0% found this document useful (0 votes)
84 views67 pagesch5 Gtu
Uploaded by
Rohit JasoliyaCopyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF or read online on Scribd
0% found this document useful (0 votes)
84 views67 pagesch5 Gtu
Uploaded by
Rohit JasoliyaCopyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF or read online on Scribd
/ 67
troduction
5.1 In Ms.
Time response of a system is defined as the output of a system when subjected to an input which
is function of time. Time response analysis means subjecting the control system to inputs that ag
functions of time and studying their Outputs which are also functions of time.
5.1.1. Time Response :
‘A control system generates an output or response for a given input. The input represents the
‘sired response while the output is the actual response of the system. Let us take an example of a
levator. When we press the 11" floor button of the lift, the input is provided to the system. The motor
inthe elevator starts moving. In the interest of passenger safety we would not like the elevator to mime
the suddeness of the input response and shoot off to the 11" floor. We would instead prefer that the
clevator reaches the desired floor in a smooth manner. The other important thing is that we would like
the elevator to come exactly in alignment with the desired floor. None of us like it when the elevator is
Stuck between two floors. Let us list down the two specific requirements for passenger comfort that the
levator should achieve.
1. The elevator should gradually rise to the desired floor.
2 Itshould stop exactly at the desired floor.
We draw a few elevator time responses based on our discussion.
1. Fig. 5.1.1(@): In this the output rises very gradually. The elevator is extremely slow. Apart from
this the elevator stops between the 10" and 11" floor.
2. Fig.5.1.1(b): Here the response of the elevator is faster than that in Fig. 5.1.1(a) which is good,
but the elevator still stops between the 10” and the 11" floor.
3. Fig.5.1.1(@): In this the response is very fast and it initially oscillates between the 10” and the
i 12 floor before settling down on the 11" floor.
Fig 5.1.1(d): This represents the best response. The elevator rises in a smooth manner and
stops exactly at the 11" floor, ;
Each of these four figures is the time response of the control system (In this case elevator.)
Scanned with CamScanner Time R
EB controt Engineering (GTU-Mech,) 52 Ime Response Aaj,
Elevator response
*Timo
© @
Fig. 5.1.1 : The dotted line represents the input while the bold line represents the output response
(elevator response)
As defined earlier, Time response is the response of control system as a function of time.
Letus take the time response in Fig. 5.1.1(b) and define a few terms.
Floor
Input signal
response tats eetnao
Fig. 5.1.2
The entire time response is divided into two parts (I and 11). In part I, the output i ing wit
: pane 5 , tput is chi with
respect to time while in part HL, the output is almost constant. Part I is called the transient reponse whl
Part II is called the steady state response. Hence transient response is that part of the time response
where the output is changing with respect to time while the Steady state response is that f the tit
response where the output is more or less constant, artofthe tine
We note that the output settles down at 10.5 instead i
e 0 . of 11. Thi i
error and is associated with the steady state response (more on this late) thre ey sm
Let us identify the transient response, th
Fig. 5.1.3(a) and Fig. 5.1.3(b). Pome, the steady state response and the steady state error in
Scanned with CamScanner
) (Contr eee teem oe 5-3 Time Response Analysis
‘Steady state error = 0
‘Steady stato
error = 0
Transient
response
Steady
State responsa™ Transient —r#— Steady —>
response state response
® (b)
Fig. 5.1.3
note: The transient response may be exponential or oscillatory.
Hence the total time response, c(t) is the transient response c(t) followed by the steady state
response Cy (1).
c(t) = transient response + steady state response |
=e) = (+e, |
We shall discuss cach of these in detail. But before doing that we should spend some time
discussing the input signal.
: 7 Inputs Supplied to a System :
ei Ze
For time response analysis of control systems, we need to. subject the system to various test
inputs. Test input signals are used for testing how well a system responds to a known input. It is not
necessary to choose complicated input signals to analyse a system behaviour.
Some of the standard test signals that are used are :
1. Impulse 2. Step 3. Ramp
4. Parabola 5. Sinusoid
1. Impulse input :
Impulse represents a sudden change in input. An Impulse is 8(t)
infinite at t = 0 and zero everywhere else. The area under the
curve is 1. A unit impulse has magnitude 1 at t= 0
r=5)
1 feo Time
=0 t#0 Fig. 5.2.1
In the Laplace domain we have
Li] = LiSw)=1
Impulse inputs are used to derive a mathematical model of the system.
Step input
A step input represents a constant command such as position. The input given to an elevator is a
Sep input. Another example of a step input is setting the temperature of an air conditioner.
