The University Of Nottingham

Malaysian Campus

Department of Chemical and Environmental

Engineering

Laboratory Report
Experiment number: CTR2
Title: Flow control
Group number: 6
Group members: Lai Kar Chiew (013319)
Lim Wei Xuen (013312)
Azman Arrif Imran (013303)
El Shanawany Hosam Samy (013396)
Name of tutor: Dr. Ong Sze Pheng
Date of experiment: 26th March 2015
Date due: 28th April 2015

Summary
The aim of this report is to explain the conceptual basis of fluid control and to discuss the
observed trends which occurred during the procedures carried out to produce the final
data. In CTR2, a flow controller (G.U.N.T RT020) is used to test the means of flow
control. The RT020 with the assistance of an adjustable pump asserts flowing fluid from
a storage tank into a piping loop. The flow can be manipulated by changing the charge at
which the electromagnetic proportional valve is set to. The integration of a rotameter
and turbine wheel flow sensor helps compute the data through a software controller.
A few tasks were carried out to determine open-loop behaviour, limitations of process,
influence of disturbance to process, controller settings by using Ziegler-Nicholas closedloop method and to evaluate the performance of controllers.
Reliability of the overall process can be questioned when considering that external
disturbances could influence the manner in which the data is processed through the
software. When scrutinizing flow, it is important to consider factors of the
experimentation that could affect results.

Results

Figure 1: CTR2-1

Figure 2: CTR2-2

Figure 3: CTR2-3

Figure 4: CTR2-4

Figure 5: CTR2-5

Figure 6: CTR2-6

Figure 7: CTR2-7

Figure 8: CTR2-8

Figure 9: CTR2-9

Figure 10: CTR2-10

Discussion
Control systems are used to keep the process operating at desired conditions by
manipulating certain process variables to adjust the variables of interest. A number of
benefits are offered by automatic control of a process: enhanced process safety, meet
product quality specification, efficient process and improved profitability. In this flow
control experiment, the variable of interest (controlled variable) is flowrate, X while the
manipulated variable is the setting of the electromagnetic proportional valve and the
regulation ratio, Y is the percentage of opening of the electromagnetic proportional
valve.
Task 1: To Determine Open-Loop Behaviour and Limitation of Process
Graph CTR 2-1 shows that the flowrate is constant at a regulation ratio of 20%. When
the regulation ratio was increased from 20% to 30% the flowrate is higher then it
becomes constant throughout the similar regulation ratio. The constant flowrate is an
indication of steady-state. Besides, it can also be seen that when Y changed, the
flowrate, X does not change immediately, instead it changes after sometime. Besides,
only first order system can react immediately to the excitation as others dont. Hence,
we can say that this is a first order process plus dead time (FOPDT).
From graph CTR2-2, it can be observed that when Y changes, X changes accordingly.
However, the increment in X is getting smaller and smaller for every constant increase in
Y. For instance, when Y changes from 30% to 40%, X changes from 72 Lhr-1 to 95 Lhr-1,
with an increment of 23 Lhr-1; when Y changes from 60% to 70%, X changes from 138
Lhr-1 to 145 Lhr-1, with an increment of 7 Lhr-1. Moreover, when Y reaches 100%, in
which the valve is fully opened, the maximum flowrate, Xmax is 160 Lhr-1. The limitations
of the process are the maximum achievable flowrate of the fluid as well as the limit of
the pump performance and the maximum opening of the electromagnetic proportional
valve.
Task 2: To Determine Influence of Disturbance to Process
In graph CTR2-3, the thinner blue line represents the disturbance signal, Z. From graph
CTR2-3, several observations were made. Firstly, when Z changes from 0% to 20%, X
doesnt change, remains constant at 130 Lhr-1. This indicates that the disturbance signal,
Z at this stage is not strong enough to implicate the process. Secondly, when Z reaches
20% and onwards, X is decreasing. This is because it is an open-loop system, there isnt
any controller to bring the process back to the set point. Thirdly, the decrement in X is
getting bigger as Z is stronger. For example, when Z changes from 90% to 100%, X
drops from 16 Lhr-1 to 6 Lhr-1; when Z changes from 20% to 30%, X drops from 65 Lhr-1

to 62 Lhr-1. The possible disturbances that could occur in a flow process are as followed:
presence of bubbles, leakage from the transportation line, pump speed, loose valve
fitting and valve failure. Pump speed will probably vary from time to time which would
certainly affect the fluid flowrate.
The signal block diagram for the open-loop system is shown below:
Disturbances:
Bubbles, pump speed

Manipulated variable:
Setting of electromagnetic proportional valve

Flow Process

Desired flowrate, X

Figure 11: Signal Block Diagram for Open-loop System.

Task 3: To Determine Controller Settings by Using Ziegler-Nichols Closed-Loop

Method
From graph CTR2-4, a sustained oscillation in response (X) is observed at Kp = 0.6. And
this Kp value is recorded as Kcrit. Furthermore, the period of sustained oscillation, Tp =
0.92s is also obtained from the graph CTR2-4. The value of the sustained oscillation, T p
is determined by measuring the time interval between two peaks of the graph. The
parameters of P, PI and PID controllers are shown in Table 1.
Table 1: Ziegler-Nichols Closed-Loop Optimum Controller Settings.

