You are on page 1of 11

CDB 3062 Chemical Engineering Laboratory III

September 2016 Semester

Experiment 5: Ratio Control

Group :8

Group Members : Assyamril bin Dzulkurnain 20110

Baldip Kaur A/P Ajit Singh 19907

Lim Chien Chiang 19789

Nur Hanif Bin Awang Da 20365

Siti Zulaikha Binti Mohd Affandi 19917

Lecturer : Dr Nasser Mohamed Ramli

Lab Instructor : Mr Muhammad Yasir Shamim

Date of Experiment : 27th September 2016

Date of Submission : 4th October 2016

1
Contents
1. Summary............................................................................................................................................ 3
2. Introduction & Theory ....................................................................................................................... 3
3. Results & Discussion .......................................................................................................................... 4
4. Conclusion ....................................................................................................................................... 10
5. References ....................................................................................................................................... 10
6. Appendixes ...................................................................................................................................... 11

2
1. Summary
This experiment was conducted to demonstrate the characteristics of Proportional Band, Integral
Action and Derivative Action on a flow process control loop. Also, the experiment was to allow us to
understand better about the characteristic of ratio control.

In the first experiment, we would like to know the changes in the flow process of control loop
when the Derivative, D value is manipulated by 1, 5 and 8 s while the Proportional Band, PB and
Integral, I value remained fixed at PB = 200 and I = 6s while a disturbance is introduced for 3s for both
flow rates of 50 and 75 LPM.

In the second experiment, we would like to investigate the changes in the flow process of
control loop when the Proportional Band, PB is adjusted. The values that we had selected was 1000,
800, 600, 50, 25, 20 and 12.5.

For the third experiment, we wanted to test for ratio control. The pre-set ratio was 2:1.
Therefore, should be getting a graph showing that for every gallon of the uncontrolled variable
flowing, two gallons of the controlled variable is allowed to flow. We tried several settings in order to
get this desired ratio.

In conclusion, all of the objectives of the experiment were obtained. We were able to
determine the characteristics of Proportional Band, Integral Action, Derivative Action and also Ratio
Control.

2. Introduction & Theory


The experimental equipment is designed to show the relationship between ratio control and a single
loop flow control. In this experiment, we shall investigate the effect of Proportional Band (PB),
Integral (I) Action and Derivative (D) Action on flow process loop control. We also would like to find
the best controller settings to obtain the desired ratio.

Proportional Band (PB), also known as throttling range, in definition is how much changes in
the controlled variable is needed to drive the loop output from minimum percent to maximum
percent. Any system that has changes in set point will need a wider PB in order to obtain stability in
control during the system upsets. Hypothetically, the larger the PB the slower the loop response. This
is because decreasing the PB will increase the amount of overshoot.

3
Integral (I) action corrects the error caused by the offset by continuing to increase or decrease
the output as long as the offset still exists. The error is determined by the following formula:

Derivative (D) action works by adding a portion of the error to the calculated output to
increase the output response. This, however, will only happen if the changes in error are more than a
specific percentage during a specific period of time. D action will counter the PI control action. Thus,
resulting in decrease of overshoot.

3. Results & Discussion


Experiment 11.5.1 PID Flow Control

The purpose in this part of the experiment is to observe the changes in the flow process of control
loop when the Derivative, D value is manipulated by 1, 5 and 8 s while the Proportional Band, PB and
Integral, I value remained fixed at PB = 200 and I = 6s while a disturbance is introduced for 3s for both
flow rates of 50 and 75 LPM.

DISTURBANCE
DISTURBANCE

DISTURBANCE

75 LPM 50 LPM

Figure 1.0

For D = 1s in Figure 1.0, the disturbance changes were reduced to 24% of the set point value
of 50 LPM while for 75 LPM, the disturbance decreases to 33.3% of its initial set point changes.
Furthermore, the disturbance peak for 75 LPM is found to be higher than the disturbance peak for 50

LPM and consequently the time taken to reach the initial set point is longer of about time unit for a

higher flow rate of 75 LPM while it takes 1 time unit for a smaller flow rate of 50 LPM.