Scanned with CamScanner [FP contro! Engineering (GTU-Mech,) 5-4 Time Response Aan,
A step signal is given by the formula, «)
r=u) = A t20 A
=0 otherwise
IfA= 1, itis called a unit step. '
Tn the Laplace domain, we have Fig. 5.22
A
Lin] = R(s)=¢
Incase of a unit step, we get L[ 1()] =R(s) = 1
3. Ramp input :
The ramp input represents a linearly increasing input command. It is given by the formula,
nt) = At 120; Here A is the slope.
=0 t<0
If A = 1, itis called a unit ramp.
In the Laplace domain we have ,
A
Lir®] = Ros) ="
s Time
Incase of unit ramp, we have R(s) = 5 Fig. 5.2.3,
s
Systems are subjected to Ramp inputs when we need to study the system behaviour for linearly
increasing functions like velocity.
4, Parabolic input :
Rate of change of velocity is acceleration. Acceleration is a parabolic function. It is given by the
formula, rt)
rt) = ae t20
= 0 t
Fig. 5.2.5
We had stated in the earlier section that the entire time response is made up of two parts viz;
t response c(t) and steady state response ¢,(t)
ie. c(t)=e,(t) +e,(t)
‘We shall discuss the steady state response first as it is easier and then follow it up by transient
response.
53” Steady State Response :
BAA
‘The steady state response is that part of the output response where the output signal remains
constant. The parameter that is important in the steady state response is the steady state error (¢,,). In
fact the steady state error is the only thing that characterizes the steady state response.
Error in general is the difference between the jy, Ets)
input (desired) and the output (actual). Steady state
error is the error at t °°.
e = lim (Error)
toe
We shall now derive the formula for steady state error.
Consider a simple closed loop system shown in Fig. 5.3.1. Fig. 53.1
‘The notations used are as follows :
R(s) — Laplace of input r(t) [Here r(t) will be step, ramp and parabola.]
C{s) > Laplace of output e(t)
Bis) — Laplace of feedback signal b(t)
E(s)_— Laplace of error signal e(t). This is generalized error. ic. the difference between
input and output. This is not the steady state error that we are interested in,
G(s) —> Transfer function of main system
H(s) — Transfer function of feedback element
From the Fig. 5.3.1 we have
E(s) = R(s)-B(s) But B(s) (9) Hs)
E(s) = R(s)—C(s)H(s) Also C(s) = E(s) - G(s)
“ E(s) = R(s)-E(s) G(s) H(s)
+ BS) [1+ G(s) H()] = Rs)
Scanned with CamScanner 56 Time Response Analysi,
(53.1
Rs
ES) = T+ G(s) H(s)
‘This is the equation for Jeneralized error in the Laplace domain. The steady state error that we
are interested in is the crror that occurs at t — e209) + along time).
ie. (5.32)
ie.
x —
Here e({) is the generalized erro =
We will now make use of the final value theorem in the Laplace transfor
> Tim f() = tim sF(s)
830
toe
rm which is as follows,
‘Which means that the value of f(t) at te» can be found out by multiply the Laplace transform
of f(t) with s such that s > 0
lim e(t) = lim sE(s) (533)
topes 830
From Equations (5.3.2) and (5.3.3) we have
eq = lim e(t)= lim sB()
ve 830
Ys oe(5.3.4)
ees)
cas im §“TEG6)HO)
‘The advantage of the final value theorem and hence Equation (5.3.4) is that we get the value of
error at t- © from the Laplace transform itself without having to take the inverse Laplace transform.
The next few pages would be devoted to the formula of the steady state error ¢,,. From Equation (5.3.4),
we note that e,, depends on two factors :
1. Input R(s) 2. Open loop transfer function G(s) H(s)
We shall discuss each one in detail.
5.3.1 Effect of Input R (s) on Steady State Error :
We have studied the standard test inputs. We will now see the effect of step, ramp and parabolic
inputs on the steady state error.
Static error coefficient :
© The response that remains after the transient response has died out is called steady state
response.
© The steady state response is important to find the accuracy of the output.
© The difference between the steady stat Sa
ae ly state response and desired response gives us the steady state
‘A control system has the following steady state error
: const ae
and acceleration. These error constants help us to calculate the suaiytanceate zh Poenen rome
()_kp= Positional error constant (i) i,
(ii) j= Acceleration error constant,
k, = Velocity error constant
‘These error constants are called as static e1
. rror coefficients. The i oe
the steady state error. We shall discuss the effet of change in input on eae of era a
steady state error.
