Experiment Title: Control Tuning

Introduction

The fundamental purpose of this experiment is the tuning of the pump flow system.

Background

In everyday life, there are different control mechanisms utilized in various chemical engineering

applications. Control mechanism is classified into two major categories of feedback control and

feed forward control. The type of control mechanism in which the measurements are utilized for

the manipulation of a variable in order to get the desired results, is termed as feedback control. It

is also called as anticipative control which actually works via prediction of the effects generated

by the measured disturbances in order to take the corrective actions to minimize the disturbances

and get the desired results. In the control mechanism of feedback controller, first of all the

variable being control is measured with the help of a measuring device and then it is compared

with the target value given to the system as the set point. This difference in comparison is called

as error and afterwards, variable is manipulated in order to get the desired result with minimizing

the error.

In general, the feedback control is used in simple thermostats at home. It is also used in the

natural processes i.e. in the regulation of blood sugar level in the body.

Figure 1: General Schematic of Feedback Control

For a controller, the control constants need to be set. One approach for choosing the “best”

values is Ziegler Nichols tuning. Ziegler Nichols is an experimental tuning method based on

quarter amplitude damping (QAD). Ku and Pu are experimentally determined and used with the

values in Table 1 below to calculate the tuned parameters based on whether P-only, PI or PID

control is used. Ku is the value of the controller gain under P-only control at which oscillations

are first observed. Ku is reached by starting with a very small gain and increasing the value in

small increments until oscillations are observed. Pu corresponds to the period of the oscillations

observed at Ku. The Ziegler-Nichols tuning rules are generally regarded as too aggressive for

most process control applications. In 1992, based on collaborative research between the Dow

Chemical Company and Lehigh University, Bjorn Tyreus and William Luyben proposed the

following modification for the traditional tuning rules.

Table 1: Ziegler-Nichols PID Setting Values

Controller Kc τ1 τD

P 0.5Ku - -

PI 0.45Ku Pu/1.2 -

PID 0.6Ku Pu/2 Pu/8

Procedure

1. Wear appropriate lab safety equipments.

2. Switch on the pump in order to begin the flow into the system.

3. Firstly, let the flow begins to run all over the system in order to check the leakages or any

problem in the system.

 4. Now, perform the tuning of the controller in order to get the data with the changing set

point with the alteration of the control method.

5. Afterwards, test the Ziegler-Nichols (ZN) proportional method with the help of an output

value and the operating range and set point.

6. Check the ZN-P and ZN-I for testing.

7. In case of third experiment, test the proportional plus integral time method for the

absolute value of error.

8. At the end, the proportional test will be done for ultimate controller gain (K cu) instead of

Kc.

9. Each test was performed for three times in order to get the data collected for consistent

average estimation.

Results

Table 2: Calculated value of Controller Gain and Integral time for each control testing method with the exception of Ultimate
gain test

Z-N Kc tau-i

P 14.47319
PI 13.15745 1.593126
ITAE Kc tau-i
PI 1.703989 0.561558
Ultimate gain Kcu 28.94638
 P-ZN Proportional only trial 1
25 100
90
20 80
70
Output (lb/min)

15 60
50
10 40
30
5 20
10
0 0
-47 -37 -27 -17 -7 3 13 23 33 43

Time (s)

Setpoint (lb/min) Flow Rate (lb/min) Steady State

Time to Steady State Pump Power (%)

P-ZN Proportional only trial 3 error

25

20

15

10

0
-51.9 -48 -44.1-40.2-36.3-32.4-28.5-24.6-20.7-16.8-12.9 -9 -5.1 -1.2 2.7 6.6 10.5 14.4 18.3 22.2 26.1 30 33.9
time

Flow Rate (lb/min) Difference Flow Rate (lb/min) Setpoint (lb/min)

 PI-ZN Proportional only trial 1
25 100

90

20 80

70
Output (lb/min)

15 60

50

10 40

30

5 20

10

0 0
-47 -37 -27 -17 -7 3 13 23 33 43

Time (s)

Setpoint (lb/min) Flow Rate (lb/min) Steady State

Time to Steady State Pump Power (%)
 ZN PI trial 1 error
30

25

20

15

10

0
1.9 4 6 8 101214161820222426283032343638404244
time

Flow Rate (lb/min) Difference Flow Rate (lb/min) Setpoint (lb/min)

 PI-ITAE only trial 1
25 100
90
20 80
70
15 60
50
Output (lb/min)

10 40
30
5 20
10
0 0
6 6 6 6 6 6 9 1 4
4. 4. 4. 4. 4. 4. 99 00 15
.
-6 -5 -4 -3 -2 -1 9 00
99 00
99 00
99 00
99 0
. 59 40
-4 5.

Time (s)

Setpoint (lb/min) Flow Rate (lb/min) Steady State

Time to Steady State Pump Power (%)

PI ITAE error trial 1

30

25

20

15

10

0
3.3 6.6 10 13.316.619.923.226.529.833.136.439.7 43
Time

Flow Rate (lb/min) Difference Flow Rate (lb/min) Setpoint (lb/min)

 P Ultimate Gain Proportional only trial 1
25 100

90

20 80

70
Output (lb/min)

15 60

50

10 40

30

5 20

10

0 0
-47 -37 -27 -17 -7 3 13 23 33 43

Time (s)

Setpoint (lb/min) Flow Rate (lb/min) Steady State

Time to Steady State Pump Power (%)

P Ultimate Gain Error Trial 1

16

14

12

10

0
3.4 6.7 10 13.316.619.923.226.529.833.136.439.7 43
Time

Flow Rate (lb/min) Difference Flow Rate (lb/min) Setpoint (lb/min)

Conclusion

The experiment was done successfully with the control tuning of the pump flow system with the

complete demonstration of the control variable for the application of the proportional and

integral time. A series of trials have been done for testing the system for data collection to get the

study on the variable effects. Each method is done with the respective formula for the quarter

decay ratio and ITAE variables. The first experimental test utilizes the Z-N proportional method

to get the output results from the power input via controller of the pump. The results has been

represented by the figures drawn with the set point and step change for the output flow and input

power of the system with respect to time. It results in the steady state within shorter time span

and poor performance of the controller with little change in the flow output. The second

experimental testing was done with the Z-N method with the integral time along with the

proportional control. This controller will reduce the effect of high gain and the controller will get

a steady state this time equal to the set point as displayed by the resulting line on the graph.

In the third experimental series, the ITAE method was used along with the PI control in order to

get the results of integral time and appropriate gain. The results were more accurate with reliable

pump flow output. It is concluded that both the results have zero error with little deviation. The

flowrate is less sensitive to the variation.