Department of Chemical Engineering

BMS College of Engineering

Process Control Engineering

Review of systems:

Review of first order system:

Examples:

1. Liquid tank system

Change of height of the tank (output) and change of inlet flow rate (input) related by a
first order differential equation.

2. Thermometer & thermocouple

Change of thermometer temperature (output) and Change of environment temperature

(input) related by a first order differential equation.

3. Mixing process

Change of solute outlet concentration (output) and Change of inlet concentration (input)
related by a first order differential equation.

4. Stirred Tank Reactor

Aim:

How output varies with time when system experienced change in input.

Definition: Any system whose input and output are related by a first order differential equation
is called first order system.

General form of first order system:

1
Prepared by: Soumen Panda
 Department of Chemical Engineering
BMS College of Engineering
Process Control Engineering

Where,

We need to know how y and f related each other

Apply Laplace transform to get relation between y and f

( ) ( )

Apply Laplace transform

̅ ̅ ̅

y’(0) = 0 [ at steady state (t = 0) change of y and derivative of change of y is zero]

̅
̅

Transfer function: Transfer function is defined as the ratio between output and input in Laplace
domain

f’(t) is the change of input. Convert f’(t) from time domain to Laplace domain and find y’(s)

̅ ̅

Apply inverse of Laplace and find y’(t) [Change of y in time domain]

f’(t) ̅
Unit impulseδ(t) 1
Unit pulse u(t) 1/s
Linear t u(t) 1/s2
2
Quadratic t u(t) 2/s3
Exponential e-at u(t) 1/(s+1)
Sin sin(at) u(t) a/(s2+a2)
Cos cos(at) u(t) s/(s2+a2)

If f’(t) = B u(t)
2
Prepared by: Soumen Panda
 Department of Chemical Engineering
BMS College of Engineering
Process Control Engineering

̅ . /

. /

Special case:
If b = 0;

̅
̅

̅ ̅

If f’(t) = B
̅

3
Prepared by: Soumen Panda
 Department of Chemical Engineering
BMS College of Engineering
Process Control Engineering

Thermometer:

Assumptions:

1. All the resistance to heat transfer resides in the film surrounding the bulb (i.e., the
resistance offered by the glass and mercury is neglected).
2. All the thermal capacity is in the mercury. Furthermore, at any instant the mercury
assumes a uniform temperature throughout.
3. The glass wall containing the mercury does not expand or contract during the transient
response.

Apply energy balance in the thermometer,

The rate of flow of heat through the film resistance surrounding the bulb causes the internal
energy of the mercury to increase at the same rate.

h depends on surrounding fluid flow and properties and bulb dimension.

At steady state,

In deviation form,

( )

( )

Where,

4
Prepared by: Soumen Panda
 Department of Chemical Engineering
BMS College of Engineering
Process Control Engineering

̅ ̅̅̅

̅
̅̅̅

= transfer function of the system. It is ratio of the Laplace transform of the deviation in the
thermometer reading to the Laplace transform of the deviation in the surrounding temperature. It
relates the cause and effect.

If Ta’(t) = Bu(t)

̅ . /

. /

Note: . /

( )
Slope of graph is
( )

( )
Initial slope
( )

5
Prepared by: Soumen Panda
 Department of Chemical Engineering
BMS College of Engineering
Process Control Engineering

Plot y(t) vs t/𝞃 :

0.95
0.865

0.632

Note:

1. At t = , y = 0.632Bkp

2. If the initial rate of change of output were maintained, the response would be completed
in one time constant.

. /

Take derivative w.r.t. time

0 1

Straight line with slope and originates from 0,

P is 1 when t=

6
Prepared by: Soumen Panda
 Department of Chemical Engineering
BMS College of Engineering
Process Control Engineering

Significance of time constant and steady state gain

Higher time constant – slower response (reaches new steady state faster)

At new steady state change of output is equal to change of input

Problem 1. A thermometer having time constant of 0.1 min is at steady state temperature of 90
0
C. At time t = 0, the thermometer is placed in a temperature bath maintained at 100 0C.
Determine the time needed for the thermometer to read 98 0C.