Scanned with CamScanner NS
control Engineering (GTU-Mech
57 Response Analysis
(Step input :
Consider a step of itude "AY j, 7
secon ‘magnitude "A" is applied as a reference input tothe system, This means the
rO =A Por t > 0
hs and r() = 0 For t< 0 (5.3.5)
We know that the Laplace transform of a step input is,
A
Ros) = a (5.3.6)
Using Equation (5.3.4) we have,
a s RG)
Bs = soo Teaue +-(53.7)
eer Lt s- Als
ss" s30 1466)-H®
a HM _~asa__
“= s0 T¥G() HO ~~
Here ‘A’ is magnitude, which is constant.
A
we 253.9)
1+, 9 G(s)-H(s)
Letus introduce the term position error coefficient. It is denoted by k, and is defined as,
Li
yt “5 GH) (5.3.10)
Putting this value in Equation (5.3.9) we get, ty
Ge = Teg 65D
This Equation indicates that when a system is
subjected to a step input, the steady state error is
contolled by position error coefficient k,. Refer
Fie. 53.2. Ignore the transient part. The input is shown by
feted line while response is shown by a firm line. Fig. 5.3.2
Note: The initial transient shown as oscillatory could also have been exponential.
Ramp Input :
Consider a ramp input having slope
is applied as a reference input to a system. This ramp
"NPitis expressed as,
r(@) = At For t> 0
and r() = 0 For t< 0 (5.3.12)
The Laplace transform of r (t) is given by,
A
R(s) = L(r@)=9 3.13)
Ut 0 -R@) (53.14)
Using Equation (5.3.4) we have, “590 14+G(HG)
Yon
Scanned with CamScanner 5-8 Time Response Analysis
ntrol Engineering (GT ‘Mech.
Using Equation (5.3.13) in Equation (5.3.14) we get,
Lt s+ Als?
&ss = 5-90 1+G(s) H(s)
Lt Als
es = 590 eae
Lt
—
830 s45°G() HO)
>t"
ess = Lt a
s205ts Ks) ®) oft) A
Lt ky
But 559 8 =O
Cs = Te A +2(5.3.15)
psG@HO
Let us introduce the term, k,. It is the velocity
error coefficient and is defined by,
k= er) s G(s) H(s) (5.3.16) Fig. 5.3.3
Putting this value in Equation (5.3.15) we get,
(53.17)
ek
‘Thus for the ramp input ; the velocity error coefficient k, will control the steady state error.
Refer Fig. 5.3.3. Ignore the transient part. The input is shown by a dotted line while the response is
shown by a firm line.
Note: The initial transient shown as oscillatory could
@) Parabolic Input :
have been exponential.
Consider a parabolic reference input of slope 4. Itis expressed as,
r= $e For tz o
and r(t) = 0 For t< 0 ce
Its Laplace transformis, R(s) = 4 (5.3.
Now the steady state error is given by, (5.3.19)
c= sR(s)
‘ss = Lt
830 T3G@H - As
u )H(S) = “340 THGQH@
Se a
MGOH® “sso Za
>
A GH)
149% 4 2
s0* G(s) H(s)
Scanned with CamScanner +-(5.3.20)
s30* G6)H(s)
Let us define the term, accelerati
Itis denoted by k, and is expressed as,
Lt
k= 5 9° GOH) 253.21)
‘Thus Equation (5.3.20) becomes,
ss = i, (5.3.22)
Thus when parabolic input signal is applied, the
acceleration error coefficient controls the steady state error of
the system.
Fig. 5.3.4
yy a firm line. Ignore the
In Fig, 5.3.4 the parabolic input is shown by a dotted line and response b
transient part.
Note :_ The initial transient shown as oscillatory could also have been exponential.
Summary of steady state error :
The equations of steady state errors in terms of different error coefficients is given below :
ye eG e «for a Step input
(i) Gs = a «for a Ramp input
i) og =e & --ofor a Parabolic input
53.2 Effect of Open Loop Transfer Function G(s) H(s) on Steady State
Error e,,:
Type of System :
As seen in Equation (5.3.4), apart from the input, ¢,, also depends on G(s) H(s). Actually ¢
depends on the Type of the system G(s) H(s). Let us define what we mean by type of a system.
Type :
‘Type of a system is defined by the number of open loop poles i.e. poles of G(s) H(s) that are
Peseular ts origin.
20(s+1)
Hence G(s) H(s) = t sae (a6) i8 4 Type O system as there are no poles at the origin.
S
s+10
#641) (6+2)
; ‘Type 1 and Type 2 systems
Sintaly Gi) HG) = Gare GOH) pees
"spectively.
Scanned with CamScanner [FP contro! Engineering (GTU-Mech 5-10 Time Response Analysis
We shall show that e,, depends on the type of the system and hence depends on the number of
poles of G(s) H(s) present at the origin. The open loop transfer function can be written in two standard
forms - the pole-zero form and the time-constant form, The pole-zero form is written as
K( +2) (8+ %) ---
s"(S+P,) (S+P2) ++
‘The time constant form is written as
K+ 78) + Ts) +
GOMES) = FETS) + T 9)
Here n is the number of poles atthe origin. It is very easy to obtain one from the other. Let us
take an example.
