Dynamic process behaviour

Process Control Loop

Process control loop 1st Order or 2nd Order or Higher order??

Pressure
Flow
Level
Process Temperature
pH

dP cell
Capacitance I/P
Sensor Radar, Sonic
Magnetic
Resistance
IR/Laser

4-20 mA
Transmitter 1-5 Vdc

PID
Controller Fuzzy logic

4-20 mA
Transducer 3-15 psig

Linear
Control valve Equal percentage

© Abdul Aziz Ishak, Universiti Teknologi MARA Malaysia (2009)

Common input changes
1. Step Input

A sudden change in a process variable can be approximated by

a step change of magnitude, M:

The step change occurs at an arbitrary time denoted as t = 0.

• Special Case: If M = 1, we have a “unit step change”. We

give it the symbol, S(t).

• Example of a step change: A reactor feedstock is suddenly

switched from one supply to another, causing sudden
changes in feed concentration, flow, etc.
 Response of first order system
• First order differential equation
dy (t )
a1  a0 y (t )  bx(t )  c
dt
dY (t )
a1  a0Y (t )  bX (t )
dt   a1 / a0
dY (t ) K  b / a0
  Y (t )  KX (t )
dt
• General first order transfer function

 K 
Y (s )    X (s )
s  1
 Process Characteristics of First
Order System
 K 
Y (s )    X (s )
s  1

Process characteristics : K and 

K = process gain, - model parameter that describes how much
the measured output (Y) changes in response to changes
in the input (X)
 = time constant, – value of time which the response
is 63.2% complete
•All first order systems forced by a step function will have
a response of this same shape.

 K  x
Y (s )   
Output response with Step function
s  1 s as the input
 • Process Gain can be used to determine the ultimate value of the
response (new steady state) for a unit step change in the input
• To calculate the process gain from the graph:

y
Process Gain, K 
x

New steady state

Old steady state

• To determine the ultimate value of output
from transfer function
Y= Gain x change of magnitude X
 Exercise 1
• Consider a first order process with the transfer
function
 4 
Y ( s) / X ( s)  
 5s  1
Determine the ultimate value of Y when a step
change with magnitude of 4 is introduced into the
process
Response with time delay

X(t)

Y(t)

t=0 t=t0

Θ=Time delay/dead time

First-order-plus-dead-time (FOPDT)

 Ke st0 
Y (s )    X (s )
 s  1 

t 0 = time delay or dead time, – delay from when an input

(X) signal is issued until when the output (Y) begins to
respond
Process characteristics
Process deadtime (time delay)
Period of time between the starting of input (load) change and the intersection of tangent line
and old steady state baseline. (Zieglar-Nichols, 1942)

PV ,% new s.s.
Upper base line

old s.s.
Lower base line
Td

time
MV, %
MV

time

© Abdul Aziz Ishak, Universiti Teknologi MARA Malaysia (2009)

Process characteristics
Time constant
Tangent and point
A period of time between the [email protected]% and the time at cross-section of old s.s. baseline and
tangent line.
TC,63.2% = [email protected]% - Td

PV ,% new s.s.
PV63.2%

TC,63.2%
old s.s. [email protected]%
Td
time
[email protected]%
MV, %
MV

time
© Abdul Aziz Ishak, Universiti Teknologi MARA Malaysia (2009)
Exercise 2

Estimate Td, TC and gain of the process

Summary of result

45
PV, %

T d, s TC,63.2% Gain of the

40 B C process

35

30
Transfer function=
25

20

15

10 A
0 20 40 60 80 100
Time, sec

© Abdul Aziz Ishak, Universiti Teknologi MARA Malaysia (2009)

Exercise 2

Step by step guide

1. Locate point of max. gradient.
45
PV, %

2. Draw a tangent line at the max. gradient.

40 B C 3. Draw a straight line at Pvoss (lower base line)

4. Draw a straight line at Pvnss(upper base line)

35
5. Calculate 0.632 (PVnss – PVoss)
30 6. Draw a straight line from [email protected]% to
the response curve
25 7. Bring the line to x-axis (Time)

20 7. Time @ [email protected]% - Td =Tc

15

10 A Summary of result
0 20 40 60 80 100
Time, sec T d, s TC, s Gain

© Abdul Aziz Ishak, Universiti Teknologi MARA Malaysia (2009)