Solution:

Time constant τp= 0.1 min. Steady state temperature Ts = 90 0C.

Sudden change of temperature bath = change in input (step Input) B = 100 – 90 = 10 0C.

Kp= 1.

. /

Need to find time when temperature rises from 90 to 98 0C.

Hence T(t) = 98 – 90 = 8 0C

Therefore, 8 = (10)(1)(1 – exp(-t/0.1))

t = 0.161 min

Problem 2. A thermometer of time constant 10 s, initially at 30 0C, is suddenly immersed into

water bath at 100 0C. How long will it take for the thermometer reading to reach 90 0C.

Solution:

We need to determine time when thermometer reading reaches 90 0C.

Hence T(t) = 90 – 30 = 600C

B = 100 – 30 = 700C

Therefore, 60 = (60)(1)(1 – exp(-t/10))

7
Prepared by: Soumen Panda
 Department of Chemical Engineering
BMS College of Engineering
Process Control Engineering

t = 19.5 s.

Problem 3. When a thermometer at 30 0C is placed in a water bath at 900C, the initial rate of rise
in thermometer temperature is found to be 20C/s. What is the time constant of the thermometer?

What will thermometer read after one minute?

Solution:

B = 60 0C

. /

( )

(60/ )(1) = 2

After one minute (60 s) temperature

T = 30 + (60)(1-exp(-60/30)) = 81.9 0C

8
Prepared by: Soumen Panda
 Department of Chemical Engineering
BMS College of Engineering
Process Control Engineering

Response of first order system with impulse input:

Impulse function: Mathematically, the impulse function of magnitude A is defined as

Where, is the unit impulse function

A first order system can be expressed as

( )

With the application of impulse input of magnitude A

( )

9
Prepared by: Soumen Panda
 Department of Chemical Engineering
BMS College of Engineering
Process Control Engineering

Response of first order system with sinusoid input:

The function is represented mathematically by the equation

, where, A is amplitude, is frequency in radian,

frequency (cycles per time)

This equation can be solve for Y(t) by means of partial fraction. The result is

The equation can be written in another form applying the trigonometric identity

Where,

10
Prepared by: Soumen Panda
 Department of Chemical Engineering
BMS College of Engineering
Process Control Engineering

As t , the first term on the right side of the equation vanishes and leaves only the ultimate
periodic solution, which is sometimes called the steady state solution

Comparing the result with the input function, we see that

Response of first order system with Pulse input: (Scan)

Response of first order system with Ramp input: (Scan)

11
Prepared by: Soumen Panda
 Department of Chemical Engineering
BMS College of Engineering
Process Control Engineering

Liquid tank system:

Assumptions: Constant liquid density, Outlet flow is linearly varies with tank level.

Apply mass balance,

At steady state,

Multiply both sides by ‘R’

Where,

̅ ̅

12
Prepared by: Soumen Panda
 Department of Chemical Engineering
BMS College of Engineering
Process Control Engineering

If (t) = Bu(t)

̅ . /

. /

The ultimate change in for a unit change in input is ‘R’

R is simply the conversion factor that relates h(t) and F0(t) when the system is at steady state.
This R is called steady state gain. This is the ratio between change of output and change of input
at steady state.

The above system is a self-regulating system because the output is limited for a sustained change
in input. For a sustained change in input the height will change with time till inlet and outlet flow
become same.

Relate inlet and outlet flow

̅ ̅

Substitute ̅ ̅

̅ ̅

If (t) = Bu(t)

̅ . /

Here, Kp=1 (dimensionless) which is expected because the input and output variable have the
same unit.

