Pid Tuning Report

PID TUNING REPORT
CPE622 : PROCESS CONTROL

PRACTICES
NAME MATRIC NO. GROUP

AMIRUL AIZAD EH2206G
SALEHUDDIN
MOHAMAD NOR FARIZWAN EH2206G
NOR HANANI BT ROSLAN 2015635312 EH2206G
NURUL SYAHEERAH BT CHE 2015230056 EH2206G
HASNAN
SEMESTER MARCH 2017-JULY 2017

CONTENT PAGE
ABSTRACT 3
INTRODUCTION 3
OBJECTIVE 4
THEORY 5
PROCEDURES 11
RESULTS AND DISCUSSIONS 20
CONCLUSION 46
RECOMMENDATIONS 47
REFERENCES 48
APPENDIX 49
2
1.0 ABSTRACT
The objective of this lab is to perform open loop test, closed loop test, load disturbance test
and set point test using DCS: Foxboro and DCS: Delta-V Emerson controllers systems. For
DCS: Delta V-Emerson, the control system was performed on Gas Pressure Control Plant,
PIC92. Meanwhile for the DCS:Foxboro, the control system was performed on Flow,
FIC21. In performing closed loop system, there are four types of method to calculate Time
Delay (Td), time constant ( Tc @ t ) and response rate which are Tangent method,
Reformulated Tangent Method, Tangent and Point Method and Two Point method. The
values of P,I and D are calculated using Ziegler Nichols, CHR, Cohen Coons and
Takahashi tuning rules. The P, I and D values are then used for closed loop test, load
disturbance and set point test.
2.0 INTRODUCTION
Process control relates to statistics and engineering used for maintaining the output of a
certain process within the desired range. There are five types of control system which are
feedforward, feedback, feedforward/feedback, cascade and ratio. The process is to identify
whether involves flow, level, temperature or pressure.
Tuning is the adjustment of the controller parameters to obtain a specified closed-loop

response. There are two methods of tuning which are on-line or closed-loop tuning and step-
testing or open-loop tuning. The open loop control system is a non-feedback system in which
the control input to the system is determined using only the current state of the system and a
model of the system. There is no feedback to determine if the system is achieving the desired
output based on the reference input or set point. The closed loop control system is
a system where the actual behavior of the system is sensed and then fed back to the controller
and mixed with the reference or desired state of the system to adjust the system to its desired
state.
In performing closed loop system, there are four types of method to calculate Time Delay
(Td), time constant ( Tc @ t ) and response rate which are Tangent method, Reformulated
Tangent Method, Tangent and Point Method and Two Point method. This value is used in
3
order to find the PID. Then, after tuning is done, load disturbance test and set point changes
are performed.
In tuning process, if the response oscillates:

a) Kc is reduced -an excessively large proportional gain will lead to process instability and
oscillation.
or
b) I is increased an excessively small integral term can cause the present value to overshoot
the set point value
3.0 OBJECTIVE
The objective of the lab word is to perform an open loop test, closed loop test, load
disturbance test and set point test using Emerson and Foxboro controller systems.
4
4.0 THEORY
There are four types of method to calculate Time Delay (Td), time constant ( Tc @ t ) and
response rate which are Tangent method, Reformulated Tangent Method, Tangent and Point
Method and Two Point method.
a) TANGENT METHOD
Figure 1 : Tangent Method
The response rate, RR is calculated by:
PV/t
=
MV
Where,
() = ()
() = ()
5
b) REFORMULATED TANGENT METHOD
Figure 2 : Reformulated Tangent Method
The response rate, RR is calculated by:

tan a
=
MV
6
c) 2 POINT METHOD
Figure 3 : 2 Point Method
Tc is calculated by :
1.5 [63.2% () 28.3% ()]
Whereas,
63.2% () = 0.632
22.3% () = 0.283
7
PID TUNING
To calculate the value of PID, Ziegler-Nichols and Cohen Coons and Chien, Hrones &
Reswick (CHR) tuning rule was used.
a) ZIEGLER-NICHOLS TUNING RULES
In Ziegler-Nichols tuning rules, only requires values of RR and Td. These values are the
ones calculated from tangent, reformulated tangent or discrete tangent method.
Performance tests are in terms of set point and load disturbance.
Below is a table for calculating respective values of P,I and D.
Mode PB (%) I (Time) D(Time)

P 100 RR Td
PI 111.1 RR Td 3.33 Td
PID 83.3 RR Td 2 Td 0.5 Td
Table 1: Tuning Rules by Ziegler-Nichols
8
b) COHEN-COON TUNING RULES
Another tuning rule is Cohen-Coon Tuning rules. For this tuning rules require values of RR,
Td and Tc. Performance test is in terms of disturbance in load variable.
Below is a table for calculating respective values P,I and D.
Mode PB (%) I (Time) D (Time)

P 100

1+3
PI 100
1 + 11
3.33[ ]
1 + 11 11
1+
5
PID 100 0.37
1+
2.5[ 5 ]
1.35(1 + ) 3 1+
5 1+ 5
5
Table 2 : Cohen Coon Tuning Rule
9
c) CHIEN, HRONES & RESWICK (CHR) TUNING RULES
Another tuning rule is CHR Tuning rules. For this tuning rules require values of RR and Td
Performance test is in terms of set point.

