2.

Process Dynamic
Oleh Hamid A and Fauzan Amir.

1
Session Outlines & Objectives
Outlines
 What is dynamic modeling?
 Why do we need dynamic models?
 Dynamic modeling methods

 Mathematical dynamic modeling of controlled

processes

Objectives
 The importance of process dynamic
 Know the methods used to model a process
 Know various type of common mathematical dynamic

model of industrial process

2
Dynamics vs. Steady State Model
 Chemical process are generally designed from a
steady-state point-of-view
◦ Steady state model
 Steady state: No further changes in all variables
 No dependency in time: No transient behavior

 Chemical processes are dynamically changing

continuously
◦ Dynamics is the time varying behavior of processes

 Steady-state change indicates where the process is

going and the dynamic characteristics of a system
indicates what dynamic path it will take

3
Why do we need dynamic model?
 People’s emotion: An Everyday Example of Dynamic
Modelling

Do we give the same approach to handle both situations?

• What is your objective?

Laughing Mad

4
Why do we need dynamic model?
Do the bus and bicycle have different dynamics?
• Which can make a U-turn in 1.5 meter?
• Which responds better when it hits a bump?

Dynamic performance
depends more on the
vehicle than the driver!

The process dynamics

are important than the computer
control!

5
Why do we need dynamic model?
Feed material is delivered periodically, but the process
requires a continuous feed flow
• How large should the tank volume be?

Periodic Delivery flow

Continuous
Feed to process

Time We must provide

process flexibility for
good dynamic performance!

6
Why do we need dynamic model?
The cooling water pumps have failed
• How long do we have until the exothermic reactor runs away?

F
Dangerous

Temperature
T L
A
Time

Process dynamics
are important
for safety!

7
Why do we need dynamic model?
 In order to be able to design a controller, to achieve
a specific objective, for the system under concern
we must know which system we are dealing with
◦ The process dynamics are important than the computer
control
 To improve understanding of the process
 To select the right controller
◦ We must provide process flexibility for good dynamic
performance
 To optimize process design/operating conditions
 To accommodate the possibility of the change of process
dynamic  ROBUST!
◦ Process dynamics are important for safety

8
How to Obtain a Dynamic Model?
 Model can be described using verbal, table,
mathematic, etc.
 The rationale for mathematical modeling
◦ To improve understanding of the process
◦ To train plant operating personnel
◦ To design the control strategy for a new process
◦ To design the control law
◦ To select the controller setting
◦ To optimize process operating conditions
 Main mathematically modeling methods:
◦ Fundamental (Physical) modeling
 Based on physicochemical law
◦ Empiric (Identification)
 Based on process data analysis
◦ Semi-empirical models
 Combined approach of fundamental and Empiric

9
 Why do we develop mathematical models?

Input change, e.g., step Process Affect on

in coolant flow rate output variable

T L
A

How does the • How far?

process input Math models
• How fast help us answer
influence the
response? • “Shape” these questions!

10
Fundamental (Physical) modelling (1)
 Total mass balance
[Rate of mass accumulation within CV] = [Rate of mass in from
surroundings] –
[Rate of mass out to surroundings]

 Component mass (molar) balance

[Rate of mass accumulation within CV] = [Rate of mass in from
surroundings] –
[Rate of mass out to surroundings] + [Rate of mass creation within CV]

 Total energy (enthalpy) balance

?
[Rate of energy accumulation within CV] = [Rate of energy in from
surroundings]
- [Rate of energy out to surroundings]

11
Physical modelling (2)
 Illustrative Example: Open Liquid Storage
Tank qiOpen Tank

h
q

12
Physical modelling (3)
 Follows conservation laws
 Requires a specific expertise
 Difficult to develop
 Expensive
 Time consuming
 Too complex model
 Found in the fields that need accurate model,

e.g. aerospace shuttle, aircraft

13
Identification
 Based on the operation data
 Easy to develop
 Requires well designed experimental data
 The behavior is correct only around the

experimental condition
 Obtained model obtained usually quite simple

for control purpose

 Found in the industrial process control area

14
Process Dynamic Characteristics
 Self-regulating and non self-regulating
 Lag
◦ 1st order
◦ 2nd order
◦ Higher order
 Dead time
 Interacting and non interacting