13
Prepared by: Soumen Panda
 Department of Chemical Engineering
BMS College of Engineering
Process Control Engineering

Special Case:

At steady state,

̅ ̅

If Fi’(t) = Bu(t)
̅

Here response grows without limit for a sustained change in input. Such a system called non-
regulating system. A first order self-regulating system with infinite resistance behaves as non-
regulating system.
̅

If
̅

14
Prepared by: Soumen Panda
 Department of Chemical Engineering
BMS College of Engineering
Process Control Engineering

Mixing process:

Assumptions:
Well mixed system
Constant hold up
Inlet flow rate = out let flow rate = q (volumetric rate)
Mole balance

At steady state

In terms of deviation

Assume V = constant

̅ ̅
Where,

Linearization (Scan)

15
Prepared by: Soumen Panda
 Department of Chemical Engineering
BMS College of Engineering
Process Control Engineering

Stirred Tank Reactor (STR)

Assumptions:
Well mixed system
Constant hold up
Inlet flow rate = out let flow rate = q
First order reaction
Mole balance

At steady state

In terms of deviation
̅̅̅
̅̅̅̅ ̅̅̅ ̅̅̅
Taking Laplace transform
̅̅̅̅̅̅̅
̅̅̅̅

̅̅̅̅̅̅̅
̅̅̅̅

16
Prepared by: Soumen Panda
 Department of Chemical Engineering
BMS College of Engineering
Process Control Engineering

First order system in series:

Non - interacting:

Interacting:

Non - interacting:

Apply mass balance in 1st tank at any time t-

At steady state

In deviation form, above equation can be written as follow

17
Prepared by: Soumen Panda
 Department of Chemical Engineering
BMS College of Engineering
Process Control Engineering

Apply Laplace transform both side

̅̅̅
̅̅̅ ̅

̅̅̅ ̅

̅̅̅ ̅

Similarly, apply mass balance in 2nd tank at any time t-

Apply Laplace transform both side

̅̅̅ ̅̅̅
̅̅̅

̅̅̅ ̅

̅̅̅ ̅

̅̅̅ ̅

̅̅̅ ̅

̅̅̅ ̅

18
Prepared by: Soumen Panda
 Department of Chemical Engineering
BMS College of Engineering
Process Control Engineering

Where, √ √

Case 1:

̅̅̅ ̅

For a step change in input,

̅̅̅

Interacting Tank:

Apply mass balance in tank 1 and express equation in deviation form

Apply Laplace transform both side

( ̅̅̅ ) ̅̅̅ ̅ ̅̅̅̅

̅̅̅ ̅ ̅̅̅̅

̅ ̅̅̅̅
̅̅̅

Apply mass balance in tank 2 and express in terms of deviation form

Multiply both sides by

19
Prepared by: Soumen Panda
 Department of Chemical Engineering
BMS College of Engineering
Process Control Engineering

Apply Laplace transform both sides

[ ̅̅̅ ] ̅̅̅ ̅̅̅̅

Substituted ̅̅̅̅

̅ ̅̅̅
̅̅̅ ̅̅̅

Multiply both sides by

̅̅̅ ̅̅̅ ̅ ̅̅̅

̅̅̅ ̅̅̅ ̅̅̅ ̅ ̅̅̅

Divide both sides by

̅̅̅ ̅̅̅ ̅
̅̅̅ ̅

̅̅̅ ̅

̅̅̅ ̅

Where, √ √

20
Prepared by: Soumen Panda
 Department of Chemical Engineering
BMS College of Engineering
Process Control Engineering

Response of Non- interacting system with Step input

For same time constant and different time constant

Response of Interacting system with Step input

21
Prepared by: Soumen Panda
 Department of Chemical Engineering
BMS College of Engineering
Process Control Engineering

Review of Second order system:

Examples:

1. U tube manometer

2. Damped vibrator

22
Prepared by: Soumen Panda
 Department of Chemical Engineering
BMS College of Engineering
Process Control Engineering

Definition:

Any system that can be represented by a second order differential equation is a second order
system.

