
R (s ) = lim s
= lim
s
0
s
0
1 + G (s )H(s ) s
1 + G (s )H(s )
1 + G (s )H(s )
(5-10)
If one defines
K p = lim G (s )H(s ) ,
(5-11)
s0
where K p is the step error constant (or position error constant). Then, the equation (5-10) can
be simplified as
R
R
R
.
=
=
s 0 1 + G (s )H (s )
1 + lim G (s )H(s ) 1 + K p
e ss (t ) = lim
(5-12)
s 0
One can easily realize that to have the steady-state error to be zero when the input is a step
function, K p must be infinite. The steady-state error due to a unit step function input is also
called the steady-state position error.
Page 13 of 25
www.PDHcenter.com
(2)
www.PDHonline.org
R
.
s2
(5-13)
1
1
R
R
.
2 = lim
=
s
0
1 + G (s )H(s ) s
s + sG (s )H(s ) lim s G (s )H(s )
(5-14)
s 0
If one defines
K v = lim s G (s )H(s ) ,
(5-15)
s 0
where K v is the ramp error constant (or velocity error constant). Therefore, the equation (5-14)
can be simplified as
e ss (t ) =
R
.
Kv
(5-16)
Therefore, to have zero steady-state error, Kv needs to be infinite. The steady-state error due to a
unit ramp function input is called the steady-state velocity error.
(3)
R
.
s3
(5-17)
1
1
R
R
.
3 = lim 2 2
=
2
s 0 s + s G (s )H (s )
1 + G (s )H(s ) s
lim s G (s )H(s )
(5-18)
s0
If one defines the parabolic error constant (or acceleration error constant),
K a = lim s 2 G (s )H(s ) .
(5-19)
s0
(5-20)
K (s + 6 )
s(s + 2)(s + 4 )
Page 14 of 25
C(s)
www.PDHcenter.com
www.PDHonline.org
(A) For K = 10, what is the steady-state error to a unit step input?
(B) For K = 10, what is the steady-state error to a unit ramp input?
(C) For K = 10, what is the steady-state error to a unit parabolic input?
(D) Find the value of K such that the steady-state error to a unit ramp input is 5%.
Solution:
(A) For a unit step input and K = 10:
10(s + 6)
K p = lim G (s )H(s ) = lim
1 = .
s 0
s 0 s(s + 2 )(s + 4 )
1
1
e ss (t ) =
=
= 0.
1+ Kp 1+
(B) For a unit ramp input and K = 10:
10(s + 6 )
10(s + 6 )
60
K v = lim s G (s )H(s ) = lim s
1 = lim
=
= 7.5
s 0
s 0
s
0
(s + 2)(s + 4) 8
s(s + 2)(s + 4 )
1
1
e ss (t ) =
=
= 0.1333 = 13.33% .
K v 7.5
(C) For a Parabolic input and K = 10:
10(s + 6 )
10(s + 6)
K a = lim s 2 G (s )H(s ) = lim s 2
1 = lim s
= 0.
s 0
s 0
s
0
(s + 2)(s + 4)
s(s + 2)(s + 4)
1
1
e ss (t ) =
= = .
Ka 0
(D) From part (B), for a unit ramp input,
K (s + 6)
K (s + 6)
6K
1 = lim
=
K v = lim s G (s )H(s ) = lim s
s 0
s 0
s
0
(s + 2)(s + 4) 8
s(s + 2)(s + 4 )
1
1
e ss (t ) =
=
= 5% = 0.05 .
K v 6K
8
8
80
K=
=
.
6 (0.05) 3
K (1 + a 1s )(1 + a 2 s ) (1 + a m s )
,
s T (1 + b1s )(1 + b 2 s ) (1 + b n s )
(5-21)
where that K, all as and bs are constants and the T = 0, 1, 2,.. The type of feedback control
system is referring to the order of the pole of G(s)H(s) at s = 0. Therefore, the system is of type
T. For instance, a feedback control system with
Page 15 of 25
www.PDHcenter.com
G (s )H(s ) =
www.PDHonline.org
5(1 + 2s )
s(1 + 3s )(1 + 4s )
is of type 1.
