Proportional-Integral Controller Retuning to Improve

Controller Performance in Formaldehyde Production

Process at PT X

Abdul Wahid1, and Dyah Kunti Surya Andari2

12
Chemical Engineering Department, Universitas Indonesia, Depok, Indonesia
a)
Corresponding author: wahid@che.ui.ac.id
b)
dyah.kunti@ui.ac.id

Abstract. A proportional – integral controller retuning is performed on formaldehyde production process at PT X.

Retuning is carried out to improve the controller performance on the process. The process model is estimated by a first
order plus dead time model (FOPDT), and then the controller parameters is tuned using the Ziegler-Nichols (PRC),
Wahid-Rudi-Victor (WRV), Cohen-Coon, autotuning, and fine tuning. The controller performance is tested using set
point (SP) tracking and disturbance rejection with integral of square error (ISE) as performance indicator. There are three
controllers that are tested, i.e., the steam flow controller (FIC – 102), air temperature controller (TIC – 101), and tank
level controller (LIC – 102). Fine tuning method gives the better control performance for FIC – 102 and TIC – 101, while
autotuning method gives the better control performance for LIC – 102 compared to the previous settings in the field.
Performance improvement for set point (SP) tracking are 81,59% (FIC – 102), 94,11% (TIC – 101), and 85,61% ( LIC –
102). While performance improvement for the disturbance rejection are 95,5% (FIC – 102), 94,53% (TIC – 101),
93,16% ( LIC – 102). Thus, retuning PI controllers work well and by using the suggested controller tuning, the better
controller performance of PT X can be achieved.

INTRODUCTION
Formaldehyde is aldehyde with the chemical formula of H2CO, which is in the form of gas or liquid. One of the
industries that produces formaldehyde in Indonesia is PT X. Over time, the performance of a plant will decrease,
causing the old controller tuning no longer appropriate to be used. Objective of this study is to improve controller
performance by changing the tuning parameter to a better one by retuning. Therefore, it is necessary to do a
controller retuning to improve the performance of the controller in formaldehyde plant of PT X , so stability will be
achieved faster.
Veronesi and Visioli [1-2] have conducted a proportional-integral (PI) and proportional-integral-derivative (PID)
controller retuning. It proved that controller performance could be improved by retuning. Wahid and Mahdi [7]
have also conducted a proportional-integral controller retuning on CO2 removal process from natural gas in PT X
Subang field. Controller retuning is done for set point (SP) tracking and disturbance rejection. The model used for
retuning is based on an open loop model testing, and results on a FOPDT model. In this study, an open loop model
testing will be conducted with the same model result, i.e. FOPDT, while the retuning method used will be searched
for the correct method. Open loop model testing is used because this study intends to obtain empirical model of the
process, then use it to do the retuning.
 METHODOLOGY

Field Data Collection

A plant visit will be undertaken to collect data required for the research, including PFD, process data, valve size,
and controller data. PFD of the plant [3] is shown in Figure 1.
The process that is used in this plant is BASF process. Formaldehyde is produced by passing methanol gas over
silver catalyst in reactor. Formaldehyde that is produced has liquid concentration of 37%. The designed
formaldehyde production of the plant is 40 m3/day. However this plant can only produce formaldehyde by 36
m3/day nowdays.

FIGURE 1. PFD of Formaldehyde Plant PT X

The tuning parameter used in PT X are in % PB (Proportional Band) and Ti (Integral Time). To do proportional-
integral controller retuning, the data needed are Kc and Ti. Therefore percent PB must be converted to Kc. The
relationship between PB and Kc [4] is given by equation 1.

100
Kc  (1)
PB

The tuning parameter of proportional-integral controller in PT X can be seen in table 1.

TABLE 1. Field Tuning Value
Controller % PB Kc Ti (min)
FIC – 102 100 1 0,2
TIC – 101 100 1 0,2
LIC – 102 100 1 0,3
 Simulation Preparation
This research is carried out with the help of a process simulator, and will run on the dynamic mode of the
simulator. Before the dynamic simulation is made, the simulation will be made on steady state mode first, until the
simulation describes the real condition in the field.
After the steady state simulation is completed and the result matches the field condition, then valves and PI
controllers will be added according to their placement on the field. Then, the valve size from the field will be
entered, so the simulation of controller describes the actual condition. After the steady state simulation, valves, and
controllers are ready, then the simulation will be switched to dynamic mode. In dynamic mode, the simulation will
run until stable condition is obtained. To achieve stability faster, the autotuner feature will be used to provide initial
tuning for the controllers.
Configuration for simulation of the plant can be seen in Figure 2, where this study focused only on FIC – 102,
TIC – 101 and LIC – 102, which have been circled in red on Figure 2. FIC – 102 controls steam flowrate that will
enter the reactor, TIC – 101 controls the air temperature, LIC – 102 controls the formaldehyde tank level.