4
DISTURBANCE

DISTURBANCE

75 LPM 50 LPM

Figure 1.1

For D = 5s in Figure 1.1, the result shows a similar decrease in percentage reduction as D = 1s
for 50 LPM. However, for 75 LPM its initial value reduced slightly higher approximately 36%.
Moreover, it also demonstrate an identical trend as D = 1s in terms of the difference in disturbance
peak height and the time taken to reach its initial value when the set point was changed from 50 LPM.
Despite the similarities, the time taken for 75 LPM was faster with 0.9 time unit.

DISTURBANCE

DISTURBANCE

75 LPM 50 LPM

Figure 1.2

For D = 8s in Figure 1.2, there is a fluctuation with regard to the reduce changes once a
disturbance was introduce. Although it was analysed initially to reduced similarly to the response in
D=1s, it decreases further forming an inverted double peak for 50 and 75 LPM. Additionally, for both
flow rates, the disturbance peak shows the same height while the time taken is longer for 75 LPM.

Overall the results show a variation in terms of the percentage reduction of set point flow
rates due to the ineffectiveness in the introduction of disturbances by the closing of the HV537 valve
for 3s. The responses also illustrates that a lower derivative value is more stable compared to a higher
5
derivative value as theoretically small values in derivative tends to lower the maximum deviation from
the set point value and the rate of oscillation. Higher value of derivative is not favourable as it can
amplify the measurement noise, the random error in Process Variable, PV signal, resulting in an
oscillatory response which can be observed from the response when D = 8s. Therefore, an
intermediate value of D = 5s is desirable as it shows a more stable and faster response in reverting the
flow rates to the initial set point.

Experiment 11.5.2 PID Flow Control Loop Tuning

PB = 600 % PB = 800 % PB = 1000 %

Figure 2.0

In this experiment, the time integral value and the time derivative value is maintained at I = 1000s and
D = 0s. Initially, we input the value for the proportional band (PB) at PB = 1000. Then, we measure the
response of the system. Next, the experiment is repeated at different values of PB which are 800, 600,
25, and 12.5.

For PB = 800 and PB = 600, we are unable to determine the natural period because the system
doesnt oscillate even after we introduce the disturbance as shown in Figure 2.0.

PB = 12.5 % PB = 25 %

2 mm
Figure 2.1

6
As shown in Figure 2.1, only when the PB is set at PB = 25, a response that continuously
oscillate around the set point is obtained. We also experimented with PB value of 12.5. However, we
would use PB = 25 because it produces a more stable response compared to PB = 12.5.

Then, from PB = 25, we measure the natural period of the oscillation. The period length is
measured to be 2mm. The rate of the recorder is set to be 1200mm/hr. Thus, we are able to calculate
the natural period in term of time.

2mm 3600s
Natural Period (P) = x 6s
1200mm/hr 1hr

Thus the new setting for the PID controller is as follows:

PB = 1.67(PB*) = 1.67(25) = 41.75%


TI = 0.5(P) = 0.5(6) = 3
TD = 0.125(P) = 0.125(6) = 0.75

Where PB = Proportional Band

TI = Time Integral

TD = Time Derivative

Next, we would tune the system using the new setting that we have obtained. However, by
using the setting that we have calculated, we are unable the desired response as shown in Figure 2.2.
Thus, we would vary the value of the PB at 10, 20 and 50. As shown in Figure 2.2, the ideal PB setting
would be PB = 20 as it has the smallest oscillation even though it takes a slightly longer time to reach
the set point.