G(s) H(s) =
200 (s +2)
: tl i t =p +2) —
Ex. 53.1: Convert the given transfer function to the time constan form Gis) =F 6 4 4)(6+5)"
Soin. : For the time constant form we write every term as (1 + Ts). Hence we have
200 x2(1+3)
G(s) = Ey ser he
2 s s
Pea(ied)x3( 148)
s
20(1+$)
G(s) = ST ECAY GRA ae form
#(1+9)(144)
We should note that converting the transfer function from the pole zero form to the time constant
form does not change the type of the system. It merely changes the constant (In this ease from 200 fo
20). Since the type does not change, e,, will be the same for both the forms. In our discussion we will
use the time constant form.
k(1+T,,8) (1 + TS).
GO)HO) = Fe Ts) 1+ T8)
Ifn =O, it is a Type O system.
Step: Output for step input *
Ifn=1, itis a Type 1 system. Ramp: i ‘Output for ramp input
Ifn=2, it is a Type 2 system. Parabolic’ (Output for parabolic inputs
Fig. 5.35
We shall now take Type 0, Type 1 and Type 2 je ramp and parabolic
tee Sr 8
5.4/ Subjecting a Type 0 System to a Step, Ramp and Parabolic Input :
5.4.1 Step Input to a Type 0 System :
From the earlier section we know that e,, for a step input is
where k, = lim G(s) H(s)
s0
Scanned with CamScanner ae
Engineering (G'
Bh cota gineering (GTU-Mech.) 511 Time Response Analysi
Kd 4Ty9 (147,
type 0 system G(s) His). =
ee S04 Ty) (14 Ts).
since k, =
k, =
k, =
A
Ss = T+k,
= AL
&= = Tek
Hence when a Type 0 system is subjected to a step input we get a constant steady state error.
geferFig. 5:41
© «y
: |
Fig. 54.1
The output response could also be as shown in
Fig. 54.2. Whether the transient is exponential or
oscillatory will be seen in the later part of the chapter,
but what is definitely known is that there will always
A
besome steady state error ¢,
542 Ramp Input to a Type 0 System :
From the earlier section we know that ¢,5 for a ramp input is
ne wherek, = lim sG(s) H(s)
ky 830
k(1+T,s) (1 +Ty8)
Fora Type O system Gs) H() = “GT (l+ Tas). Ths) dT)
since k, = lim sG(s) H(s)= lim s
s0 ss0
k= 0
Scanned with CamScanner ‘N
Time Resp,
512 Onse A,
Control Engineering (GTU- aly,
ALS Om
Ge = 0
input, the steady state error j
stem to a ramp inpul increag,
Hence when we subject a Type 0 §) stable for ramp inputs.
Continuously. Hence Type 0 systems are not sui
r)
Fig. 543
As stated earlier we are not sure at this moment whether the transient response is exponential op
oscillatory. We, for convenience, assume an exponential transient.
5.4.3 Parabolic Input to a Type 0 System :
Form the earlier section we know ¢,, for a parabola input is ¢,, = 4 ;
where k, = lim s’G(s)H(s)
so
k(1+T,s) (14T,,5)...
For a type 0 system G(s) H(s) = 4,904 T)
z Se
k(+Tys) (1+ Ts
since = lim s’G(s) H@)= li ‘|
Me iy | OLB) tn (1+T 8) (1+ Tyas).
830
k= 0
A
fee
o>
e = ©
When we subject a type 0 system to _ .
Continuously. Hence type 0 systems are not suitable Ensen) input, the steady state error increases
hen the input is parabolic in nature.
1)
Fig. 5.4.4
Scanned with CamScanner — oe
(conta Engineering (GTU-Mech.) 5-13 Time Response Analysis
We thus conclude that Type 0 system gives constant ¢,, when subjected to step inputs and gives
sgt ey When subjected {0 ramp and parabolic inputs. Hence type 0 systems are not suitable for Ramp
fod Pasabobe INPUTS.
nt)
a
Fig. 5.4.5
5.5 Subjecting a Type 1 System to a Step, Ramp and Parabolic Input :
7
‘We proceed in a similar manner as for type 0.
55.1 Step Input toa Type 1 System :
We know that ¢,, for a step input is
ea = te where k, = lim sG(s) H(s)
ss0
k(1+Tys) (1+T,9)...
8 #T,:5) 1 +Ty5).
