School of Chemical Engineering

The University of New South Wales

CEIC3006
Process Dynamics and Control

Week 3
Lecturer: Prof. Jie Bao
 Development of Transfer functions

 Steps to develop transfer functions

1. Define deviation variables:

2. Take Laplace transform on the time domain ODE

3. Rearrange the equation such that Y(s) can be represented
as a function of U(s)
 Important parameters for first order systems
 Steady state gain
 Time constant

CEIC3006 Term 2, 2022

 Transfer functions
To determine y (t) :
With U (s ), the Laplace transform of u (t), and G(s ), the transfer
function, you can work out the output in the Laplace domain:

The time domain output y (t):

What is the catch?

The solution is only valid with the zero initial value y (0)=0!

CEIC3006 Term 2, 2022

 Transfer function
 What is a transfer function – a quick recap
 An algebraic expression for the dynamic relationship between input
and output of a process model.
Dynamic model

u(t) ODE y(t)

U (s ) G(s ) Y (s )

CEIC3006 Term 2, 2022

Recap: Development of Transfer functions

CEIC3006 Term 2, 2022

 To use a transfer function to determine the process output
To determine y (t) :
1. To calculate

2. Get the Laplace transform

3. Calculate the response:

4. Find the inverse Laplace transform:

5. Calculate the actual process output:

CEIC3006 Term 2, 2022

 Example
At the nominal steady state, the inlet stream temperature is Ti=70°F
and the heater input is Q=1920Btu/min. Additional information:

200

(1) If at t = 0, inlet temperature Ti is changed from 70 to 90°F,

calculate the exit stream temperature response, T(t).

From the steady-state model:

T(0)= 100 °𝐹𝐹
The input step changes:

CEIC3006 Term 2, 2022

 Example
At the nominal steady state, the inlet stream temperature is Ti=70°F
and the heater input is Q=1920Btu/min. Additional information:

200

(1) If at t = 0, inlet temperature Ti is changed from 70 to 90°F,

calculate the exit stream temperature response, T(t).

From the steady-state model:

𝑇𝑇� = 100 °𝐹𝐹
The input step changes:

CEIC3006 Term 2, 2022

 Example

At the nominal steady state, the inlet stream temperature is Ti=70°F

and the heater input is Q=1920Btu/min. Additional information:
200

(2) If at t = 0, the heater input Q is changed from 1920 to 1600 Btu/min,

calculate the exit stream temperature response, T(t).

From the steady-state model: 𝑇𝑇� = 100 °𝐹𝐹

The input step changes:

CEIC3006 Term 2, 2022

 Example

At the nominal steady state, the inlet stream temperature is Ti=70°F

and the heater input is Q=1920Btu/min. Additional information:
200

If at t = 0,
(1) inlet temperature Ti is changed from 70 to 90 °F, and
(2) the heater input Q is changed from 1920 to 1600 Btu/min,
calculate the exit stream temperature response, T(t).
𝑇𝑇(0) = 100 °𝐹𝐹
The input step changes:

CEIC3006 Term 2, 2022

 Transfer functions
 Advantages
 Transfer function representation can be used to analyze the dynamic
feature without solving the model
 Very convenient in model-based control system design

 Typical linear (or linearized) processes

 First order systems
 Time delay systems
 Integrating systems
 Second order systems

CEIC3006 Term 2, 2022

 First order systems

We have seen them!

CEIC3006 Term 2, 2022

 First order Processes
 Examples:
Fi
A Surge Tank (linearized version)
h

CEIC3006 Term 2, 2022

 First Order Processes
 Speed of a Car u = traction force
v = speed

 Stirred-tank heater

Note:

CEIC3006 Term 2, 2022

 First order processes

All first order processes produce the same type of

responses to one input signal (as a function of time)
regardless of what physical systems they are and the
fundamental mechanisms.

Liquid level Speed Temperature

CEIC3006 Term 2, 2022
 First order processes
 What we look for? (Again!)
 Process Gain: Steady-State Response

 Process Time Constant:

τ = Time required to reach 63.2% of final value

 What do we need to identify process gain and time constant?