P 333 RR Td
PI 286 RR Td 1.2 Td
PID 167 RR Td Td 0.5 Td
Table 3 : CHR Tuning Rules
d) TAKAHASHI TUNING RULES
Another tuning rule is Takahashi Tuning rules. For this tuning rules require values of RR and
Td Performance test is in terms of disturbance in load variable.

P 110 RR Td
PI 110 RR Td 3.3 Td
PID 77 RR Td 2.2 Td 0.45 Td
Table 4 : Takahashi Tuning Rules
10
5.0 PROCEDURES
5.1 DCS Delta- V Emerson
Open Loop Test for PIC92
1. The Liquid Flow Plant, PIC92 was selected.
2. After the overall diagram of the plant has opened, the controller of PIC92 was double
clicked to open the faceplate.
3. The Process History View was clicked to see the trend of the process.
4. The process need to be stabilized in manual.
5. The initial value of MV was recorded and the step change of 10% was taken.
6. After the response has reached the steady state, the response graph was printed and
calculation for obtaining the RR, td and tc.
7. The calculated value of PI controller was obtained by using Ziegler-Nichols and Cohen-
Coon method.
Closed Loop Test for PIC92
1. The controller was set to auto mode.
2. The detail icon was clicked at the faceplate in order to set the controller setting.
3. The calculated value of Gain, Kc and Reset, I was entered in the detail.
4. After the response has reached the steady state, the value of the optimum PI controller was
recorded.
5. Load disturbance test was done by turn the process into Manual mode and make a step
change of MV by 10% for three second and change into Auto mode again.
6. After the process has stabilized, the process response after changed the MV was observed.
7. The set point test was done by adding the previous set point with 10% from total value
which is 0.6m3/h.
8. After the process response has reached the steady state, the response was observed.
11
9. Both graph for load disturbance test and set point test was printed.
5.2 DCS Foxboro
Open Loop Test for FIC32
1. The Flow Control Plant, (WLF922) was selected.
2. FIC31 was selected for the control loops
3. For viewing the trending, the step are by clicking the File>Additional FoxView>Change
Env(The environment was change to operator).
4. Double clicked at the controller at the new windows of WLF922, then trend button was
clicked.
5. The process response was stabilized by setting the process in Auto mode.
6. The initial value of MV was recorded.
7. The process was changed to manual mode and step change of 10% of MV from the initial
was done.
8. After the slope can be calculated, the process was changed to Auto mode again.
9. The data for the process was printed by selecting the AIM Historian Data Display and
desired data was tagged which are SP, PV and MV.
10. Numerical analysis was done to obtain the data of RR, td and tc.
11. The determination of PI controller setting was calculated using the Ziegler-Nichols
method.
12
Close Loop Test for FIC31
1. After the calculated value of PI was done, the value then was inserted into the controller
setting of PB and I.
2. After the response has reached the steady state, the optimum controller setting was
recorded.
3. The load disturbance test was done by changing the process response into Manual mode
and change 10% of the current value was done for three seconds and set to Auto again.
4. The response was observed until it stabilized.
5. The set point test was done by adding the previous value set point with 0.1 of total value
which is 1.18m3/h.
6. The response was observed until it stabilized.
7. The process response than was printed by following the step

START>Program>HyperSnap-DX>HyperSnap-DX>Capture>Active Window.
8. The trending active window was selected and then the image was inverted to black and
white. The process response than was printed.
5.3 Level Controller, LIC92
Open Loop Test for LIC92
1. The process was stabilized in manual mode.
2. The initial value of manipulated variable ( ) and the initial value of process
variable ( ) were recorded.
3. A step change of 10 % was made to the manipulated variable (MV)
4. After the process has increased to about 7 10 % of process span, the controller is set
to automatic mode and let to stabilize to the original operating condition.
5. After the response has reached the steady state, the response graph was printed
13
6. Response Rate (RR), Dead Time ( ) and Time Constant ( ) were calculated by
using Reformulated tangent method, Tangent method and Tangent and Point method.
7. The average value of Response Rate (RR), Dead Time ( ) and Time Constant ( )
from three methods were obtained.
8. The calculated value of PI controller was obtained by using Ziegler-Nichols and

