MAK305 09 TransientResponseAnalysis
MAK305 09 TransientResponseAnalysis
●
TIME RESPONSE of a system is defined as the behavior of a system under the effect of an input
which is assumed to be applied at t=0 with zero initial conditions.
●
TIME RESPONSE consists of two parts:
1. Steady-state response : The manner in which the system behaves as t → ∞.
2. Transient response : The part of response between initial state and final state.
●
In the analysis and design of control systems, a common basis of comparison exist for the transient
performances of various systems. This is usually done by using particular test input signals and
comparing the responses.
INPUT LAPLACE TR. RESPONSE
Unit Impulse 1 Impulse Response
used in most
transient Unit Step 1/s Step Response commonly
response Unit Ramp 1/s2 Ramp Response used
analysis Unit Parabolic 1/s3 Acceleration Response
Sinusoidal Frequency Response
Random Statistical Response
●
For Linear Time-Invariant (LTI) systems, if the response to any one of these test inputs is given, the
others can be obtained.
If Y ( s)=GYX ( s) X ( s) then Y imp ( s)=GYX ( s)=Y stp ( s) s=Y rmp (s) s2 =Y acc ( s) s3
δr
0 td tc tp ts t
tr
●
The major requirements of the transient response is:
1.Speed of response : Fast reaction of the system to the inputs.
2.Relative stability : Response is free from excessive oscillations, both in size and duration.
δr
0 td tc tp ts t
1. Delay time (td): tr
●
Time for the response to reach half of its final value for the first time.
●
y (t d )=0.5 y ss
● Measure of speed of response (td ↘ , speed ↗)
2. Rise Time (tr):
● Time for the response to reach from δr∙yss to (1-δr)∙yss
● Measure of speed of response (tr ↘ , speed ↗)
● Common values for δr are 0 (underdamped systems) and 0.05 & 0.1 (overdamped systems)
● If δr=0, tr = tc (crossing time)
δr
0 td tc tp ts t
tr
3. Peak time (tp):
●
Time for the response to reach the first peak value.
● Measure of speed of response (tp ↘ , speed ↗)
4. Percent overshoot (P.O) or Maximum overshoot (MP):
●
If y ss =1 → M P = y (t p )−1
y (t p )− y ss
If y ss ≠1 → P.O= x 100[%]
y ss
● Measure of relative stability (P.O or MP ↗ , relative stability ↘ )
δr
0 td tc tp ts t
tr
5. Settling time (ts):
● Time for the response to reach and stay within a range ±δs∙yss
● Generally δs is 0.02 or 0.05
● Measure of relative stability (ts ↗ , relative stability ↘ )
●
Speed and Relative Stability requirements conflict with each other !!!
●
Any measure taken to increase the speed of the system usually results in a decrease in its
relative stabilityand vice versa.
2
(T 0 s+1) ω n
G (s)= 2 2
s +2 ζ ω n s+ ω n
●
Roots of characteristic equation (i.e. Poles)
Im
ζ=0
underdamped 0 < ζ < 1
ωn
0<ζ<1 P1
ωd
P 1,2=−ζ ω n ∓ j ω n √ 1−ζ 2
ωn
ζ=1
real imaginary β
β 0 Re
ω d =ω n √ 1−ζ 2 ζ>1 ωn
-ωd
-ζ ωn P2
-ωn
sin β = √ 1−ζ 2 cos β =ζ tan β =
√ 1−ζ 2
ζ ζ=0
y (t )=1−e
−ζ ω n t
( cos ω d t +
ζ −T 0 ω n
√ 1−ζ 2
sin ω d t
)
or
where
1 /2
A=
[ 1−2 ζ ω n T 0 +(ω n T 0 )2
1−ζ 2 ] and φ =tan −1
[√ ]
ζ −T 0 ω n
1−ζ 2
1
for T 0=0 → A=
√ 1−ζ 2
and φ =tan−1
[ ζ
]
=
√ 1−ζ 2 2
π −β
−ζ ω n t r
from (2) → 1=1− A e cos(ω d t r −φ )
1
→ cos(ω d t r −φ )=0 → ω d t r −φ = π → t r= ω ( π + φ )
−ζ ω n t r
→ Ae cos(ω d t r −φ )=0
2 d 2
−1
1 π− β π−cos ζ
If T 0=0 → t r= ω (π/2+π/ 2− β ) → t r = ω =
d d
ω n √ 1−ζ 2
ẏ (t p )=0
−ζ ω n t p −ζ ω n t p
from (2) → ẏ (t p )= A ζ ω n e cos(ω d t p −φ )+ A ω d e sin (ω d t p −φ )=0
1
If T 0=0 → t d = ω (π /2+ β +π/ 2− β ) → t p = ωπ
d d
→ M P =−A e sin( β )
2
−ζ −π ζ MP = M Im
( π +β + π − β )
1 √1−ζ 2 2 2 √1−ζ 2
If T 0=0 → M P= e sin ( β )=e
√ 1−ζ 2 MP > M
MP < M β*
● For small MP (i.e. better relative stability) Re
β*
● ζ should be large (Conflicting with the requirement for tr and tp !!!)
MP > M
● MP is independent of ωn !!!