General form of second order system:

( ) ( ) ( )

0 ̅ 1 ̅ ̅ ̅

̅
̅

is damping factor

Response of Second order system:

Step response:

̅ ̅

If f’(t) = B

23
Prepared by: Soumen Panda
 Department of Chemical Engineering
BMS College of Engineering
Process Control Engineering

√ √
Where,

Case 1:

⁄ √ √
( )
√

Case 2:

⁄
⁄

Case 3:

√ √

Taking Laplace transform,

⁄ √ √
( )
√

Derivation of response of second order system with step input:

Case 1:
24
Prepared by: Soumen Panda
 Department of Chemical Engineering
BMS College of Engineering
Process Control Engineering

Let,

√ √

Therefore,

Note: Laplace transform of * +

( √ ) ( √ )

Taking Laplace transform

( √ ) ( √ )
( √ ) ( √ )
[ ]
( √ ) ( √ )

( √ ) ( √ )
( √ ) ( √ )
[ ]
√

(√ ) (√ ) (√ ) (√ )
( ) √ ( )
[ ]
√

, (√ ) √ (√ ) -
[ ]
√

0 2 (√ ) (√ ) 31
√

25
Prepared by: Soumen Panda
 Department of Chemical Engineering
BMS College of Engineering
Process Control Engineering

Note:

Case 2:

Case 3:

√ √

Let,

√ √

Therefore,

Note: Laplace transform of * +

( √ ) ( √ )

26
Prepared by: Soumen Panda
 Department of Chemical Engineering
BMS College of Engineering
Process Control Engineering

( √ ) ( √ ) √

Taking Laplace transform

( √ ) ( √ )
( √ ) ( √ )
[ ]
( √ ) ( √ )

(√ ) (√ ) (√ ) (√ )
( ) √ ( )
[ ]
√

, (√ ) √ (√ ) -
[ ]
√

0 2 (√ ) (√ ) 31
√

Note:

√
Here, √
√ √

Therefore,
√
[ { (√ ) }]
√

√ √
[ {( ) }]
√

√ √

27
Prepared by: Soumen Panda
 Department of Chemical Engineering
BMS College of Engineering
Process Control Engineering

Response of second order system with impulse and sinousoidal input:

See Book Coughnaowr Page 98, 99 and 100

28
Prepared by: Soumen Panda
 Department of Chemical Engineering
BMS College of Engineering
Process Control Engineering

Response analysis of second order system when system experienced step input (Under damped
system):

Overshoot: How much response exceeds the ultimate value ( )

. /
√

Decay Ratio:ratio of the two successive peaks ()

( ) . /
√

Rise time: Time required for the response to first reach its ultimate value.

29
Prepared by: Soumen Panda
 Department of Chemical Engineering
BMS College of Engineering
Process Control Engineering

Response analysis of second order system when system experienced step input:

30
Prepared by: Soumen Panda
 Department of Chemical Engineering
BMS College of Engineering
Process Control Engineering

Transportation Lag

31
Prepared by: Soumen Panda
 Department of Chemical Engineering
BMS College of Engineering
Process Control Engineering

Block Diagram:

X2
+
+ Y = X1 + X2 – X3
X1 Y
-

X3
X2
-
+ + Y Y = (X1 - X2)+ X3
X1
+
X3

X1 G Y Y = GX1

X1 G1 G2 Y Y = G1G2X1
G G

32
Prepared by: Soumen Panda
 Department of Chemical Engineering
BMS College of Engineering
Process Control Engineering

X1 G1
+
Y Y = G1X1 - G2X2

-
X2 G2

X + GF Gp
GC Y
- Y
M

GM

Develop Transfer function of the above block diagram:

d Gd
+ 33
Prepared by: Soumen Panda
X + GF Gp
GC Y
+
 Department of Chemical Engineering
BMS College of Engineering
Process Control Engineering

Develop transfer function for the above block diagram:

Determination of Y(s) is difficult in direct method, so we determine the Y(s) value once omitting
d(s) and then omitting X(s)

Omitting d(s): i.e. d(s)= 0

X + GF Gp
GC Y
- Y
M

GM

Develop Transfer function of the block diagram:

Omitting X(s): i.e. X(s)= 0

34
Prepared by: Soumen Panda
 Department of Chemical Engineering
BMS College of Engineering
Process Control Engineering

d(s) +
Gd GP Y
- Y
M

GF GC

Develop Transfer function of the block diagram:

When, X(s) and d(s) are not equal to zero,

When d(s) = 0 (No disturbance in the system)

This type of problem called regulatory problem Regulatory problem

When X(s) = 0 (No change in set point in the system)

35
Prepared by: Soumen Panda
 Department of Chemical Engineering
BMS College of Engineering
Process Control Engineering

This type of problem called servo problem Regulatory problem

Transient response:

Variation of system response with time before reaching steady state is called transient response.