For atype 1 system G(s) H(s) =
(1+ Tas) 14 Ty
+79) 04 Tyas
since k, = lim G(s)H(s)= lim
840 830
Since there is a sin the denominator
= Teh Teo 6 @& £0
Scanned with CamScanner BF contra Engineoring (GTU-Moc! 5-14 Time Response Analy
As stated earlier, at this stage we are not sure whether the transient response will be exponential
OF oscillatory. We assume it to be exponential only for convenience. Hence type I systems are excellent
for step inputs as the steady state error is zero,
mt)
Fig. 5.5.1
5.5.2 Ramp input to a Type 1 System:
‘We know for a ramp input,
A
a wherek, = lim s G(s) H(s)
s0
k(+Tys) (1+ Ty).
SI +T,8) + Ty).
For a type I system G(s) H(s) =
(1+ Tys) (1+ Ty
(4T 9) 0+ 7,25)
since k, = lim sG(s)H(s)= lim i[
s370 s70
k=
k
A 4 a
k, sk
When we subject a Type 1 system to a Ramp input, we get a constant steady state error.
()
f
Fig. 5.5.2
5.5.3 Parabolic Input to a Type 1 System:
For a parabolic input,
A
wee wherek, = lim s* G(s) H(s)
* 80
Gey = $i ttad+Tys.
For a type 1 system, G(s) H(s) = ‘SC#T, 9) d+ Ts).
Scanned with CamScanner a
ering (GTU-Mech
ges Engineering ( Time Response Analysis
since ky. = lim s? G(s) H() = ee
so
4(1+T,8) (14+ T 9).
since there is a's" in the numerator,
s Keo
A
ae
ky ne, S02
Hence when a parabola input is given to a type 1 system, we get infinite steady state error.
ype I systems are thus not suitable for parabolic inputs.
no)
=]
Time
Fig. 5.5.3
tt)
a
ze
ou
it)
Fig. 5.5.4
rato 1 system gives e,, = 0 for step inputs and constant error for ramp inputs. It is not suitable for
inputs as ¢,, > 00,
Scanned with CamScanner a8 Time Response Anaiysis
(GTU-Mech.
a Type 2 System to a Step, Ramp and Parabolic Input :
5.6 _ Subjecting
\ -
‘56.1 Step Input to a Type 2 System
‘We know for a step input,
+k, rerek, = lim G(s) His)
' 840
kU #Ty9) 1+ T 5).
_ ewe
For atype 2system GOO = TTT 6) 4 Ts).
since, = Him G(s) HG)
k(1+T ys) (1+T,)...
eke io
soo 8 (1+Tys)(1+T,5)-.-
Since there is a s* in the denominator,
k, =
A___A
fe = Tek, 1+
e, = 0
Therefore when a type 2 system is subjected to a step input, it gives 0 steady state error.
) o)
Fig. 5.6.1
5.6.2. Ramp Input to Type 2 System :
For a ramp inputs, we have
Se =k, where k, = lim s G(s) H(s)
s30
Fora 2 system G(s) H(s) =
a S14 Ts) (1+Ty9)...
k(1+T,,s) (1 +T,3).
8 (L+T,8) (1 + Tyo
since k, = lim sG()HG)= if
a0
Scanned with CamScanner g cont Engineering (GTU-Mech,)
517
since there is aS in the denominator,
k= 0
AA
ee BLA
' : kee fm = 0
Hence when we subject a type 2 syst ii
tt) 'ystem to a Ramp ae Bet O steady state error.
‘Type 2
t
Fig. 5.6.2
563 Parabolic Input to Type 2 System :
We know that for a parabola input
A
kK where k, = lim s’G(s) H(s)
80
k(1+Tys) (14+ Ty)...
$(14T,8) (14+ T 9)...
k(1+Tys)(1+T,9).
£(1+T ys) (14 T,9). |
Fora Type 2 system G(s) H(s)
since k, = lim s’ G(s) H(s) =
830
k
A ‘i aA
5 eo”
Hence a type 2 system gives a constant steady state error for parabolic inputs.
nm
Fig. 5.63
‘©can conclude by saying that a Type 2 is excellent for step and ramp inputs and gives constant
Parabolic inputs.
W
efor
Scanned with CamScanner Time Response Ang,
ie
5-18
Fig. 5.6.4
‘Type 2 systems are excellent for step and ramp inputs aS eu = 0. It gives constant error for
parabolic inputs. :
‘Let us draw tables listing all that we have discussed so far.