Cohen-Coon method and Takahashi
Close Loop Test for LIC92 (Ziegler-Nichols)
2. The calculated value of PB and I which are 3.30 % 6.00 s respectively by Ziegler
Nichols was inserted in the detail.
3. After a while, the value of I was changed to 3.00 s to optimize the controller.
4. After the response has reached the steady state, the value of the optimum PI controller
was recorded.
5. Load disturbance test was done by turn the process into Manual mode and make a
step change of Manipulated Variable (MV) by 10% from 44.6 % to 54.6 % for three
second and change to Auto mode again.
6. After the process has stabilized, the process response after changed the MV was
observed.
7. The set point test was done by changing the value from 4003 to 500 3 .
Closed Loop Test for LIC92 (Cohen Coon)
2. The calculated value of PB and I which are 2.90 % 5.60 s respectively by Cohen Coon
was inserted in the detail
14
was recorded.
observed.
Closed Loop Test for LIC92 (Takahashi)
11. The calculated value of PB and I which are 3.20 % 6.00 s respectively by Takahashi
was inserted in the detail
was recorded.
step change of Manipulated Variable (MV) by 10% from 43.1% to 53.1% for three
observed.
15
5.4 Liquid Flow Plant, FIC21
Open Loop Test for FIC21
1. The process was stabilized in manual mode.
2. The initial value of manipulated variable ( ) and the initial value of process
variable ( ) were recorded.
3. A step change of 10 % was made to the manipulated variable (MV)
4. After the process has increased to about 7 10 % of process span, the controller is set
to automatic mode and let to stabilize to the original operating condition.
5. After the response has reached the steady state, the response graph was printed
6. Response Rate (RR), Dead Time ( ) and Time Constant ( ) were calculated by
using Reformulated tangent method, Tangent method and Tangent and Point method.
7. The average value of Response Rate (RR), Dead Time ( ) and Time Constant ( )
from three methods were obtained.
8. The calculated value of PI controller was obtained by using Ziegler-Nichols and

Cohen-Coon method and Takahashi.
Closed Loop Test for FIC21 (Ziegler-Nichols)
2. The calculated value of PB and I which are 94.4306 % 12.000 s respectively by

Ziegler Nichols was inserted in the detail
was recorded.
observed.
16
6. The set point test was done by changing the value from 3.00 3 / to 3.50 3 /.
Closed Loop Test for FIC21 (Cohen Coon)

was recorded.
observed.
Closed Loop Test for FIC21 (Takahashi)

3. After a while, the value of PB and I were changed to 182.00 % and 6.00 s to optimize
the controller.
was recorded.
17
step change of Manipulated Variable (MV) by 10% from 54.5% to 64.5% for three
observed.
5.5 Gas Flow System
Open Loop Test for Gas Flow System
1. The system was switched on.
2. Process was stabilized in Automatic (A) mode.
3. Response of the process was recorded by pressing the Recorded button at the chart
plotter
4. Initial Manipulated Variable (MV), in percent, was recorded.
5. The mode was switched to Manual (M) mode.
6. Step change of 10% was made on the MV.
7. After the process reached new steady state, the Recorded button was pressed to stop
response curve.
8. Response Rate (RR), time delay (Td) and time constant (Tc) were calculated by
Tangent method.
9. PI controller setting was determined using Zingler Nicholes, Cohen Coons and Chien,
Hrones &Reswick (CHR) method.
18
Close Loop Test for Gas Flow System
1. Response of the process was recorded by pressing the recorded button at the chart
2. Controller was set in automatic mode.
3. The values of P and I obtained was inserted into at Proportional and Integral section.
4. Tuning was done when the response was oscillate to achieve stable condition.
5. The optimum PI controller setting was recorded.
6. Then the controller was set in manual mode.
7. MV was changed 10% to the current MV
8. After 3s, the controller was set in Automatic mode.
9. Wait until the response stable.
10. Then set point test was continued from the load disturbance test.
11. The set point was changed for 10% from current set point.
12. The process was waited until the response become stable.
19
6.0 RESULTS AND DISCUSSIONS
6.1 DCS Delta-V Emerson
Result Open Loop for PIC92
For Emerson, the gas pressure is chosen in this test and denote by PIC 92. For pressure, the
process behaviour is self-regulating and fast response which means that it can stabilize by
itself and have the final steady state fast.
= 45 %
= 55 %
Initial reading Final reading Change of variable
Manipulated variable 45 55 MV =
(%) = 55% - 45%
= 10%
Sample calculation to find RR and average
a) Using Tangent Method:
a = 2.5%/17mm
= 0.1471 %/ mm
b =60s/45mm
= 1.3 s/ mm
PV = -
= 51% - 35.6%
= 15.4%
X = 19.5 mm
Y = 30 mm
20
Time constant, = () x b
= 19.5 mm x 1.3 s/ mm
= 25.35 s
Dead time, = () x b
= 6 mm x 1.3 s/ mm
= 7.8 s
30%
19.5
RR = ( )/MV = 10%
= 25.35/s
b) 2-point method
a = 2.5%/17mm
= 0.1471 %/ mm
b =60s/45mm
= 1.3 s/ mm
Angle = 56
Time constant, = 1.5 ( 12.5 6 ) x 1.3
= 12.675 s
30%
19.5
RR = ( )/MV = 10%
= 25.35/s
21
c) Reformulated Tangent Method
b =60s/45mm
= 1.3 s/ mm
PV63.2= 0.632 30
= 18.96
PV28.3= 0.283 30
= 8.49
Dead time, = () x b
= 6 mm x 1.3 s/ mm
= 7.8 s
Time constant, = 25.35 s
tan 56 0.1471
RR = (tan )/MV = ( )
10% 1.3
= 25.35 s
RR average = (0.1741 + 0.01741 + 0.01678) / 3
= 0.0172
Td = 1.8
Tc average = (25.35 + 25.35 + 12.675) / 3
= 21.125 s
22
Discussion Open Loop for PIC92
For Emerson, the open loop test is first conducted in order to obtain the response rate
(RR), dead time ( ) and the time constant ( ). For pressure process, the mode choose is PI
mode because it is fast response. The slope from the open loop test is used to calculate by
adding the MV by 10%. For this process, three formula is used to calculate the RR and
which is Tangent method, 2 point method and Reformulated tangent method. And then the
RR and average from all three method is calculate to obtain the P and I. And for tuning
rule, method used is Ziegler Nichol and Cohen-Coon tuning rule.
Result Close Loop for PIC92
Optimum controller setting (Ziegler-Nichols method)
P = 111.1 RR Td
= 111.1 0.0172 x 7.8
= 14.91%
I = 3.33 Td
= 3.33 7.8
= 25.97 s
Optimum controller setting (Cohen-Coons method)
100 2
P= RR Td = = = 0.369
1+ 21.125
11
= 100 0.369 0.0172 7.8

