Scribd
Upload
120 views11 pages

Lecture - 23 Industrial Control PDF

1) The Smith predictor configuration is a control technique used to compensate for time delays in processes. 2) It uses a process model to predict the process output, allowing the controller to calculate corrections based on the predicted output rather than the actual delayed output. 3) When the process model is accurate, the Smith predictor removes the effects of the time delay from the closed loop transfer function, improving control performance compared to a standard controller. However, performance deteriorates if the process model is inaccurate.

Uploaded by

287 Jatin
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
120 views11 pages

Lecture - 23 Industrial Control PDF

1) The Smith predictor configuration is a control technique used to compensate for time delays in processes. 2) It uses a process model to predict the process output, allowing the controller to calculate corrections based on the predicted output rather than the actual delayed output. 3) When the process model is accurate, the Smith predictor removes the effects of the time delay from the closed loop transfer function, improving control performance compared to a standard controller. However, performance deteriorates if the process model is inaccurate.

Uploaded by

287 Jatin
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 11

Industrial Control

(EEPC45)

Lecture Notes 23

Unit III
Smith Predictor
L

Gd

Ysp
E E P Y
Gc Gp
+ +


Y 
Y +

1 2
-Td s
G* e -

Y-Y 2

Gv  Gm  1 ; Plant G(s) = G *e-Tds


 * (s)e-Tds  G(s)
Process model : G 
 * (s)
G is to be so chosen that, it is able to predict the effect
of control action on undelayed output. The controller than
uses the predicted response Y 1 to calculate its output signal
. Predicted output signal is also delayed by the amount of
process dead time , for comparison with actual delayed
output Y. This action corrects for modeling errors and
disturbances entering into the process.

E = E -Y (at second comparator)
1
 ) -Y
= Ysp  (Y-Y 
2 1

If the model is perfect:   Y


Y2

and , E =Ysp -Y 1
 controller responds to error signal that would occur
as if no time delay is present.

With a modifucation an alternate form is obtained as:


L

Gd

Ysp E E P Y
Gc G
+ +
 -Y
Y 
1 2
 * (1-e-Tds )
G
P Gc
inner loop TF G  = 
E 
1+G c G*(1-e -sTd
)
Y GG
and 
Ysp 1  GG
GGc
=

1  G c G*(1-e -sTd
)  GGc
 -sTd G
G*e c

 G (G*e
1  G c G*-  -sTd )  GG
c c

 the linear undelayed part of plant T.F. correctly modelled


If G*,
we get ,
Y GcG *e-sTd

*
Ysp 1  G c G

whereas, without the modification, closed T.F. for the plant


Y GcG GcG *e-sTd
would have been  
Ysp 1  Gc G 1  G c G *e-sTd

By comparison , it is observed that , with the modification ,


the characteristic equation of smith predictor is free from
time delay effect !
Unfortunately , the advantage is lost if process model is not
correct i.e. G *  G *
Practically , it works for differences even up to < 30% of
actual value.
{ Reference: Morari & Zafirion- Robust process control
(PH), 1989}

Sample response For unit step set point change


Y

Td=0 Td =2

t
0 5 10 15 20 25 30
Initial time delay will be observed but the difference is
settling time is not much different !!

Smith predictor is effective if changes in model are small.


For more changes , controller performance will
deteriorate.

If time -delay varies significantly, some adaptive controller


must be used.
Regulatory response of Smith predictor

Y Gd Gd
 
D 1+GG Gc
1  -sTd
Gp
1+G c G*(1-e )


G d [1+G c G*(1-e -sTd
)]


(1+G c G*-G  -sTd )  G G
G*e
c c


G d [1+G c G*(1-e -sTd
)]


1+G c G*
A typical response is shown in fig.
Thus, denominator and characteristic equation is same as in
case of SPC problem though the numerator is very
different .

Smith predictor configuration is useful for handling


disturbances. However due to numerator terms, in certain
cases, simple PI may give better performance.

The approach can be improved by incorporating 2 DOF


IMC controller configuration and give better
performance.
PI

Smith

You might also like