The significance of system types is as follows (when restricted to unity feedback systems):
Type 0:
Type 1:
Type 2:
The following summary of the Steady-State Errors due to Step, Ramp and Parabolic inputs can
be constructed, as shown in Table 5-1, where T denotes the type of system.
Table 5-1: Steady-State Error analysis table.
Step
Ramp
Parabolic
(Acceleration)
e ss =
e ss =
R
Kv
e ss =
Kp
Kv
Ka
T=0
T=1
e ss = 0
T=2
e ss = 0
e ss = 0
T3
e ss = 0
e ss = 0
Type
e ss =
R
1+ Kp
e ss =
e ss =
e ss = 0
Example 5-2. For the given system with its block diagram listed below:
R(s)
+
_
5
T
s (1 + s )
Page 16 of 25
R
Ka
C(s)
www.PDHcenter.com
www.PDHonline.org
5
T
s (1 + s )
C(s)
s
H(s) = 1
One may first apply case 3 in Table 1-1 to obtain the transfer function for the inner
closed-loop as
5
T
5
s (1 + s )
= T
.
G (s ) =
5
(
)
s
1
+
s
+
5s
s
1+ T
s (1 + s )
Then, use case 3 once again with a unity feedback (H(s)=1), therefore, the overall
transfer function can be obtained as
5
5
C(s )
s (1 + s ) + 5s
= T
=
.
5
R (s )
(
)
s
1
+
s
+
5
s
+
5
1
1+ T
s (1 + s ) + 5s
T
Or, one can use Mason Rule and get the overall transfer function as
C(s )
=
R (s )
(B)
5
s (1 + s )
T
1+
5
5
s + T
1
s (1 + s )
s (1 + s )
5
.
s (1 + s ) + 5s + 5
T
5
5
=
.
(1 + s ) + 5s 1 + 6s
To find the steady-state position error, the input is assumed as a unit step input,
R (s ) =
1
, and the position error constant can be calculated as
s
Page 17 of 25
www.PDHcenter.com
www.PDHonline.org
5
1 = 5 .
K p = lim G (s )H(s ) = lim
s 0
s 0 1 + 6s
1
1
1
=
= .
1+ Kp 1+ 5 6
To find the steady-state velocity error, the input is assumed as a unit ramp input,
R (s ) =
1
, and the velocity error constant can be calculated as
s2
s 0
5
= 0.
1 + 6s
1
1
= =.
Kv 0
5
5
=
.
s(1 + s ) + 5s s(s + 6)
To find the steady-state position error, the input is assumed as a unit step input,
R (s ) =
1
, and the position error constant can be calculated as
s
5
K p = lim G (s )H(s ) = lim
1 = .
s 0
s 0 s(s + 6 )
1
1
=
= 0.
1+ Kp 1+
To find the steady-state velocity error, the input is assumed as a unit ramp input,
R (s ) =
1
, and the velocity error constant can be calculated as
s2
s 0
5
5
5
= lim
= .
s
0
s(s + 6)
s+6 6
1
1
6
=
= .
Kv 5
5
6
Page 18 of 25
www.PDHcenter.com
www.PDHonline.org
First-Order Systems:
The transfer function of a first-order system can be expressed as
G (s ) =
C(s )
1
,
=
R (s ) s + 1
(5-22)
1
R (s ) .
s + 1
(5-23)
The unit step response of the system in time domain can be obtained by inverse of the Laplace
transform, or using the formula listed in Table 4-1, namely,
1
1 1
=1 e
= L -1 s +
t 0,
(5-24)
c(t ) = L -1 [C(s )] = L -1
1
s+
s + 1 s
where is the time constant of the given control system. By plotting the output signal in time
domain, Figure 5-5 can be obtained.
c(t)
1
0.982
0.95
0.865
0.632
(5-25)
Page 19 of 25
www.PDHcenter.com
www.PDHonline.org
where
> 1 : overdamped,
= 1 : critically damped,
< 1 : underdamped,
= damping ratio
and
1
1
e n t sin 1 2 n t + ,
(5-26)
where
1 2
= tan 1
(5-27)
The time domain plot for such a second-order system with a unit step input is shown in Figure 56.