 Process at steady-state
 Step input of magnitude M
 Measure process gain from new steady-state
 Measure time constant

CEIC3006 Term 2, 2022

 First order processes
 Step change response

0.9

0.8
t y(t)
0.7 0.632 0 0
τ
0.6
63%
0.5

0.4
2τ 86%
0.3
3τ 95%
0.2 4τ 98%
0.1 5τ 99%
0
0 τ1 τ
2 3 τ 4τ 5τ τ
6

CEIC3006 Term 2, 2022

 First order processes
 Ramp response:

Using partial fraction expansion:

5
Input

0
-1 0 1 2 3 4 5 6 7 8 9
Time

CEIC3006 Term 2, 2022

 First order processes
 Ramp response:

Ramp input of slope a

4.5

3.5

2.5

1.5

0.5

0
0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5

CEIC3006 Term 2, 2022

 First order Processes
 Sinusoidal response

CEIC3006 Term 2, 2022

 First Order Processes
 Steady-state response

1.5

0.5
AR
0

-0.5

-1 φ
-1.5
0 2 4 6 8 10 12 14 16 18 20
t/τ

The steady-state responses of first order system to

sinusoidal input is another sinusoidal signal with the same
frequency.
CEIC3006 Term 2, 2022
 First order Processes
 Frequency response

Amplitude Ratio Phase Shift

CEIC3006 Term 2, 2022

Thermocouple example

CEIC3006 Term 2, 2022

Thermocouple example

CEIC3006 Term 2, 2022

 Thermocouple Example
 Computer simulation
 Using Simulink

CEIC3006 Term 2, 2022

 Properties of Transfer Functions

CEIC3006 Term 2, 2022

 Converting an ODE to transfer function

Laplace Transform

CEIC3006 Term 2, 2022

 Converting an ODE to transfer function

Laplace Transform

CEIC3006 Term 2, 2022

 Converting an ODE to transfer function

Why?

CEIC3006 Term 2, 2022

 Properties of Transfer Functions

R1

U1(s) U2(s) Y(s)

G1(s) G2(s)

R2

Two surge tanks in series

CEIC3006 Term 2, 2022

 Time delay systems

CEIC3006 Term 2, 2022

 Time Delays

M
Tin, w

TT1 TT2

Q
T, w

Time delays are commonly occur in process control because of

the transport time required between process units.

If: length of pipe = 10 m, velocity of the fluid = 0.5m/second.

There is a time-delay (or called dead time) of θ=20 seconds.

CEIC3006 Term 2, 2022

 Time Delays

Readings from Sensors 1 and 2

55

54

Sensor 1 (Deg)
53

52

51

50
0 10 20 30 40 50 60

55

54
Sensor 2 (Deg)

53

52

51

50
0 20 40 60 80 100 120
Time (sec)

CEIC3006 Term 2, 2022

 Time Delays
fd = 0 for t < θ. fd and f are related by:
Dead time

Delayed function Unit step function

 Laplace transform of a time-delayed function:

CEIC3006 Term 2, 2022

 Time Delays
 (t − θ) is the artificial variable of integration, it can be replaced by t*

 yielding the Real Translation Theorem.

CEIC3006 Term 2, 2022

 Time Delays

CEIC3006 Term 2, 2022

 Integrating Processes

CEIC3006 Term 2, 2022

 Integrating Processes
Example: Liquid storage tank Fi

If F is constant,

If Fi is constant,

Integrating Processes
CEIC3006 Term 2, 2022
 Integrating Processes
 Transfer functions of integrating systems
Fi

CEIC3006 Term 2, 2022

 Process Modeling
 Step input of magnitude M

Slope = KM
unbounded

Time Time
An integrating system does not reach a new steady-state when
subjected to step changes (if there is no constraints).
Non-self-regulating systems
CEIC3006 Term 2, 2022
 Integrating Processes

 Rectangular pulse response

Time Time

CEIC3006 Term 2, 2022

 Integrating Processes

 Unit impulse response

Time Time

CEIC3006 Term 2, 2022

 Example
 A vented cylindrical tank is used for storage between a tank car unloading facility
and a continuous reactor that uses the tank car contents as feedstock.
 Reactor feed flow rate: 0.02 m3/s.
 Tank height: 5 m
 Cross-sectional area: 4 m2

CEIC3006 Term 2, 2022

 Example
 Suppose that after a long period of operation, the storage tank level is 2 m at
the time the tank car empties. How long can the reactor be operated before
the feed tank is depleted?
 For this system, there is no unique steady-state level corresponding to a
particular value of input and output flow rate.
 Suppose the initial level is h = 2 m and the constant flow rate from the
feed pump to the reactor, q = 0.02 m3/s, is the basis for defining
deviation variables for h, q, and qi.