MP = M
If MP ≤ M is desired, where M is a given value:
−π ζ
√1−ζ 2 −π ζ √ 1−ζ 2 ≤ −π π =β *
M ≥e →
√ 1−ζ 2
≤ln M →
ζ ln M
→ β ≤tan−1
[ ln (1/ M )]
MAK 305 System Dynamics and Control Dr. Yiğit Taşcıoğlu
http://groups.google.com/group/mak305 Spring 2014-2015
SPECIFICATIONS FOR AN UNDERDAMPED SECOND ORDER SYSTEM
● Settling Time (ts)
y(t)
1+A
−ζ ω n t
1+ A e
δs
The response curve always remains
1
within a pair of envelope curves !!!
δs
−ζ ω n t
1− A e
0
1-A ts t
−ζ ω n t s −ζ ω n t s
from the top curve → 1+ A e −1=δ s from the bottom curve → 1−(1− A e )=δ s
−ζ ω n t s δs 1 A
Ae =δ s → −ζ ω n t s =ln → t s= ln
A ζ ω n δs
1 1
If T 0=0 → t s= ln
ζ ωn δ √ 1−ζ 2
s
ln (1 /δ s ) ln (1/ 0.02) 4
t s= = = for 2%
ζ ωn ζ ωn ζ ωn
ln (1 /δ s ) ln (1/ 0.05) 3
t s= = = for 5%
ζ ωn ζ ωn ζ ωn
● The above expressions for ts are fairly good for 0 < ζ <0.9
● For small ts (i.e. better relative stability)
● ζ should be large
● ωn should be large Im
ts < T ts > T
ω 2n
G (s)= 2 2
s +2 ζ ω n s+ ω n
a)Determine and plot the permissible region of poles in the complex plane so that the system
response to a unit step input has less than 20% overshoot occuring at a time shorter than 5 s with
a settling time less than 8 s using 2% criteria.
b)Determine the values of ζ and ωn so that MP=0.05 and ts=2 s (δs=2%)
c)Determine the corresponding tp and system poles and mark the locations of the poles on the same
plot drawn for part (a).
4 4
t s =2s (2%) → 2= → ω n≥ =2.9 rad/s
ζ ωn 2⋅0.69
c)
t p= ωπ = π = π =1.5s
d
ω n √ 1−ζ 2
2.9 √ 1−0.69
2
a
q
aj r
b k (s +ζ k ω k )+c k ω k √ 1−ζ 2k
Y (s)= + ∑ +∑
s j =1 s+ p j k =1 s 2 +2 ζ k ω k s+ ω 2k
or in time domain:
q r
steady-state y(t)=a+ ∑
j=1
aje −p jt
+ ∑ e−ζ
k =1
k ωkt
[ b cos(ω √ 1−ζ
k k
2
k t)+c k sin( ω k √ 1−ζ 2k t) ]
output
1st order responses 2nd order underdamped responses
(exponential decays) (exponential decaying oscillations)
MAK 305 System Dynamics and Control Dr. Yiğit Taşcıoğlu
http://groups.google.com/group/mak305 Spring 2014-2015
TRANSIENT RESPONSE OF HIGHER ORDER SYSTEMS
1.The necessary and sufficient condition for a system to be stable is that all its poles must have
negative real parts.
2.The poles closer to the imaginary axis dominate the system behavior (DOMINANT POLES)
3.Zeros cancel the effect of poles near them.
If the ratio of the real parts is more than five, and if there is no zero near it,
the pole further away from the imaginary axis may be neglected.
Im y(t)
Re
0
0 t
Im y(t)
Re
0
0 t
MAK 305 System Dynamics and Control Dr. Yiğit Taşcıoğlu
http://groups.google.com/group/mak305 Spring 2014-2015
EXAMPLE
A control system has the following closed-loop transfer function:
Y (s) 10 s +29
G (s)= = 2
X (s) ( s +2 s+2)(s2 +28 s+75)
10(s+2.9) 10 0.4
G (s)= ≈ =
(s 2 +2 s+2)( s+3)(s +25) (s 2 +2 s+2) 25 s2 +2 s +2
2 1
ω n=√ 2 , ζ = = =0.707<0.9
2ω n √2
4 4
for 2% t s= = → t s=4
ζ ωn 1
It is desired to have an underdamped closed -loop system with the following requirements.
● Frequency of damped oscillations (ωd) is between
(ω )
d 2 rad/s and 3 rad/s,
● Damping ratio (ζ) between 0.7 and 0.8.
a) Show the region on the complex s-plane satisfying both of these requirements.
b) For the step response of this closed-loop system, determine and indicate the pole locations on your
plot to satisfy the following requirements.
i. Minimum undamped natural frequency (ωn)
ii. Minimum rise time (tr)
iii. Smallest maximum overshoot (MP) and settling time (ts)
c) Find KP and KT for the smallest maximum overshoot and settling time.
d) If the controller gain is adjusted so that the undamped natural frequency is 4.2 rad/s, where can the
closed-loop poles be located in order to satisfy the requirements given above.
MAK 305 System Dynamics and Control Dr. Yiğit Taşcıoğlu
http://groups.google.com/group/mak305 Spring 2014-2015
SOLUTION
Im
ωd = 4.2
ζ=0.7
ζ=0.8
b) pole ii a) allowable region
b) pole iii for poles
ωd = 3
d) allowable region
for poles b) pole i
ωd = 2
β=45.57o
Re
-4.00 -2.94 -2.67 -1.96 β=36.87 o