 Response of second order system
• Second order differential equation
d 2 y (t ) dy (t )
a2 2
 a1  a0 y (t )  bx(t )  c
dt dt a2
 
a0
d 2Y (t ) dY (t )
a2 2
 a1  a0Y (t )  bX (t ) 
a1

a1
dt dt 2a0 2 a0 a2

d 2Y (t ) dY (t ) b
K
 2
2
 2  Y (t )  KX (t ) a0
dt dt

• General second order transfer function

8/3s2+5s+1
 K 
Y (s )   2 2  X (s )
 s  2s  1
Composed of two first order subsystems (G1 and G2)

K1K 2
Y (s ) 
 1s  1 2s  1
K  1 overdamped
G(s) = 2 2
 s  2s  1 0   1 underdampe d
  1 2  1 critically damped

1   2 2nd order ODE model

=
2 12 (overdamped)
K
    12 G(s) =
roots: 1 2s 2 + (1   2 )s + 1

1. Step response

Y (s )  2 2
K K x
 s  2s  1 
X (s ) Y (s ) 
 
 2s 2  2s  1 s
 Second Order Step Change
a. Overshoot – fraction of the final steady-state change
by which the first peak exceeds this change
a   
os=  exp  
b  1 2 
 
b. time of first maximum-time required for the output
to reach its first maximum value

tp  8/(3s2+5s+1) 8 / (3s2 +5s
1 2
+1)

c. decay ratio-ratio which the amplitude of the sine

wave is reduced during one complete cycle
c  2  a 2
 exp   2
a  1 2  b
 
d. period of oscillation, P – time between two
successive peaks of the response.
2
p
1 2

e. Rise time, tr – time taken for the process output to

first reach the new steady state value.

f. Settling time – time it takes for the output to come

within a band of the final steady-state value and
remain in this band
 Exercise 3

• a, b and P in the second order response are 20ºC, 50ºC and 6 s,

respectively. Determine the second order transfer parameters and
write down the complete 2nd order transfer function.
Unit 1: Terminology
Terminology
Process response curve: The trend of process measurement.
Rise time
Overshoot
PV
Period

SP2

Undershoot

SP1
Settling time

0 Time
1. Settling time is the period of time taken by the process reach stability.
2. Period is the time interval between two succesive peaks.
3. Overshoot is the process response (PV) surpasses the set point.
4. Undershoot is the process response (PV) goes below the set point.
5. Risetime is the period of time taken by the process to reach the setpoint.
© Abdul Aziz Ishak, Universiti Teknologi MARA Malaysia (2009)
process responses under automatic control.
Terminology
Ideal response: The desired process response is achieved at an instantaneous time.

Ideal
response
PV2

PV1

SP2

SP1

Time

© Abdul Aziz Ishak, Universiti Teknologi MARA Malaysia (2009)

Terminology
Stable: The process response stabilized at (near) the set point .

Ideal
response
PV2

PV1

SP2

SP1

Time

© Abdul Aziz Ishak, Universiti Teknologi MARA Malaysia (2009)

Terminology
Unstable: The process response could not be stabilized at the set point.

Ideal
PV2 response

PV1

SP2

SP1

Time

© Abdul Aziz Ishak, Universiti Teknologi MARA Malaysia (2009)

Terminology
Quality limits: A range, set values above and below the set point, whereby the process
is allowed to oscillate. Product quality is acceptable within these limits.

PV
Out of spec
UCL
SP2

LCL
Out of spec
SP
1

0 Time
1. LCL = Lower control (quality)
limit.
2. UCL = Upper control (quality)
.
limit

© Abdul Aziz Ishak, Universiti Teknologi MARA Malaysia (2009)

Various shapes of process responses under automatic
control.

QAD Oscillatory

Underdamped

Overdamped

Offset

© Abdul Aziz Ishak, Universiti Teknologi MARA Malaysia (2009)

Unit 1: Process settling criteria
Terminologies
Settling criteria: A response curve that meet any of the following criteria (criterion)
is considered settle.

1. Response time
2. Settling time
3. Rise time
4. Quarter Amplitude Damping (QAD)
5. Quality limits (BEST for product quality control)
6. No overshoot or no undershoot (BEST for
temperature and pH control)
7. Minimum IAE, ITSE, etc.

© Abdul Aziz Ishak, Universiti Teknologi MARA Malaysia (2009)

 Unit 1: Introduction
Let us look at a simple PID control loop.
Assuming all instrumentations have been configured correctly, the process has
been designed accordingly and installation has been made as per manual /
guidelines.

Characteristics PB, I & D An optimum PB, I & D

Optimum would drive the process
PB, I & D? toward the setpoint.
Fine Tuning


Tuning Rules

RR, Td & Tc

I/P
Openloop Test

© Abdul Aziz Ishak, Universiti Teknologi MARA Malaysia (2009)