1.5
O ver Shoot
+2%
90%
0.5
t r
10%
0
tp
10
ts
Fig. 5-6. The typical unit step input response of a second-order system.
There are some important performance criteria that are typically used to characterize the
transient response to a unit step input include percent (maximum) overshoot, peak time, rise time
and settling time.
Page 20 of 25
www.PDHcenter.com
Percent overshoot:
%OS = e
1 2
100% .
(5-28)
seconds.
n 1 2
Rise time:
tr =
Peak time:
tp =
www.PDHonline.org
(5-29)
The time required for the output to rise from 10% of the final value to
90% of the final values.
0.8 + 2.5
seconds.
n
(5-30)
Settling time: The time it takes for the system to settle within certain percentage of the
final value.
ts =
4
n
(5-31)
Example 5-3: The transfer function of the second order system given below is obtained in
R(s) +
-
4
s +1
1
s
C(s)
C(s )
4
= 2
. What are the percent overshoot, peak time, rise time and the
R (s ) s + s + 4
settling time for this system?
Solution:
By comparing the actual transfer function of the system with the classical second-order
equation, namely,
Example 1-1 as
C(s )
4
2n
=
=
R (s ) s 2 + s + 4 s 2 + 2n s + 2n
One should be able to obtained the undamped natural frequency and the ramping ratio as
n = 2 ; =
1
= 0.25
2n
Page 21 of 25
www.PDHcenter.com
Step
Sum
www.PDHonline.org
s+1
s
Integrator
Transfer Fcn
Scope
y
To Workspace
n2
C(s )
4
=
=
R (s ) s 2 + s + 4 s 2 + 2 n s + n2
1.5
O ver Shoot
+2%
90%
0.5
t r
10%
0
tp
10
ts
1 2
100 = e
0.25
% Overshoot:
%OS = e
Peak time:
tp =
=
= 1.6223 seconds
2
n 1
2 1 0.252
Rise time:
tr =
Settling time:
ts =
4
4
=
= 8 seconds (for 2%)
n 0.25 2
1 0.252
100 = 44.43%
If one uses Matlab/Simulink to model this simple system, one should be able to get the
plot shown above and confirm the answers obtained with equations (5-28)~(5-31).
Page 22 of 25
www.PDHcenter.com
www.PDHonline.org
Example 5-4: Find K and a for the closed-loop system below such that the transient response to a
step input satisfies: %OS 5% & t s 4 seconds
R(s) +
1
s(s + a )
C(s)
Solution:
The transfer function of the system can be obtained as
K
n2
C(s )
K
s(s + a )
= 2
= 2
=
.
2
K
R (s )
s
+
a
s
+
K
s
+
2
s
+
n
n
1+
1
s(s + a )
Therefore,
K = 2n ; a = 2n .
The percent overshoot is required to be less than or equal to 5%, hence,
%OS = e
1 2
100% =5%
= ln (0.05) = 2.995732
(0.95357 )
0.9093 0.9093 2 = 2
2 =
0.9093
1.9093
0.9093
= 0.47625 = 0.69
1.9093
1 2
1 2 =
To find the undamped natural frequency, one needs to use the other given condition,
namely,
ts =
4
=4
n
n = 1
n =
1
= 1.4491
0.69
Page 23 of 25
www.PDHcenter.com
www.PDHonline.org
1.2
1
0.8
0.6
0.4
0.2
0
0
10
Page 24 of 25
www.PDHcenter.com
www.PDHonline.org
References:
[1]
[2]
[3]
[4]
[5]
[6]
Page 25 of 25