Table 5.6.1
Sr.No. Constant . Equation ‘Steady State Error
1) _ | Position Error constant Lt A
° © l=. 50 SOHO” [ts “THe
(@) __| Velocity Error constant ky Lt = A
ky = 559 GOH) [Os “ky
{3 _ | Acceleration Error constant ky Lt A
= 550. GOHO [6s =k
Table 5.6.2
[[Sr.No. | Type Step Input ‘Ramp Input Parabolic Input
ky Gs ky Cus ky es
(| TypeZero| — k Any 0 ” 0 ©
T+k
@ [Type One | 0 k a 0 o
k
@ |TyeTwo | 0 5 5 5 a
K
a
Scanned with CamScanner a
trol Engineering (GTU-Mech.) 5-19 Timo Response Analy:
0 Analysis
oo
“64 pisadvantages of Static Error Coefficient Method :
‘ey can be used only for three standard test inputs, ‘They cannot give error for inputs other than
O iese
oe values of error that means °0" and ‘co are specified. The exact valve of error is not
specified.
» This method does not provide any information
0 out the variation of eror with respect otime, Rl) —$ £0) _Jo@|$ + 10)
‘his method is applicable only to the stable system -
pot to unstable systems.
‘This method is applicable only to the configuration He)
as shown in Fig. 5.6.5. Here error constants are “
defined for the loop transfer function G (s) x H (s). Fig. 5.65
9 When the configuration is different than as shown in above point then it is necessary t0 establish
the expression for error signal and apply final value theorem directly without using error
coefficients.
565 Dynamic Error Coefficients :
‘The drawbacks of static error coefficients are overcome by using dynamic error coefficient.
‘ese coefficients are also called as generalised error coefficients.
Wehave the Equation of E (s).
@
RO)
E® = T36@HO
a
- EO = TE@Heo 8 (5.6.1)
Letus express R. H. S. of Equation (5.6.1) as product of two functions F, (s) and R (s) in Laplace
denn.
—_—
Wier, FQ) =1366) HO
E(s) =F, (s) “R(s) w=s(5.6.2)
Now according to the convolution properly of Laplace transform,
1 F(@) =F, (6):
‘Then Laplace inverse of F (s) can be obtained as,
(5.63)
t
fe) = LIF @= IFO Ftv dx
0
3) to Equation (5.6.2) we can,
a Equation (5.6.2) Fy © applying Equation
Laplace inverse of E (s) that means ¢ (t)-
t
6.4)
e() =JF@ORO-o coe
0
Expanding R (t—x) by Taylor series,
1 RQy = RED—xR OFZ, RO
R” (t)
Scanned with CamScanner 5-20 Time Response Analy,
Control Engineering (GT!
dx
t Rew
ee = [reo{Rm-aR 421k"
0
t
t t Lyi. .R”
eW = Tr, qo mayede~ Ja cD 08+ 7 J Fr) ROO 565
0 0
general Equation for error and this error is given for all time from Oto,
Equation (5.6.5) is a
uu
Stendy State Error :Itisgivenby — %s = 490°
u ,
gg R-H.S. of Equation (5.6.5)
Put ky = JF, @)-dx
0
ky == fx Fy (@)- dx and y= f x7F, (x) - dx
0 tO)
(Note upper Limit is co for steady state error)
_ ut
ss ta
‘The next step is to evaluate kp, ky, kp, etc.
Tofind kg: ky = J F, (x)- dx Multiply both sides by e“*
0
k,
° ekg RO +k, R W474 RO om
c kge® = JF, (x)-e°%-dx
o
2 kge"™ = F, (8) (by definition of Laplace) (5.6.6)
Take Ltass—0
ee - Lt
tote ™ = 550% But LHS ato Kye =ky
lt
ky = g 0h 56D
To find k; =
Take derivative of ky ew. 1.8. A oxkge™® = z F,()
t
Put value of ky in LHS. am in @ o*-dked BF, 6 a kee FO)
Lt
7 ox, Ltd
Take Ltass 0 so oO
Scanned with CamScanner Get Engineering (GTU-Mech.) 521 -
Lt
d
KF ss0a FO ean
li @ "
«ilar WA) ky = =
spasiilar 82 2 = so gg Fi) A569)
ye LY PRO
3 830 ds (5.6.10)
ae Lt "F, (s)
wager 830 ast +-o(5.6-11)
ky kv ky ~~» are called dynamic error coefficients.
We shall now solve examples on steady state response and after that discuss the transient
sponse.
566 Examples on Steady State Response :
points to remember :
1, While working on the steady state response, we use the open loop transfer function G(s) H(s).
a 7 G(s)
We do not require the closed loop transfer function 7G) Fay *
Since the input will be given and the type of the system can be found out by mere observation, in
most cases the error coefficients and the e,, can be predicted.
L561: Aunty feedback system has G(s) = wats: Determine : () Type of system (i) All
error coefficients (lil) Error when subjected to a step of magnitude 2.