Aim: To find variation of desired output of a system with time when system is experienced an
external disturbance or a change in set point.

d Gd
+
X + GF YF
Gp
GC Y
+
- Y
M
36
Prepared by: Soumen Panda GM
 Department of Chemical Engineering
BMS College of Engineering
Process Control Engineering

Response of above system in Laplace domain can be expressed as following

Response of the system can be found using following steps

Step 1.Knowing values of GC, GF, GP, GM, Gd, X and d we can find Y(s) servo or regulatory
problem.

Step 2. Take inverse Laplace of Y(s) to get Y(t), response in time domain. Find initial and final
value of response with the substitution of t 0 and t ∞, respectively in the expression of Y(t).

Alternate Step 2. Find initial and final value directly from Y(s) with the application of initial and
final value theorem.

Example:

GC = K; GM = 1; GF = 1; Gd = 1/A; GP = A/(𝞃s+1)

Case 1:

Let us assume X(s) 1/s (step response) and d = 0 (regulatory problem)

Therefore,

( )

* ( ) +

Final value:

Apply final value theorem

37
Prepared by: Soumen Panda
 Department of Chemical Engineering
BMS College of Engineering
Process Control Engineering

Offset = value of set point in time domain – Value of response Y(t∞)

Offset =

Offset is defined as the amount of deviation ofultimate response from the expected value. Here,
expected ultimate response of the system is 1 and ultimate response of the system is .

Initial value:

Apply initial value theorem

Case 2:

Let us assume d(s) 1/s (step response) and X(s) = 0 (Servo problem)

( )

* ( ) +

Final value:

Apply final value theorem

38
Prepared by: Soumen Panda
 Department of Chemical Engineering
BMS College of Engineering
Process Control Engineering

Offset = 0 – Value of response Y(t∞)

Offset =

Offset is defined as the amount of deviation of ultimate response from the expected value. Here,
expected ultimate response of the system is 0 and ultimate response of the system is .

Initial value:

Apply initial value theorem

Transient response of a real system:

Stirred tank heater:

39
Prepared by: Soumen Panda
 Department of Chemical Engineering
BMS College of Engineering
Process Control Engineering

Apply energy balance:

At any time t,

At steady state,

In deviation form

Where,

Apply Laplace transform in both side and substitute , we get

( )

Where,

We can write above expression as

Where,

Response in closed loop system:

Load TF
Ti +
TR + Controll FCE Q Process
er T
+
40
- T
Prepared by: Soumen
M Panda

GM
 Department of Chemical Engineering
BMS College of Engineering
Process Control Engineering

We can write the above system as following

Ti Gd
+
TR + GF Q
Gp
GC T
+
- Y
M

GM

Ultimate response can be found for servo and regulatory problem for different value of GC

( )

( )

Here,

Assume, GM = 1, GF = 1
41
Prepared by: Soumen Panda
 Department of Chemical Engineering
BMS College of Engineering
Process Control Engineering

Servo Problem:

No change in disturbance (Ti’= 0)

Response for Proportional – Integral controller (PI)

Therefore,

* ( )+ ( )

0* ( )+ ( ) 1

Final value:

Apply final value theorem

Offset = value of set point in time domain – Value of response Y(t∞)

Offset =

Offset is defined as the amount of deviation ofultimate response from the expected value. Here,
expected ultimate response of the system is 1 and ultimate response of the system is .

Regulatory problem:

No change in point of vessel temperature (T’R = 0)

Response for PI controller

( )

* ( ) ( ) +

42
Prepared by: Soumen Panda
 Department of Chemical Engineering
BMS College of Engineering
Process Control Engineering

Final value:

Apply final value theorem

Offset = 0 – Value of response Y(t∞)

Offset =

43
Prepared by: Soumen Panda
 Department of Chemical Engineering
BMS College of Engineering
Process Control Engineering

Controller and Final control element

Transducer: Takes Input as temperature and provide output as current

Controller recorder: Takes Input as current and provide output as current

Converter: Takes Input as current and provide output pressure

Control valve: Takes Input as air pressure and provide output as flow rate

Note: In the control system an external power of 120 V is needed to run each component like
transducer, controller and converter. A source of 20 psig is needed for the converter.