(1 + )
11
= 12.98 %
23

1+
11
I = 3.33 [ 11 ]
1+
5
0.369
1+
11
= 3.33 [ 110.769 ] 7.8
1+
5
= 14.82 s
Discussion Close Loop for PIC92
For closed loop test, the P and I controller were obtained from the calculation of
Ziegler-Nichols tuning and Cohen-Coon tuning rule. For the Ziegler Nichols, the value of P
is 14.91 and I is 25.97. And for the Cohen-Coon tuning rule, the value of P is 12.98 and I is
14.82. Because of the gas pressure is hard to get stable due to open and close. For the Ziegler-
Nichol tuning rule, when undergo the disturbance test, the time for the MV to reach the set
point is not too long which is good. And for the set point test, unfortunately the MV take too
long to reach the set point. The I need to be tune again by dividing them by 4 to fasten the
MV to reach the set point. And for the Cohen-coon tuning rule, for the disturbance test is
same with the Ziegler-Nichols which is fast. And for the Set point test, unlike the Ziegler-
Nichols rule, it is fast to reach the set point.
From both tuning rule, Cohen-Coon rule resulted better response compare to the
Ziegler-Nichol for the process response for the gas pressure.
24
6.2 DCS Foxboro
Result Open Loop for FIC21
Numerical Method
Process
Time, s PV (m3/h) span PV (%) RR, 1/s Td (s) Tc (s)
(m3/h)
0 0.679 3.00 22.63 -
1 0.679 3.00 22.63 0
2 0.679 3.00 22.63 0.2055
3 0.802 3.00 26.74 0.453 2 3.06
4 0.950684 3.00 31.69 0.373
5 1.0261 3.00 34.20 0.1915
6 1.0655 3.00 35.52 0.0975
7 1.0850 3.00 36.15 0.049
8 1.0951 3.00 36.5 0.0175
9 1.0951 3.00 36.5 0
10 1.0951 3.00 36.5 -
Sample Calculation
PV1 = 31.69 %
PV-1 = 22.63%
h = 1s
MV = MVf MVI
= 45-35
= 10%

, =

31.69 22.63
RR1 =
2(1)(10)
25
= 0.453/s

, = [ ]

31.69 22.63
= 4 2(1) [ ]
31.69 22.63
= 2s

, = [ ]

36.5 31.69
= 2(1) [ ]
31.69 22.63
= 3.06
Discussion Open Loop for FIC21
As we know, the liquid flow control is one of the self-regulating type process which is
the system is kept in manual mode. After the system become stable, the system is change to
the manual mode in order to carry the open loop test before increase the value of MV by
10%. The initial value of MV was 35%, when increased to 10%, the value of MV was 45%.
Due to the fast process of the liquid flow process, the PI mode is chosen compare to the P and
PID mode. By using the numerical method, the Dead Time and Constant Time is chosen at
the highest Respond Rate. From the value obtained, the optimum value of controller setting
can be calculated by using either Ziegler-Nichols or Cohen-Coons tuning rule.
26
Result Close Loop for FIC21
I) Optimum Controller Setting (Ziegler-Nichols Method)
P = 111.1 Td RR
= 111.1 (2) (0.452)
= 100.67 %
I = 3.33 Td
= 3.33 (2)
= 6.66 s
II) Optimum Controller Setting (Cohen-Coon Method)
Td 2
= = = 0.653
Tc 3.06
100
P= RR Td
1 + 11
100
= (0.453) (2)
0.653
1 + 11
= 85.55 %