FIGURE 2. Dynamic Simulation of Formaldehyde Plant

PRC and Empirical Models

To determine the empirical models for tuning equations, Process Reaction Curve (PRC) is made by increasing
5% of valve opening (OP). Process empirical modeling will be conducted after the PRC is obtained, and the method
used is FOPDT. A transfer function is a mathematical model that define the ratio of the output of a system to the
input of a system [5]. The transfer function of FOPDT [4] is shown by equation (2)
K p es
G p ( s)  (2)
s  1
where, Kp is process gain describing the distance of PV move; τ is time constant describing PV responds time;
and θ is dead time describing the delay occurs before the PV begins to move. The process empirical model, Kp, τ,
and θ, will be determined with method FOPDT [5] in equation (3), (4), and (5).


Kp  (3)

where, Δ is the change of PV; and δ is the change of OP.
   1,5t 63%  t 28%  (4)

where, t is the time needed for PV to reach 28% and 63% of the final stable value.

    t 63%  t 0 (5)
where, t0 is the time when model testing starts.

Controller Retuning
The methods used for this study are Ziegler – Nichols (PRC), Wahid – Rudi – Victor (WRV), Cohen – Coon
(CC), simulator autotuner, and fine tuning. The equation for PI controller retuning from ZN (PRC) [4], WRV [6],
and CC [4] are shown in Table 2. Fine tuning method is conducted by refining the best tuning from ZN, WRV, or
CC to a better tuning.

TABLE 2. Tuning equation for ZN (PRC), WRV, and CC

Method Controller Gain, Kc Imtegral Time, Ti
1
0,9   
ZN (PRC)   3,33
K  

0,0433  0,8353
WRV 1,027  10,77
K
1   
1
   0,9 
 

  
 30  3  
9  20 
CC K    12 


Controller Performance
All controllers tuning performance is tested by doing set-point tracking and disturbance rejection. SP Tracking is
conducted by reducing 8% SP for FIC – 102, and 5% SP for TIC – 101 and LIC – 102. The disturbance given for the
FIC – 102 and TIC – 101 is by reducing the steam pressure to 2,8 kg/cm, and 8% reduction of methanol flowrate for
LIC – 102. Parameter used for controller performance in this study is integral of square error (ISE) [4] shown on
equation (6). After all tuning had been tested, ISE of each and every one of the tuning will be compared. Tuning
with the lowest ISE value is the best tuning.

ISE   SPt CV t 2 dt   E t 2dt ` (6)

where, SP is set point describing the value set for the variable; CV is controlled variable describing the
current variable value; and E is error describing the value difference of SP and CV. All of those parameters are in
the function of time.

RESULT AND DISCUSSION

PRC and FOPDT Models

After performing model testing for every controller, PRCs are obtained for FIC – 102 and TIC – 10. PRC of LIC
– 102 can’t be obtained, because the controller have shown instability when valve opening change occur. PRCs of
FIC – 102 and TIC – 101 are shown in Figure 3.
 (a)

(b)

Process Variable
Valve Opening

FIGURE 3. PRC results for model testing: (a) FIC – 102 and (b) TIC – 101

From the PRCs, Δ or the change of PV and δ or the change of OP are obtained. The data that have been collected
then used to determine process empirical models based on FOPDT using equation (3), (4), and (5). FOPDT models
shown in Table 3.
TABLE 3. FOPDT Models
Controller Kp
FIC-101 2,46 0,36 0,07
TIC-101 0,54 4,66 0,17

Tuning Value
From SP tracking and disturbance rejection result, WRV method gives the best controller performance between
two other method, hence fine tuning is conducted based on WRV tuning value. Fine tuning is done for FIC – 102
and TIC – 101. Fine tuning can not be done for LIC – 102, because LIC – 102 is unstabble. Tuning values for field,
autotuner, and fine tuning are shown in Table 4.
 TABLE 4. Controller Tuning
Tuning Method Tuning FIC-102 TIC-101 LIC-102
Kc 1 1 1
Field % PB 100 100 100
Ti (min) 0,2 0,2 0,3
Kc 0,26 9 3,77
Autotuner % PB 384,61 11,11 26,52
Ti (min) 0,008 0,34 0,73
Kc 0,7 8 -
Fine Tuning % PB 142,85 12,5 -
Ti (min) 0,008 0,3 -

SP Tracking
Result of controller response for set point tracking shows that ZN (PRC), WRV, and CC tuning is worse than
current tuning, so they are not displayed on the graph. Controller responses for set point tracking with the tunings
observed are shown in Figure 4.