15
Self-regulating process
pump
Levelsensor
Level sensor
dV dL
  A  Fin  Fout
dt dt
Liquid-filled
• In flow is set independent of level
tank
• Out flow is dependent of level

Level

Fin

Fout

time
16
Non Self-regulating process
Levelsensor
Level sensor
dV dL
  A  Fin  Fout
dt dt
Liquid-filled
• In flow and out flow are set independent of level
tank
• Also known as Integrating Processes
pump

Level

Fin

Fout

time
17
1st Order Processes (1)
The basic equation is:

dy (t )
p  y (t )  K p u (t )  Differential equation
dt
 Transfer function
Kp
G p ( s) 
 ps  1  Note that gain and time constant
define the behavior of a 1st order
process.
Kp = steady-state gain
p = time constant

18
1st Order Processes (2)
 Same Kp, different tp  Different Kp, same tp
KpA = KpB = KpC
tpA < tpB < tpC
KpA < KpB < KpC
tpA = tpB = tpC

u(t) u(t)

Always overdamped

19
1st Order Processes (3)

20
 2nd Order Processes (1)
The basic equation is:

2
2 d y (t ) dy(t )
p  2  p  y (t )  K p u (t )  Differential equation
dt 2 dt

Kp  Transfer function
G p (s) 
 2p s 2  2  p s  1

 Note that the gain, time constant, and the damping factor define the
dynamic behavior of 2nd order process

21
2nd Order Processes (2)
 Underdamped vs. Overdamped

underdamped

critical damped

overdamped

22
2nd Order Processes (3)
 Effect of z on Underdamped Response

23
2nd Order Processes (4)
 Effect of z on Overdamped Response

24
2nd Order Processes (5)
 Characteristics of an Underdamped Response
 Rise time
 Overshoot (B)
 Decay ratio (C/B)
 Settling or
response time
 Period (T)

25
2nd Order Processes (6)

26
Underdamped Processes
 Many examples can be found in mechanical and
electrical system
 Among chemical processes, open-loop underdamped
process is quite rare
 However, when the processes are controlled, the
responses are usually underdamped
 Depending on the controller tuning, the shape of
response will be decided
 Slight overshoot results short rise time and often
more desirable
 Excessive overshoot may results long-lasting
oscillation

27
Dead Time
 Fluid transportation through a pipe
◦ Also called as distance-velocity lag, transportation
 = dead time
lag, time delay
Yout (t ) = Uin (t - q )
-q s
Yout (s) = e Uin (s)
Yout

Uin

time
28
 Structure of Process Systems (1)
Non-Interacting Series
 The output from an

element does not

influence the input to
the same element
 Common example is
T
tanks in series with
pumped flow between
v(s)Block diagram as shown

F (s)
0 T (s) 1 T2(s) Tmeas(s)

Gvalve(s) Gtank1(s) Gtank2(s) Gsensor(s)

29
 Structure of Process Systems (2)
Non-Interacting Series
v(s) F0(s) T1(s) T2(s) Tmeas(s)

Gvalve(s) Gtank1(s) Gtank2(s) Gsensor(s)

Y(s) n
In general:  Gi(s)
X(s) i1 • overall gain is
product of gains
With each element a Y
(s) n K i
• no longer first
first order system:  order system
X(s
) i
1(
is 1
) • slower than any
single element
30
 Structure of Process Systems (3)
 High Order Processes
 The larger n, the more
sluggish the process
response (i.e., the larger the
effective deadtime)
n=3  Transfer function:

Kp
PVm

G p (s) 
n=5 n=15
 p s  1
n

time

31
Structure of Process Systems (5)

5
Controlled Variable

0
0 5 10 15 20 25
time

5
Manipulated Variable

0
0 5 10 15 20 25
time

32
Structure of Process Systems (7)
 Two plants can have different intermediate
variables and
have the same input-output behavior!

33
Structure of Process Systems (8)
Interacting process
 Many chemical processes exhibit interacting

nature
 The output from an element influence the

input to the another element and vice versa

34
Session Summary
 A good understanding on controlled process
dynamic is very important to achieve a good
control
 Process models can be obtained using
physical, identification and combined
approaches
 Many industrial processes can be represented
by simple mathematical process dynamic, i.e.
1st order, 2nd order (with dead time)

35