1 + 11
I = 3.33 [ ] Td
11
1+
5
0.653
1+
11
= 3.33 [ ] (2)
11(0.653)
1+
5
= 2.88s
27
Discussion Close Loop for FIC21
To obtain the controller setting value, the method used is Ziegler-Nichols tuning
rules in order to calculate the value of P and I. The calculated value is used for closed loop
test. For this system which is Foxboro, band is used instead of Kc and the band value is equal
to Kc. Because of the band is equal to Kc, the value of P is directly inserted. But for the
integral value, the reset is used which is equal 1/I. After the calculated value is inserted in the
system, the graph was observed. From the graph, there is no oscillation occur. Thats mean
the value of band inserted which is 100.67% is right and no tuning need to be done.
And then, the disturbance test was carry out as soon the value of controller setting was
obtained. In order to carry out the disturbance test, the process was set into the manual mode
first and then increase the value of MV 10 % so it change from 44.1% to 54.1%. Within 3
second, the mode is change back to the auto mode. From the graph, it returns to the set point
back without any oscillate but the time it returns to the set point is quite long. That means the
value of reset need to be tuned a little bit and improved in order to get the optimum value.
And lastly the set point test was carry out by changing the set point from 1.08m3/h to
1.18m3/h. It shows that the increasing is fast and oscillate does not occur.
28
6.3 Level Controller, LIC92
Result Open Loop for LIC92
I. Reformulated Tangent Method
a = 10 % / 10 mm
= 1 %/mm
b = 3600 s / 2000 mm
= 1.8 s / mm
= 12
= 30 mm
= 10 %
= 1 mm
= 1 mm x 1.8 s / mm
= 1.8 s
tan 12 1
RR =
10 1.8
= 0.0118 / s
= 30 mm x 1.8 s / mm
= 54 s
29
II. Tangent Method
a = 1 %/ mm
b = 1.8 s / mm
= 30 mm
= 10 %
= 10
= 1 mm
= 1 mm x 1.8 s / mm
= 1.8 s
1030 1
RR =
10 1.8
= 0.0185 / s
= 30 mm x 1.8 s / mm
= 54 s
30
III. Tangent and Point Method
a = 1 %/ mm
b = 1.8 s / mm
= 30 mm
= 10 %
= 10
= 1 mm
= 1 mm x 1.8 s / mm
= 1.8 s
1030 1
RR =
10 1.8
= 0.0185 / s
63.2 = 0.632 x 10
= 6.3 mm
63.2 = 21 mm
= (21 1) x 1.8 s / mm
= 36 s
31
Discussion Open Loop Test for LIC92
Open loop test was being conducted in order to identify the open loop process
response such as the response rate (RR), dead time ( Td ) and time constant ( Tc ). The process
slope can be used to calculate the time constant, dead time and response rate. Reformulated
tangent method, tangent method and point and tangent method was used in order to calculate
the value of RR, and . The average value of all three of the method used was then
obtained in order to calculate the optimum controller setting in closed loop later on. The
value of RR, and are 0.0163/s, 1.8 s and 48 s respectively.
RR
Reformulated Tangent Method 1.8 s 0.0118 / s 54 s
Tangent Method 1.8 s 0.0185 / s 54 s
Tangent and Point Method 1.8 s 0.0185 / s 36 s
Average 1.8 s 0.0163 / s 48 s
Table 5 : The value of Response Rate (RR), Dead Time (Td) and Time Constant (Tc)
Result Close Loop for LIC92
I. Ziegler-Nichols
PB = 111.1 (0.0163) (1.8)

= 3.2597 %
I = 3.33 (1.8)
= 5.994 s
II. Cohen-Coon
= 1.848
32
= 0.0375
100
PB = 0.0375 (0.0163) (1.8)
1+
11
= 2.9240 %
0.0375
1+
11
I = (3.33) 11(0.0375) (1.8)
1+
5
= 5.5561 s
III. Takahashi
PB = 110 (0.0163) (1.8)

= 3.2274 %
I = 3.3 (1.8)
=5.94s
33
Discussion Close Loop Test for LIC92
PI mode is chosen because the process is fast and no further adjustments need to be
made. Proportional Band (PB) and Integral Time (I) is calculated by using 3 tuning rules;
Ziegler-Nichols, Cohen Coon and Takahashi.
The value of PB and I calculated by Ziegler-Nichols are 3.2597 % and 5.9940 s