CEIC3006 Term 2, 2022

 Example
 The process model for the tank:

 At the time the tank car empties

 Inversion to the time domain gives h′(t) =

−0.005t and h(t) = 2 − 0.005t. The length of
time for h(t) to go to zero is t = 2/0.005 = 400 s.
CEIC3006 Term 2, 2022
 Example
 Another tank car is moved into place and connected
to the tank, while flow continues into the reactor at
0.02 m3/s. If flow is introduced into the tank just as
the tank level reaches 1 m, how long can the transfer
pump from the tank car be operated? Assume that it
pumps at a constant rate of 0.1 m3/s when switched
on.
 For the tank-filling period and using the same basis
for deviation variables

CEIC3006 Term 2, 2022

 Example

 Inversion to the time domain yields h(t) = 1 + 0.02t.

Thus, the transfer pump can be operated for 200 s
until h(t) = 5 m, when the tank would overflow.
 Note that this time (as well as the time to empty the
tank in (a)) can be calculated without using Laplace
transforms, simply by using the constant rate of
inflow (or outflow) and the tank volume.

CEIC3006 Term 2, 2022

 Second Order Processes

CEIC3006 Term 2, 2022

 Second Order Processes
Three types of second order process:
1. Multi-capacity processes (physical)
Two first-order processes are connected in series.
Two surge tanks in series

2. Inherently second order processes

Fluid or solid mechanic processes possessing inertia and subjected to
some acceleration
e.g., A pneumatic valve

3. Processing system with a controller

Presence of a controller induces oscillatory behavior
e.g., Feedback control system
CEIC3006 Term 2, 2022
 Second Order Processes
 Multi-capacity Second Order Processes
 Naturally arise from two first order processes in series
e.g. two tanks connected in series:

U(s) Outlet Y(s)

Inlet flowrate Outlet
flowrate flowrate
of tank 1 of tank 2

 By multiplicative property of transfer functions

U(s) Y(s)

CEIC3006 Term 2, 2022

 Second Order Processes
 Multi-capacity Second Order Processes
R1

R2

CEIC3006 Term 2, 2022

 Second Order Processes

 Inherently second order processes

 Mass on a spring
Hookes Law - Spring Force

Responses very different to that of a 1st order process!

What if there is frictional force?
CEIC3006 Term 2, 2022
 Second Order Processes

 Damping coefficient
 Mass on spring with mechanical damping (if there is frictional force):

CEIC3006 Term 2, 2022

 Second Order Processes

 Second order process:

The general form

CEIC3006 Term 2, 2022

 Step Responses of Second order Processes

 Three families of processes

CEIC3006 Term 2, 2022

 Step Responses of Second order Processes

 Three families of processes

CEIC3006 Term 2, 2022

 Step Responses of Second Order Processes

 Three families of processes

CEIC3006 Term 2, 2022

 Step Responses of Second Order Processes

 Three families of processes

CEIC3006 Term 2, 2022

 Second order processes

 Step response

2
1.8
ζ=0
1.6 ζ=0.2
1.4
y/KM

1.2 ζ=1
1
0.8
0.6
ζ=2
0.4
0.2
0
0 1 2 3 4 5 6 7 8 9 10
t/τ
CEIC3006 Term 2, 2022
 Second order process
Observations
 Damping ratio ζ indicates the amount of damping in the system.
 That is the degree of oscillation in the process response after a perturbation.

 Small value of ζ implies little damping.