PB = 50 PB = 20 PB = 10 PB = 10 WITH DISTURBANCE

Figure 2.2

7
Experiment 11.5.3 Ratio Control

In this experiment, we set the PID controller to Setting A which is PB = 20, I = 3s, and D = 1s.
Then, we would introduce another loop flow into the system. From the graph, there are two different
colours which indicate different flows; the green indicates the uncontrolled flow, while the red
indicates the controlled flow. As can be seen in Figure 3.0, the patterns of the two flows are different.
This is because when a step-up change is introduced for the uncontrolled flow, the controlled flow
does not change.

SETTING A SETTING B RECOMMENDED


SETTINGS WITH RATIO 1:2

Figure 3.0

Hypothetically, the flow rate of the controlled should change to always maintain a 2 to 1 ratio
of controlled flow to uncontrolled flow. We have also tried a different setting that is Setting B which is
PB = 25, I = 3s and D = 1s and also the recommended settings from our Lab Evaluator which is PB = 38,
I = 13s, and D = 2s. However, we are still unable to obtain the desirable result of 2 to 1 ratio.

8
Questions

1. Discuss the significance of step (1) in experiment 11.5.1 and step (2) in experiment 11.5.2.
What is PB, I and D represents? Why is this step necessary, and why does the values differ
from one experiment to the next?
- The significance of step (1) in experiment 11.5.1 is to observe the changes the flow
response by manipulating the derivative value. PB is the proportional band, I refers to
integral action while D is derivative action.
- The significance of step (2) in experiment 11.5.2 is to be able to provide on Proportional
Band control only when the Time Integral is set to a maximum while the Time Derivative
is set to a minimum. Then, the PB value is varied to allow us to find the natural period.

2. For each of the experiments above, discuss the results that you have obtained, and identify
the key differences in the each of the controller type and the effect of tuning parameters in
controlling the ratio of the flows to the tank. Which controller suits best for ratio control?
Why?
- For experiment 11.5.1, the key differences that can be observed, is in terms of the
stability and the response time taken to reach initial value. It was found that D = 5s, have
faster time response compared to D = 1s and 8s and are more stable. Hence, the
intermediate value is highly suitable for tuning parameter value.
- For experiment 11.5.2, the value of PB was varied. Firstly, the value that we selected for
PB was 1000, 800, 600, 400 and 200. However, after we have conducted the experiment
for PB = 600, the pattern still remains unchanged. Thus, we would reduce the value of PB
from 400 and 200 to 25 and 12.5. Only with values of PB = 25 and 12.5 we are able to
observe an oscillatory response. However, PB = 25 is more stable as shown in Figure 2.1.
- For experiment 11.5.3, we are unable to find the suitable setting to obtain the 2 to 1 ratio
of controlled flow to uncontrolled flow. However, it is best to use PID controller if the
tuning can be done properly since it eliminates offsets and produces less oscillation.

9
4. Conclusion
The objectives of this experiment were to demonstrate the characteristics of Proportional Band,
Integral Action and Derivative Action on a flow process control loop including the characteristics of
ratio control.

We were able to conclude that with lower value of derivative time, the system would be
stable as the maximum deviation is lower to the set point. Besides that, we were also able to conclude
that a lower proportional band is more suitable compared to a higher one as the lower one causes the
system to respond faster. Apart from that, in the ratio control experiment we were not able to prove
that the ratio is 1:2 after we had introduced the step change. But from the data we had, the ratio can
only be obtained initially when there is no step change occurring. In a nutshell, all the objectives were
achieved in this experiment.

5. References
By constraining CO, the controller output loses its symmetry, causing the PV to wander. (n.d.).
Derivative Action and PID Control. Retrieved October 03, 2016, from
http://controlguru.com/measurement-noise-degrades-derivative-action/

Seborg D.E., T.F. Edgar and D.A. Melliechamp, Process Dynamics and Control, John Wiley and
Sons, New York, 1989, pp 116-118

10
GROUP 8

6. Appendixes

Figure 4.1: PID Apparatus Setup

Figure 4.0: PID Control Panel

Figure 4.2: PID Apparatus Setup


11

You might also like