respectively. These values were inserted in the controller for closed loop process response.
After few seconds, the process is considered as taking a long-time stable and reach the set
point. The value of I is optimized from 5.9940 s to 3 s. As a result, a shorter time obtained for
the process to reach the set point. Load disturbance test changing from 44.6 % to 54.3 % and
Set point test changing from 400 to 500 are made. The result is shown in the Figure 2.
The value of PB and I calculated by Cohen Coon are 2.9240 % and 5.5561 s
respectively. These values were inserted in the controller for closed loop process response.
After few seconds, the process is stable without further tuning optimization needed. Load
disturbance test changing from 47.6 % to 57.6 % and Set point test changing from 400 to 500
are made. The result is shown in the Figure 3.
The value of PB and I calculated by Takahashi are 3.2274 % and 5.94 s respectively.
These values were inserted in the controller for closed loop process response. After few
seconds, the process is stable without further tuning optimization needed. Load disturbance
test changing from 43.1 % to 53.1 % and Set point test changing from 400 to 500 are made.
The result is shown in the Figure 6.
Tuning Rules PB I
Ziegler-Nichols 94.4306 % 12.000 s
Cohen Coon 83.7932 % 9.0250 s
Takahashi 62.5284 % 11.988 s
Table 6 : The value of Proportional Band (PB) and Integral Time (I)
34
6.4 Liquid Flow Plant, FIC21
Result Open Loop for FIC21
I. Reformulated Tangent Method
a = 20 % / 20 mm
= 1 %/mm
b = 3600 s / 1000 mm
= 3.6 s
= 84
= 30 mm
= 10 %
= 1 mm
= 1 mm x 3.6 s / mm
= 3.6 s
tan 84 1
RR =
10 3.6
= 0.2643 / s
= 3 mm x 3.6 s / mm
= 10.8 s
35
II. Tangent Method
a = 1 %/ mm
b = 3.6 s / mm
= 3 mm
= 10 %
= 24 mm
= 1 mm
= 1 mm x 3.6 s / mm
= 3.6 s
243 1
RR =
10 3.6
= 0.222 / s
= 3 mm x 3.6 s / mm
= 10.8 s
36
III. Tangent and Point Method
a = 1 %/ mm
b = 3.6 s / mm
= 3 mm
= 10 %
= 24
= 1 mm
= 1 mm x 3.6 s / mm
= 3.6 s
243 1
RR =
10 3.6
= 0.222 / s
63.2 = 0.632 x 24
= 15 mm
63.2 = 2 mm
= (15 2) x 3.6 s / mm
= 46.8 s
37
Discussion Open Loop Test for FIC21
Open loop test was being conducted in order to identify the open loop process response
such as the response rate (RR), dead time ( Td ) and time constant ( Tc ). The process slope can be
used to calculate the time constant, dead time and response rate. Reformulated tangent method,
tangent method and point and tangent method was used in order to calculate the value of
RR, and . The average value of all three of the method used was then obtained in order to
calculate the optimum controller setting in closed loop later on. The value of RR, and are
0.2361/s, 3.6 s and 22.8 s respectively.
RR
Reformulated Tangent Method 3.6 s 0.2643 / s 10.8 s
Tangent Method 3.6 s 0.2220 / s 10.8 s
Tangent and Point Method 3.6 s 0.2220 / s 46.8 s
Average 3.6 s 0.2361 / s 22.8 s
Table 7 : The value of Response Rate (RR), Dead Time (Td) and Time Constant (Tc)
Result Close Loop for FIC21
I. Ziegler-Nichols
PB = 111.1 (0.2361) (3.6)

= 94.4306 %
I = 3.33 (3.6)
= 12 s
38
II. Cohen-Coon
= 3.622.8
= 0.1579
100
PB = 0.1579 (0.2361) (3.6)
1+
11
= 83.7932 %
0.1579
1+
11
I = (3.33) 11(0.1579) (3.6)
1+
5
= 9.0250 s
III. Takahashi
PB = 110 (0.1579) (3.6)

= 62.5284 %
I = 3.3 (3.6)
= 11.988 s
39
Discussion Close Loop Test for FIC21
PI mode is chosen because the process is fast and no further adjustments need to be
made. Proportional Band (PB) and Integral Time (I) is calculated by using 3 tuning rules;
Ziegler-Nichols, Cohen Coon and Takahashi.
The value of PB and I calculated by Ziegler-Nichols are 94.4306 % and 12.000 s

respectively. These values were inserted in the controller for closed loop process response. After
few seconds, the process is stable without further tuning optimization needed. Load disturbance
test changing from 58.2 % to 68.2 % and Set point test changing from 3.00 to 3.05 are made. The
result is shown in the Figure 2.
The value of PB and I calculated by Cohen Coon are 83.7932 % and 9.0250s
respectively. These values were inserted in the controller for closed loop process response. After
few seconds, the process is stable without further tuning optimization needed. Load disturbance
test changing from 58.3 % to 68.3 % and Set point test changing from 3.00 to 3.05 are made. The
result is shown in the Figure 3.
The value of PB and I calculated by Takahashi are 62.5284 % and 11.988 s respectively.
These values were inserted in the controller for closed loop process response. After few seconds,
the process is stable without further tuning optimization needed. Load disturbance test changing
from 54.5 % to 64.5 % and Set point test changing from 2.66 to 3.10 are made. The result is
shown in the Table 8
Tuning Rules PB I
Ziegler-Nichols 94.4306 % 12.000 s
Cohen Coon 83.7932 % 9.0250 s
Takahashi 62.5284 % 11.988 s
Table 8 : The values of Proportional Band (PB) and Integral (I).
40
6.5 Gas Flow System
Result Open Loop for Gas Flow System
Response Rate (RR), s-1 Time delay (Td), s Time constant (Tc), s
Tangent Method (TM) 0.03125 7.2 28.8
Two Points (2P) 0.03125 7.2 15.0768
Reformulated Tangent 0.03967 7.2 28.8
(RTM)
RRAvg = 0.034056/s
Tc, Avg = 24.2256 s
Td = 7.2 s
Discussion Open Loop for Gas Flow System
For open loop test, the response was not give a stable response as it is in gas phase. Gas
phase is rarely stable, therefore the response when increase in MV is not smooth. It is overshoot
than after few seconds, it reached to the set point. Then, the RR, Td and Tc were calculated for
the close loop test. Three methods was used to calculate the RR, Td and Tc which are tangent
method (TM), two points method (2P) and reformulated tangent method (RTM). Then, three
tuning rules used in order to achieve the objective of the experiment which is to compare the
response for each tuning rules.
41
Result Close Loop for Gas Flow System
Proportional, P Integral, I
ZN tuning 27.2421 23.976
CC tuning 23.8755 14.8907
CHR (20% overshoot) 40.9489 28.8
SAMPLE OF CALCULATION
a = 10% / 10mm b = 3600s / 1000mm
= 1 %/mm = 3.6 s/mm
y = 9 mm x = 8 mm
The calculation of Time delay (Td), is same for all three methods by using equation, which is:
s
Td = 2 mm X 3.6
mm
= 7.2 s
42
RR Tc
TM 98 1
= 0.03125 8 3.6 = 28.8
10 3.6
2P 98 1 Tc = 1.5 (T63.2 T28.3)