Sol: Since there are no poles at the origin, the given system is Type 0. Hence it is easy to predict that
K =constant ; k, = 0, k, = 0
2
A
and Gy = T+k, “1+k,
convert the transfer function to the time constant form
We shall now solve the sum. We
10(1+4)
G(s) H(s) = aro(i+s)
iM me system is Type 0
lating the error coefficients :
k= aes G(s) H(s) = i wes
k, = 10 :
x= io sG(@)H@=s Ferrero
k= 0
Scanned with CamScanner eee PONSE Analyy
is,
Control Engineering (GTU- jech.)
10.1 + 8/3)
k, = lims ? G(s) H(S) = § [ial
30
k= 0
i have
(3) Since the input isa step signal of magnitude 2, we M
i fer function to the ti
You would get exactly the same answer without converting the tran: © the time
constant form !!
Let us do it without converting to time constant form
20(s+3)
G(s) Hs) = ahs
+3). 60 _
k, = lim Gis) His)= | ti GED GO F = 10
s30
k, = 10
43)
ky = im 5G H)= tim . [eeheeal
$70
k, = 0
(+3)
k, = lim (9) H()= fim. s as]
$90
o keo0
Since input is step signal of magnitude 2, we have
A _2
fe = T+k,” +10 eq 0.18
Let us draw the response. At this stage, we are not aware of the kind of transient response that the
system would generate. Let us take an exponential as well as a damped oscillation response.
gg = 0.18
(a) Exponential transient
Fig. 5.6. Damped oscillatory transient
Scanned with CamScanner rv
a nee ing (GTU-Mect a
Time Response Analy:
feedback system hes Gio) =—40(642)
a .
pelermine the typo of system, ar on et)
s
6
eae ©rror coefficionts and error for a ramp input with
Gis) Ho) = ~AOG+2)
et S(841) (644)
since there One pole at the origin, itis a type 1 system,
‘
' Jim G(s) H(s)= tim 406 +2)_
a * a0 40 S641 644°"
= lim sG(s) H(s)= lim ¢-22-G4+2)__
woo lim tT e+ n+
= jim G(s) H(@)= lim ¢-—40.8+2
int (8) HO) = lim 8 EE yea =
a. Sealy sate err fora ramp with magnitude 4 is
4.4
Gm = R= 79702
{5831 What information is contained in the specification k, = 1000 ?
1000 implies the following things :
The given system is Type 1, ic. the system has one pole at the origin.
{® Thesystem is subjected to a Ramp input.
The steady state error of the system is going to be a constant ic.,
e, = Alk,
Where A is the slope of the ramp input.
4 = _A/1000
20 "
%564: The control system having unity feedback has G(s) = s7+4s) (T+a)* Determine =
2
(0 Type of systom (i) Static error coeficients (ji) Steady state error if input r) = 2 + At +f
Prediction : Since the given system is type 1, ky =, ky = constant, K, = 0. Also the input is
eta step, ramp and parabola. Since itis a type 1 system, e, for step will be 2670, Cu for ramp
Constant and e,, for parabola will be ©.
4 bie solve the example ,
ince there is one pole at the origin, the system is Type 1.
lating static error coefficients.
. 20,
Fri 6) H)= lin ray TH)
a
Scanned with CamScanner Time Re:
5.24 SPONSE Any
[EP contro Engineering (GTU-MEd [ 20
lim G(s) H) fae 9 71 +45) (1 +5),
5390
k, = 20 os aa
= lim s'G(s)H(s)= lim s liasantea
870 so
k= 0
2
t
(iii) Calculating the steady state error when subjected to input r(t) = 2+At4y
‘There are two ways of doing thi
Method 1:
2
Wt) = 24445
We take Laplace transform.
2.4.1
~ Ro = 244445
‘We use the main formula of e, i.e.
20404
= tim —SRO__ y, fed+3]
se = 1m TGQ) HOA SOT 20
S90 145 (44s) (1 +5)
= li 2 Als us?
= lit ——4_ Is
“i 20
sd+4)(+3) '*s(44)G4+9 | +5445) 05.
- 2
= te bo fim [dg
S991 + qyaydsa| 87° “aS
+ lim 1
$30] 2 s +20
S44 45) (5)
= 29 += 042 400
qe
Method 2
is is a much easier way of calculating the ste
s bi tude 2, ramp of magnitude 4 and la of
‘magni
gni ignitude 1,
paces 0
Scanned with CamScanner 6 at nsnering (GTUMech
Time Response Analysis
6. for parabola =
3)
& = 0%
Acertain Feedback contr
; an rol system is described by the following transfer function,
Gis)= Ferm +30) HO
Determine steady stato error coefficients and
also det
steady state eror to 10 units due to input) = 1+ tor ti the value of k to limit the
agus The steady state error coefficients are,
Lt
y= 550 GOH) =
nk
$905%% +20) (6430) =”
-u.. Lt k
Ky = 549 8: GOH) = 8056420) (6430) =?
