EPF 3304, SEM 1, 20/21

SEMESTER 1, 2012/2021

EPF3304-1
PROCESS CONTROL

ASSIGNMENT 8: CLASS EXERCISE FOR TUNING 10.2

LECTURER’S NAME: PROF. MADYA DR. FARAH SALEENA

BINTI TAIP

GROUP: 2
SUBMISSION DATE: 9/1/2021

GROUP MEMBERS MATRIC’S NUMBER

NURUL JANNAH BINTI MOHD SHUKERI 198968
DIYANAH NABIHAH BT MOHD SHAFEK 194897
NURUL NAIMAH BINTI OTHMAN 196801
SITI NUR AFIEKAH BINTI MOHD GHANI 197702
WAN NUR SYAFIZA BINTI WAN ZULKIFLI 198850

ASSIGNMENT 8 CLASS EXERCISE FOR TUNING 10.2

 EPF 3304, SEM 1, 20/21

DATE OF SUBMISSION : (WK 9)

DATE OF PRESENTATION: WEEK 9
LENGTH OF REPORT: MAXIMUM 35 PAGES (EXCLUDING APPENDIX)
PLEASE SEND THE SOFTCOPY OF REPORT BY THE END OF WEEK 9

Consider a plant with a model given by

Find the parameters of a PID controller using the different methods. Based on the value
of tuning parameters, discuss the predicted performance of the output responses

2(s+0 .5 ) −5 s
G(s)= e
(s+2)(2 s+1)
Use the excel file or simulation.
1. Approximate it to be 1st order process
2. Use the methods learnt in class: ZN, TL, CC and IE method for setpoint change
3. Discuss and compare the values of Kc Ti and Td obtained using the different
methods. Predict the performance of the tuning methods
4. Run the simulation
5. Discuss the performance of the tuning methods based on several performance
criteria such as overshoot, rise time and settling time

SOLUTIONS:
1. First Order Equation
A general first order system can be written as:

A general Transfer function:

 Note that gain and time constant define the behaviour of a first order
process.
Hence,
2 ( s +0.5 ) −5 s
G(s) = e
( s+ 2 )( 2 s +1 )
 EPF 3304, SEM 1, 20/21

2 s +1
G(s) = e−5 s
( s+ 2 )( 2 s +1 )
We need to cancel out the numerator and denominator (2s+1),
1
G(s) = e−5 s
(s+2)
Then, divide by 2,
½ −5 s
e 0.5 −5 s
G(s) = s+2 = e eqn.1
( 0.5 s+1 )
2
1st order system with gain K, dead time θ and time constant τ; 3 parameters to be
fitted,
K e−θs
G(s) = eqn.2
(τs+1)
Compare between equations 1 and 2 to get the value of K, dead time θ and time
constant τ,
K = 0.5
θ=5
τ = 0.5

From the simulated graph below,

Take two peaks of stable oscillation (blue color):
Pu = 184 – 164
= 20
EPF 3304, SEM 1, 20/21

Figure 1 Simulation Graph

 EPF 3304, SEM 1, 20/21

ZN METHOD

Table 1 Kc T1 Td ZN Method

Kc τ1 τD
0.6 K CU PU /2 PU /8
= 0.6(1) = (20)/2 = (20)/8
= 0.6 = 10 = 2.5

TL METHOD
 EPF 3304, SEM 1, 20/21

Table 2 Kc T1 Td TL Method

Kc τ1 τD
0.45 K CU 2.2 PU PU /6.3
= 0.45(1) = 2.2 (20) = (20)/6.3
= 0.45 = 44 = 3.17

IE METHOD
 EPF 3304, SEM 1, 20/21

Derive the relation below then we get:

Table 3 Kc T1 Td IE Method

Kc τ1 τD
K KC = A (θ/ τ)B τ/ τ 1 τ D /¿ τ

5
−0.85
= 0.796 + (-0.1465)*( 5 −0.85
= 0.965 ( ) = 0.965 ( )
0.5 0.5
5
= 0.1363 ¿ = 0.1363
0.5
Kc τD
= -0.669
= 0.1363/0.5 τ1 = 0.1363*(0.5)
= 0.2726 = 0.5/-0.669 = 0.06815

= -0.7474
CC METHOD
 EPF 3304, SEM 1, 20/21

Table 4 Kc T1 Td CC Method

Kc τ1 τD
K K C = 1.35
τ
θ +0.270 τ 1 /τ =θ τ D /¿ τ
5 [32+6 (5/0.5)] 0.37(5 /0.5)
= 1.35(0.5/5) + 0.270
= 5 = 5
13+ 8( ) 1.0+ 0.2( )
= 0.405 0.5 0.5
Kc = 4.946 = 1.233
= 0.405/0.5 τ1 τD
= 0.81 = 4.946*(0.5) = 1.233*(0.5)
= 2.4731 = 0.6167

Graph Comparison
1) ZN Method
 EPF 3304, SEM 1, 20/21

Figure 2 Simulation Graph ZN Method

2) TL Method
 EPF 3304, SEM 1, 20/21

Figure 3 Simulation Graph TL Method

3) IE Method
EPF 3304, SEM 1, 20/21

Figure 4 Simulation Graph IE Method

 EPF 3304, SEM 1, 20/21

4) CC Method

Figure 5 Simulation Graph CC Method

Question 3:
 EPF 3304, SEM 1, 20/21
Table 5 Kc T1 Td for ZN TL IE and CC Method

METHOD Kc τ1 τD
ZN 0.6 10 2.5
TL 0.45 44 3.17
IE 0.2726 -0.7474 0.06815
CC 0.81 2.4731 0.6167

The value of Kc, T1 and TD had been obtaining from the four different method. The value of the
controller gain, Kc for ZN method is 0.6, T 1 is 10 and TD is 2.5. The graph obtain was not that
stable although it reaches the set point. The value of Kc for TL method is 0.45, 44 for T 1 and
3.17 for Td. As the value of T1 increase the value Kc decrease, this will lead to faster set-point
response. Thus, the graph showed for TL method had reach the set point and least overshoot.
Next, for the IE method the value of Kc was 0.2726. T 1 value was -0.7474 and TD was 0.6167.
The graph obtain was didn’t reach the set point and there was no overshoot. Lastly was CC
method. The Kc value obtain was 0.81 which was the highest value compare to other method.
The T1 value was 2.4731 and TD value was 0.6167. The graph obtain show that it didn’t reach
the set-point also. This graph shows large overshoot.

Question 5:
Between these four graphs plotted, TL method showed the best graph shape since it gave the
least overshoot, good controller gain and reached the set point. It also had a better stability in
the control loop compared to the other methods. The performance evaluated
using time response parameters Rise Time (Tr), Settling Time (Ts) and overshoot. The general
effect of each controller parameter on closed system are shown below.

Table 6 General Effect

CL RESPONSE RISE TIME OVERSHOOT SETTLING TIME S-S ERROR

Kp Decrease Increase Small Change Decrease
Ki Decrease Increase Increase Decrease
Kd Small Change Decrease Decrease No change