= 0.03125
10 3.6 = 1.5 [3.6 (5.056-2.264)]
=15.0768
RTM tan 55 1
= 0.03967 8 3.6 = 28.8
10 3.6
RRAvg = (RRTP + RR2P + RRRTM) / 3
= (0.03125 + 0.03125 + 0.03967) / 3
= 0.034056
TC, Avg = (TC, TP + TC, 2P + TC, RTM) / 3
= (28.8 + 15.0768 + 28.8)
= 24.2256
43
Tuning Setting
For CC, the need to calculate which the equation is:
= Td / TC
= 7.2 / 24.23
= 0.2971
P I
ZN 111.1 (0.034056) (7.2) 3.33 (7.2)

= 27.2421 = 23.976
CC 100
0.034056 7.2 1 + 0.297111
0.2971 3.33(7.2) [ ]
1 + 11 11(0.2971)
1+
5
= 23.8755
= 14.8907
CHR 167 (0.034056) (7.2) 4 (7.2)

= 40.9489 = 28.8
44
Discussion Close Loop for Gas Flow System
Tuning that used are Ziegler Nichols (ZN), Cohen Coon (CC) and Chien, Hrones &
Reswick (CHR). The settling criteria for ZN and CC are quarter amplitude damping (QAD)
while CHR is 20% overshoot and minimum time response. For the performance test, ZN is use
for set point and load disturbance while both CC and CHR for load disturbance test only.
For ZN tuning, the result obtained in the Appendix section, the response oscillate when
the value of P and I was inserted. As in the procedure, when oscillation happened, tuning need to
be performed. The value of P is multiply by 4, the response gives high off-set. Then, the value of
I is reduced from 24s to 6s, the response become stable after a few seconds.
CC tuning also oscillate when the P and I calculated was inserted into the controller
hence, the value of P was changed to 95.6 while I is still the same. It takes a 288s to settle then,
the load disturbance test was changed from 74 to 64 the time taken to reach the set point is
212.4s. After that, set point test was conducted from 25% to 30%. The response was overshoot at
first then became stable after 126s.
As the response was overshoot for open loop test, the CHR (20% overshoot) was chosen.
The value obtained for P is 40 and I is 28.8s. Both values were inserted in the gas flow
controller. Since the response oscillates, the fine tuning was made. The value of P was change to
102%. The response overshoot and takes a long time to stable therefore, value of I is reduced by
dividing 3 which give the value of 9. As the I is to reduced overshoot and shorter the time to
stable, the time taken for the response to settle after changing the value of I is 216s. Then, load
disturbance test was conducted from 75% to 65%. The response takes 360s it is because the
value of I is quite high and the gas phase is basically takes a very long time to stable. For the set
point test, the change was made from 25% to 30%. The response takes 252s to settle.
Therefore, CC is much better than the other 2 tuning settings. It is because, when fine
tuning was made, CC takes shorter time than ZN and CHR, and the value of P for fine tuning is
just 95.6%. For load disturbance test, time taken for ZN and CC are same which is 252s while
CHR takes 360s. It is due to the I value of CHR is higher than ZN and CC. For set point test, the
45
time taken for CC is shorter which is 252s while Zn is 288s although the I is 6s, and CHR is 270s
with the I is 9s. Hence, we can say that CC gives better result than ZN and CHR tuning setting.
7.0 CONCLUSIONS
In conclusion, the process of each system is either self regulating or non- self regulating
process. A self regulating process, for example flow, temperature, pH and pressure, is a process
whereby the process moves to a new steady state position after making an open loop test. While
the non-self regulating process such as level, the process keep changing until the process reach
its limit after the open loop test was made.
There are four methods to calculate RR, Td and Tc which are tangent method, tangent and
point method, two points method and reformulated tangent method. The value of RR, Td and Tc
is used to calculate the tuning setting, P, I and D. P. In this experiment, the value of P and I was
calculated.
After inserting the value of P and I, if the response oscillate, Thumbs Rule is apply.
Thumbs Rule is used as a guideline to obtain the fine tuning. The value of P and I is either
divide or multiply by 4. The function of P is to eliminate the offset while I is to reduce error.
Then, the load disturbance test and set point was conducted. The load disturbance test
was performed by changing the value of MV while set point change is changing the either
increase or decrease the previous set point.
46
8.0 RECOMMENDATIONS
1. In order to minimize the error when doing test, make sure to understand the process
characteristics and procedure of the process before conducting lab work.
2. In open loop test, manipulated variable need be set to initial value after several second to
avoid the process reaching the limit.
3. Preferable to used 4 decimal places in every calculation to get accurate result of P, I and
D.
4. If the response oscillate, adjust the P, I and D setting depending on the process either
increase or decrease the value of setting.
5. When adjusting P and I value, make sure to adjust one by one to prevent extreme
oscillation which may produce large error.
47
9.0 REFERENCES
1. Inc., L. T. (n.d.). Ziegler-Nichols Tuning Rules for PID. Retrieved June 12, 2017, from
http://www.mstarlabs.com/control/znrule.html
2. Integral Action and PI Control. (n.d.). Retrieved June 12, 2017, from
http://controlguru.com/integral-action-and-pi-control/
3. PID for Dummies. (n.d.). Retrieved June 12, 2017, from