Lt sk k
a 7
B= 5 PGK = 505% +20) (5 +30) ~ 600
ee = 10
rt) = 14104200 = 141042 e
+ The input is a made up of a step of magnitude 1, ramp of magnitude 10 and parabola of
2
wee (10)
“Ca should be 10,
Sg, Kit + 25)
5+ For the following type of systom : G(S) = S55) (re omaye: Find val of Ko
‘Steady state error to 10% for an input t.
*(satype 1 system, it will have a finite error to ramp input)
1
Lt __sR(s) as sues
= soTeGeor@ RO =E(WO) °
eee
Scanned with CamScanner Time Respon
5-26 Ponse Anan
Us
ts
sls u KL +25)
2
u sis. fy
es = 5 90THGIMD” $714 TT 04g
u —aiork
&s = 540 k
“vs S#7y9) (14045)
But Cg = 10% ie. 0.1
ee . k =10 osAng,
“Kk
Ex.5.6.7: For the given system, assume the input to be 0.1 t. We need to keep @. < 0.05. Fing the
a range of k for the error to be within the specified limits.
k
Cs)
Fig. P.5.6.7
‘Soln.:
k2
G(s) Hi) = ma"! GOH = Wa
This is @ type I system. Hence it will give a constant error for a ramp input. The given input i
0.1 t which is a ramp, Since it is a ramp, we need to find only k,
k, = lim s G(s) His)
840
i kr
k= li
‘ rae #(1 +822)
(Since ramp has a magnitude of 0.1)
We need to, to be 0.05,
Ot
005 = Ot
kL OL
2 = 005 keg
i -k-=-4, 4 = 0.05
+: Fore, $0.05,weneed keg
Consider the first order system wing contoleras ay
: Shown in Fig,
() Find the steady state error 9, with unit stp ig. P. 5.6.8.
with gan ky. (88° Fig. P. 5.6.5 'nPut using a proportional controller
Scanned with CamScanner Engineering (GTU-Mech.) not
rol
re
Ris) °
) [i ] s)
Proportional
Controlior Plant
Fig. P. 5.6.8(a)
(i) Find the steady state error o,, with unit step input usit it
k. Comment on the results, (see Fig. P. 55.00) a
Res) ow
Intogral
Controller
Plant
Fig. P. 5.6.8(b)
soln:
|) Wefind the steady state error e,, with unit step using proportional controller gain ky,
Re) os)
&) z
Proportional rant
Controller
Fig. P. 5.6.8(¢)
Since the system is subjected to a unit step input , we calculate ky
k= lim G(s) H(s)
s70
1
lim ky Ts +1
530
k,
fo Ao
ee = Trey “TF
) Wenow find the steady state error wit a unit step using a integral controller gain k; «
os)
Fig. P. 5.6.8(4)
Scanned with CamScanner Time Responte Aaj
5:28
Control Engineering (GTU-Mech. is
lim G(s) H@)
4 50
oo
= lim = Feat
590
ky
k= lim Sos+D)
30
k= °o
os = Tk,
e& = 0
Hence the integral controller reduces the steady state error to zero.
Ex.5.6.9: Fora unity feedback control systems shown below, obtain steady state error for step input,
10.
(9) = ges = Ses 1400 60)
Sie) = Tas and Al) = 8 ("+ 148 +50)
Soin. :
Since the system is a unity feedback,
HG) = 1
since the input: A
Since the input isa step input R(6) = (A isthe magnitude)
10
Gy = = 2_
a 09 = Fyhase50
yA
a sR) s
fe = lim TSG() He = Him 10
s9014——10
s+ 148450
= ——A_____aA
= 10 = 10
1+ lim =—"— ,,10
so 8145450 15
b>
& =
_ 10
0) GAS) = SEF 1458450)
s
‘
saeeee Tr" a
10
Scanned with CamScanner
5 &
oe
e given system represent
gant ne ‘ooo eines a er tines ‘emperature has to be maintained at a set
steady state temperature,
oR 20000
The block diagram shown in G@rnts019 aT Oo0e
heating oven. Set point is
1000°C. What is the steady
state temperature.
sie: From the example
Ros)
8)
Fig. P. 5.6.10
20000
= 20
SO = GU+01s) +a FON HHO=1
‘constant set point of 1000° C indicates a step input of magnitude 1000... r(t) = 1000
1000
. RG) = 12
sR)
co = IM TEGQ) HO)
tx
eI lim ———999 ——
5901 +e aH aso1 UF O0EH
1000
e, = lim
20000
sao