 As ζ (with 0< ζ <1) becomes smaller the
response becomes more oscillatory
 Responses exhibit overshoot (y(t)/KM >1)
when ζ <1 ζ damping

 If ζ<0, the system oscillates without bounds

(unstable)

CEIC3006 Term 2, 2022

 Second order process
Observations
 Large ζ yields a sluggish (slow) response
 Systems with ζ=1 yield the fastest response without overshoot

CEIC3006 Term 2, 2022

 Second order process

 Damping coefficient
 Double tank:

CEIC3006 Term 2, 2022

 Second order process

 Damping coefficient
 Mass on spring:  Mass on spring with damping:

CEIC3006 Term 2, 2022

 Second order Processes
1.8

1.6

1.4

1.2 +5%
1
-5%
0.8

0.6

0.4

0.2

0
0
tr tp5 10 15 20
ts
25 30 35 40 45 50

Rise time: tr is the time the process output takes to first reach the new steady-state value.
Time to First Peak: tp is the time required for the output to reach its first maximum value.
Settling time: ts is defined as the time required for the process output to reach and remain
inside a band whose width is equal to ±5% of the total change in y.

CEIC3006 Term 2, 2022

 Second order Processes
1.8
P
1.6

1.4
a
1.2 +5%
c
1
-5%
0.8

0.6
b
0.4

0.2

0
0 5 10 15 20 25 30 35 40 45 50

Overshoot: OS=a/b
Decay ratio: DR=c/a
Period of oscillation: P is the time between two successive peaks or two successive valleys of the response.

CEIC3006 Term 2, 2022

 Second order Processes
 Characteristics of underdamped second order process

 Time to first peak, tp

 Overshoot:

 Decay ratio:
ζ
 Period

CEIC3006 Term 2, 2022

 Example of second order systems
 A stirred-tank heater with a control system
 The overall system from the feed flow w to temperature T in the tank is
second order underdamped system
TT

Tin, w
Thermocouple

Heater
Q T, w

TC

(a) Sudden change of Q from 4kW to 5 kW causes a dynamic change of T from a steady-state of 100C to
eventually 102C. What is the steady state gain of the overall system?

CEIC3006 Term 2, 2022

 Example of second order systems
(b) The operator notes that the resulting response is slightly oscillatory with maxima of 102.5
and 102.03 C occurring at times 1000s and 3060s after the change is initiated. What is the
complete process transfer function?

CEIC3006 Term 2, 2022

 Example of second order systems
(b) The operator notes that the resulting response is slightly oscillatory with maxima of
102.5 and 102.03 C occurring at times 1000s and 3060s after the change is initiated.
What is the complete process transfer function?

CEIC3006 Term 2, 2022

 Sinusoidal Response

For large values of t:

where:

The amplitude ratio:

CEIC3006 Term 2, 2022

 Example
 An engineer uses a temperature sensor mounted in a thermowell to measure
the temperature in a CSTR.
 The temperature sensor/transmitter combination operates approximately as a
first-order system with a time constant of 3 s.
 The thermowell behaves like a first-order system with a time constant of 10 s.
 The engineer notes that the measured reactor temperature has been cycling
approximately sinusoidally between 180 °C and 183 °C with a period of 30 s
for at least several minutes.
 What can be concluded concerning the actual temperature in the reactor?

CEIC3006 Term 2, 2022

 Example
 First, note that the sensor/transmitter and the transmission line act as
two first-order processes in series with an overall gain K = 1, and with
the approximate transfer function:

 Some disturbance has caused the actual reactor temperature to vary

sinusoidally.
 The cycling period is much longer than the time constants of the process
so the transients have died out.

CEIC3006 Term 2, 2022

 Example
 The frequency of the perturbing sinusoidal signal (reactor temperature) is
calculated from the observed period of 30 s:

 The amplitude of the output perturbation also is obtained from observed

results as

CEIC3006 Term 2, 2022

 Example
 The amplitude of the actual reactor temperature

A = 4.12 °C
 The reactor temperature is varying between 181.5 − 4.12 = 177.38 °C
and 181.5 + 4.12 = 185.62 °C, nearly three times the variation indicated
by the measured value.
 Overdamped system (ζ = 1.19) results sinusoidal perturbations in the
reactor temperature always will be attenuated in the measurement system
regardless of the frequency of the perturbation.

CEIC3006 Term 2, 2022