https://www.csimn.com/CSI_pages/PIDforDummies.html
4. Tuning PID control loops for fast response. (n.d.). Retrieved June 12, 2017, from
http://www.controleng.com/single-article/tuning-pid-control-loops-for-fast-
response/495b3c78823d6ccfa58f2f83d58dc85c.html
5. Ziegler Nichols method. (n.d.). Retrieved June 12, 2017, from

http://kurser.iha.dk/m/mtpri1/control/3_PID/f_ziegler/ziegler_nichols_method.html
48

Pid Tuning Report

Uploaded by

Copyright:

Available Formats

Pid Tuning Report

Uploaded by

Document Information

Original Description:

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Pid Tuning Report

Uploaded by

Copyright:

Available Formats

PID TUNING REPORT

CPE622 : PROCESS CONTROL

NAME MATRIC NO. GROUP

SEMESTER MARCH 2017-JULY 2017

RESULTS AND DISCUSSIONS 20

Tuning is the adjustment of the controller parameters to obtain a specified closed-loop

In tuning process, if the response oscillates:

Figure 1 : Tangent Method

The response rate, RR is calculated by:

Figure 2 : Reformulated Tangent Method

The response rate, RR is calculated by:

Figure 3 : 2 Point Method

1.5 [63.2% () 28.3% ()]

a) ZIEGLER-NICHOLS TUNING RULES

Below is a table for calculating respective values of P,I and D.

Mode PB (%) I (Time) D(Time)

Table 1: Tuning Rules by Ziegler-Nichols

Below is a table for calculating respective values P,I and D.

Mode PB (%) I (Time) D (Time)

Table 2 : Cohen Coon Tuning Rule

Below is a table for calculating respective values P,I and D.

Mode PB (%) I (Time) D (Time)

Table 3 : CHR Tuning Rules

d) TAKAHASHI TUNING RULES

Below is a table for calculating respective values P,I and D.

Mode PB (%) I (Time) D (Time)

Table 4 : Takahashi Tuning Rules

5.1 DCS Delta- V Emerson

Open Loop Test for PIC92

1. The Liquid Flow Plant, PIC92 was selected.

4. The process need to be stabilized in manual.

Closed Loop Test for PIC92

1. The controller was set to auto mode.

5.2 DCS Foxboro

Open Loop Test for FIC32

1. The Flow Control Plant, (WLF922) was selected.

2. FIC31 was selected for the control loops

6. The initial value of MV was recorded.

4. The response was observed until it stabilized.

6. The response was observed until it stabilized.

7. The process response than was printed by following the step

5.3 Level Controller, LIC92

Open Loop Test for LIC92

1. The process was stabilized in manual mode.

3. A step change of 10 % was made to the manipulated variable (MV)

8. The calculated value of PI controller was obtained by using Ziegler-Nichols and

Close Loop Test for LIC92 (Ziegler-Nichols)

1. The controller was set to auto mode.

Closed Loop Test for LIC92 (Cohen Coon)

1. The controller was set to auto mode.

Closed Loop Test for LIC92 (Takahashi)

10. The controller was set to auto mode.

Open Loop Test for FIC21

1. The process was stabilized in manual mode.

3. A step change of 10 % was made to the manipulated variable (MV)

8. The calculated value of PI controller was obtained by using Ziegler-Nichols and

Closed Loop Test for FIC21 (Ziegler-Nichols)

1. The controller was set to auto mode.

2. The calculated value of PB and I which are 94.4306 % 12.000 s respectively by