Type of controller

Optimum Settings
Proportional Gain

Integral Action Time

Derivative Time

0.300

P+I

0.273

0.767

P+I+D

0.353

0.460

0.115

Task 4: To Evaluate Performance of P, PI, and PID Controllers

From graph CTR2-5 to graph CTR2-10, the yellow line represents the set-point, W.
Graph CTR2-5 and graph 2-6 are associated with a P controller. Graph CTR 2-5 shows
the performance of the P controller with no disturbance. From graph CTR2-5, it can be
seen that the flowrate, X is so far away from the set-point, W, results in an off-set of
approximately 30 Lhr-1. Graph CTR 2-6 shows the performance of the P controller in the
presence of disturbance. In graph CTR2-6, again an off-set of approximately 30 Lhr -1 is
observed before introducing the disturbance. After introducing the disturbance, X first
drops from 30 Lhr-1 to 0 and then is recovered back to 12 Lhr-1 by the P controller, results
in an off-set of 48 Lhr-1.
Graph CTR2-7 and graph CTR2-8 are associated with a PI controller. Graph CTR 2-7
shows the performance of the PI controller with no disturbance. From graph CTR2-7, the
flowrate, X overlaps with the set-point, W, no off-set is observed in this case. However,
the process response becomes oscillatory. Graph CTR 2-8 shows the performance of the
PI controller in the presence of disturbance. In graph CTR2-8, no off-set is observed
before introducing the disturbance. After introducing the disturbance, X first drops from
60 Lhr-1 to 30 Lhr-1 and is recovered back to the set-point in 8 seconds by the PI
controller.
The last two graphs, graph CRT2-9 and graph CTR2-10 are associated with PID
controller. Graph CTR 2-9 shows the performance of the PID controller with no
disturbance. In graph CTR2-9, no off-set is observed and the Y curve is subjected to
abrupt changes. This is because the measured output is noisy. From graph CTR2-10, no
off-set is observed until the disturbance is introduced. After introducing the disturbance,
X first drops from 60 Lhr-1 to 27 Lhr-1 and is recovered back to the set-point in 4 seconds
by the PID controller.
After comparing the last 6 graphs (graph CTR2-5~graph CTR2-10), P controller is the
least suitable controller for flow control as an off-set of 48 Lhr-1 is observed. For PI and
PID controllers, both can recover the process back to set-point even though disturbance
is present. PID controller responds faster than PI controller because it takes into account
of now, present and future error while PI controller considers only now and present error.
However, the PID controller is seldom used for flow control because the flow
measurement tends to be very noisy and flow process is quite fast. From graph CTR2-9
and graph CTR2-10, it can be seen that X and Y curves are more oscillatory because the
measured output is noisy. In conclusion, the PI controller is the most suitable controller
for flow control as it can recover the process back to the set-point in short time and it is
a lot cheaper than that of PID controller.

The signal block diagram for closed-loop system is shown below:

Disturbances:
Bubbles, pump speed

Set-point,
W

Set-point
error, e

Controlle
r

Secondary element:
Transmitter

Electromagnetic
proportional
valve

Flow Process

Primary element:
Turbine wheel flow sensor

Figure 12: Signal Block Diagram of Closed-loop Diagram.

Desired
Flowrate,
X

Conclusion
The mechanism used in CRT2 provides efficient means of controlling flow. And again,
flow rate is highly dependent on the regulation ratio to which the electromagnetic
proportional valve is set to. The regulation ratio increases as set points and disturbances
are accounted for statistically.
For task 1, the process is a first order process plus dead time (FOPDT). Furthermore, the
limitations of the process are as followed: maximum achievable flowrate of the fluid,
limit of the pump performance and the maximum opening of the electromagnetic
proportional valve.
For task 2, the possible disturbances are presence of bubbles in fluid, leakage from the
transportation line, pump speed, loose valve fitting and valve failure.
For task 3 and 4, the performance of P, PI and PID controllers are evaluated. We
concluded that PI is the most suitable controller for flow control as it can recover the
process back to the set-point in short time and it is cheap.

Notations
P Proportional
PI Proportional-Integral
PID Proportional-Integral-Derivative
W Set point, Lhr-1
X Flowrate, Lhr-1
Y Regulation ratio
Z Disturbance singal
Kp Proportional gain
Kcrit Critical proportional gain
Tp Period of sustained oscillation, s

References
Coughanowr, D. R. & E.LeBlanc, S., 2009. Process systems analysis and control. 3rd ed.
New York: McGraw-Hill.

Appendix
Given Kp = Kcrit = 0.6, Tp = 0.92s (obtained from graph CTR2-4)
The parameters of P, PI and PID controllers are calculated using the table below.
Table 2: Ziegler-Nichols Closed-Loop Optimum Controller Tuning Formulae.
Type of controller

Optimum Settings
Proportional Gain

Integral Action Time

Derivative Time

Kcrit/2

P+I

Kcrit/2.2

Tp/1.2

P+I+D

Kcrit/1.7

Tp/2

Tp/8

For P controller:
Kp = Kcrit/2 = 0.6/2 = 0.300
For PI controller:
Kp = Kcrit/2.2 = 0.6/2.2 = 0.273
Integral Action Time = Tp/1.2 = 0.92/1.2 = 0.767s
For PID controller:
Kp = Kcrit/1.7 = 0.6/1.7 = 0.353
Integral Action Time = Tp/2 = 0.92/2 = 0.460s
Derivative Time = Tp/8 = 0.92/8 = 0.115s