170
Steam Flowrate

160
(kg/h)

150
(a)
140
0 100 200 300 400 500
Time (Seconds)
sp Field Auto Fine

115
Air Temperature

110
(°C)

105
(b)
100
0 500 1000 1500 2000 2500
Time (Sconds)
SP Field Auto Fine

55
Tank Percent Level (%)

50

45 (c)
40
0 3000 6000 9000 12000 15000 18000
Time (Seconds)
SP Field Auto
FIGURE 4. SP Tracking Results : (a) FIC - 102, (b) TIC - 101, (c) LIC -102
 Figure 4 shows that the tuning which can set the controlled variable back to set point fastest when the set point is
decreased are fine tuning and autotuning. For 8% set point (SP) tracking, retuning with fine tuning gives 81,59%
improvement for FIC – 102. For 5% set point (SP) tracking, retuning with fine tuning gives 94,11% improvement
for TIC – 101, and retuning with autotuning gives 85,61% improvement for LIC – 102. ISE value of every tuning in
every controller for SP tracking is shown in Table 5.

Disturbance Rejection
Result of controller response for disturbance rejection shows that ZN (PRC), WRV, and CC tuning is worse than
current tuning, so they are not displayed on the graph neither. Controller responses for disturbance rejection with the
tunings observed are shown in Figure 5.

162
Steam Flowrate (kg/h)

160
158
(a)
156
154
0 100 200 300 400 500
Time (Seconds)
SP Field Auto Fine

125
Air Temperature (°C)

120
115
110
105 (b)
100
0 500 1000 1500 2000 2500
Time (Seconds)
SP Field Auto Fine
Tank Percent Level (%)

50.25
50.15
50.05
49.95 (c)
49.85
49.75
0 2000 4000 6000 8000 10000
Time (Seconds)
SP Field Auto

FIGURE 5. Disturbance Rejection Results: (a) FIC – 102, (b) TIC – 101, (c)
LIC – 102
 Figure 5 shows that the tuning which can set the controlled variable back to set point fastest when disturbance
occur are fine tuning and autotuning. By reducing the steam pressure to 2,8 kg/cm, retuning with fine tuning gives
95,5% improvement for FIC – 102 and gives 94,53% improvement for TIC – 101. By giving 8% reduction of
methanol flowrate, retuning with autotuning gives 93,16% improvement for LIC – 102. ISE value of every tuning in
every controller for SP tracking is shown in Table 6.

TABLE 5. ISE Value for SP Tracking

Controllers Field Autotuner Fine Tuning
FIC 40,67 20,13 7,48
TIC 139,24 16,64 8,19
LIC 856,55 115,49 -

TABLE 6. ISE Value for Disturbance Rejection

Controllers Field Autotuner Fine Tuning

FIC 19,44 1,80 0,86

TIC 627,72 113,19 34,32

LIC 0,60 0,04 -

CONCLUSION
The result of this research shows that fine tuning method give the better control performance for FIC – 102 and
TIC – 101, while autotuning method gives the better control performance for LIC – 102 compared to the previous
settings in the field. For set point (SP) tracking, ISE values that is obtained by fine tuning are 7,48 for FIC – 102 and
8,19 for TIC – 101. Fine tuning method gives 81,59% improvement for FIC – 102 and 94,11% improvement for TIC
– 101. ISE value that is obtained by autotuning is 115,49 and gives 85,61% improvement for LIC – 102. For
disturbance rejection, ISE value that is obtained by fine tuning are 0,86 for FIC – 102 and 34,32 for TIC – 101. Fine
tuning gives 95,5% improvement for FIC – 102 and 94,53% for TIC – 101. ISE value that is obtained by autotuning
is 0,04 and gives 93,16 % improvement for LIC – 102. Thus, by using the suggested controller tuning, the better
controller performance of PT X can be achieved.

ACKNOWLEDGMENTS
We express our gratitude to the Universitas Indonesia which has funded this research through the scheme of
Hibah Publikasi Internasional Terindeks untuk Tugas Akhir Mahasiswa (PITTA B)
No.0690/UN2.R3.1/HKP.05.00/2019.

REFERENCES
1. M. Veronesi, A. Visioli, Ind. And Eng. Chem. Res, 48,2616-2623 (2009)
2. M. Veronesi, A. Visioli, Proceeding of IFAC Conference on Advances of PID Controll, 28-30 (2012)
3. PT X, Process Flow Diagram and Operasional Data for Main Process Production of Formaldehyde (1985)
4. A. C. Smith. A.B. Corripio. Principles and Preactice of Automatic Proces Control (1997)
5. T. E. Marlin, Process Control (2nd ed) (2000)
6. A. Wahid, R. Gunawan. Proceeding of National Seminar of Chemical Process Technology VII 2005, 1-9
(2005)
7. A. Wahid, Mahdi. Retuning PI controller to improve the control performance in CO2 removal process, subang
field. E3S Web of Conference, 67, 01027 (2018)