44
Prepared by: Soumen Panda
 Department of Chemical Engineering
BMS College of Engineering
Process Control Engineering

Controller: Controller may thought of as a device that receives the error signal and provide
corrective instruction to the final control element as output of it.

Final Control Element: The final control element may be regarded as a device that produces
corrective action on the process. The corrective action is regarded as mathematically related to
the output signal from the controller.

Types of Controller:

1. Proportional Controller
2. Proportional – Integral Controller
3. Proportional – Integral – Derivative Controller

Proportional Controller: The proportional controller produces an output signal that is

proportional to the error (ϵ).

This action may be expressed as

P = Output signal from controller, psig or mA

K = Gain or sensitivity

ϵ = error = set point – measured variable

PS = Output signal of the controller when error is zero

PS can be adjusted to obtain the required output signal when the control system is at its steady
state and error is zero.

At steady state P = PS

Therefore,

Taking Laplace transform

̅ ̅

̅
̅

In proportional controller instead of Gain proportional Band is used

45
Prepared by: Soumen Panda
 Department of Chemical Engineering
BMS College of Engineering
Process Control Engineering

Proportional Band (Pb): The error required to move valve from fully close to fully open. This is
expressed in percentage.

Let a controller can control range of a parameter is R, but the parameter can vary only P unit
when the valve move from fully close to fully open.

Therefore,

Example:

A pneumatic proportional controller is used to control temperature within the range of 60 to 100
0
C. The controller is adjusted so that the output pressure goes from 3 psi to 15 psi (Valve move
fully close to fully open) as the measured temperature goes from 71 to 75 0C with the set point
held constant. Find the Gain and Proportional Band.

Now assume that the proportional band of the controller is changed to 75%. Find gain and
temperature change necessary to cause a valve to go fully open to close.

As Pb is 75%, so temperature can vary up to (100-60)×0.75 = 30 0C when valve go fully open to

fully close.

Therefore,

ON – OFF Controller: This is a special type of proportional controller.

If Gain K of the proportional controller is very high, the valve will move from one extreme
position to other when controller experienced very small error. This very sensitive action is
called On – OFF action, because the valve is either open or close i.e valve acts like a switch.

The band width of an ON – OFF controller is approximately zero.

46
Prepared by: Soumen Panda
 Department of Chemical Engineering
BMS College of Engineering
Process Control Engineering

Proportional – Integral control:

This mode of controller is described by the relationship

Where,

To visualize the response of this controller, consider the response to a unit step change in error.

Therefore,

P changes suddenly by K amount and then changes linearly with time at a rate of

Apply Laplace transform to get Transfer function

̅
( )
̅

47
Prepared by: Soumen Panda
 Department of Chemical Engineering
BMS College of Engineering
Process Control Engineering

Proportional – Derivative Controller:

This mode of controller is described by the relationship

Where,

To visualize the response of this controller, consider the response top-=a linear change in error.

Therefore,

The effect of derivative action in this case is to anticipate the linear change in error by adding
additional output to the proportional action.

Apply Laplace transform to get Transfer function

̅
̅

Proportional – Integral – Derivative Controller:

This mode of controller is described by the relationship

48
Prepared by: Soumen Panda
 Department of Chemical Engineering
BMS College of Engineering
Process Control Engineering

Similarly we get Transfer function

̅
( )
̅

Motivation of addition of Integral and Derivative control mode:

Use of only proportional controller we get non zero offset that is not expected. PI controller has
the ability to make offset zero. Though PI controller provides zero offset, it produces an
oscillatory behavior.

Use of PID controller arrests rise of control variable more quickly and return rapidly to the
original value with little or no oscillation.

Note: If some offset is tolerable proportional action would likely be selected. If no offset were
tolerable, integral action would be added. If excessive oscillation needs to be eliminated,
derivative action might me added.

49
Prepared by: Soumen Panda