Tuning of PID Controllers For Continuous Stirred Tank Reactors

See discussions, stats, and author profiles for this publication at: https://www.researchgate.
net/publication/262891481
Tuning of PID Controllers for Continuous Stirred Tank Reactors
Article in Indian Chemical Engineer · September 2012

DOI: 10.1080/00194506.2012.766005
CITATIONS READS
14 2,496
4 authors, including:
Darapureddi Krishna Ravi Padma Sree

Maharaj Vijayaram Gajapati Raj College of Engineering Andhra University
58 PUBLICATIONS 106 CITATIONS 50 PUBLICATIONS 753 CITATIONS
SEE PROFILE SEE PROFILE
Some of the authors of this publication are also working on these related projects:
Process optimization studies on removal of heavy metals from waste water using low cost adsorbents from DST under WTI with an amount 26.58
Lakhs (Ongoing). View project
All content following this page was uploaded by Darapureddi Krishna on 28 November 2015.
The user has requested enhancement of the downloaded file.

This article was downloaded by: [Indian Institute of Technology - Kharagpur]
On: 26 April 2013, At: 03:12
Publisher: Taylor & Francis
Informa Ltd Registered in England and Wales Registered Number: 1072954 Registered office:
Mortimer House, 37-41 Mortimer Street, London W1T 3JH, UK
Indian Chemical Engineer

Publication details, including instructions for authors and subscription
information:
http://www.tandfonline.com/loi/tice20
Tuning of PID Controllers for Continuous

Stirred Tank Reactors
a b b b
D. Krishna , K. Suryanarayana , G. Aparna & R. Padma Sree
a
Department of Chemical Engineering, M.V.G.R. College of Engineering,
Vizianagaram, 535005, India
b
Department of Chemical Engineering, Andhra University College of
Engineering (A), Visakhapatnam, 530 003, India
Version of record first published: 13 Feb 2013.
To cite this article: D. Krishna , K. Suryanarayana , G. Aparna & R. Padma Sree (2012): Tuning of PID
Controllers for Continuous Stirred Tank Reactors, Indian Chemical Engineer, 54:3, 157-179
To link to this article: http://dx.doi.org/10.1080/00194506.2012.766005
PLEASE SCROLL DOWN FOR ARTICLE
Full terms and conditions of use: http://www.tandfonline.com/page/terms-and-conditions
This article may be used for research, teaching, and private study purposes. Any substantial
or systematic reproduction, redistribution, reselling, loan, sub-licensing, systematic supply, or
distribution in any form to anyone is expressly forbidden.
The publisher does not give any warranty express or implied or make any representation that
the contents will be complete or accurate or up to date. The accuracy of any instructions,
formulae, and drug doses should be independently verified with primary sources. The
publisher shall not be liable for any loss, actions, claims, proceedings, demand, or costs or
damages whatsoever or howsoever caused arising directly or indirectly in connection with or
arising out of the use of this material.
INDIAN CHEMICAL ENGINEER – 2012 Indian Institute of Chemical Engineers
Vol. 54 No. 3 September 2012, pp. 157179
Print ISSN: 0019-4506, Online ISSN: 0975-007X, http://dx.doi.org/10.1080/00194506.2012.766005
Tuning of PID Controllers for Continuous

Downloaded by [Indian Institute of Technology - Kharagpur] at 03:12 26 April 2013
Stirred Tank Reactors
D. Krishna1, K. Suryanarayana2, G. Aparna2 & R. Padma Sree2*

1
Department of Chemical Engineering, M.V.G.R. College of Engineering, Vizianagaram,
535005, India
2
Department of Chemical Engineering, Andhra University College of Engineering (A),
Visakhapatnam,
530 003, India
Section A
Abstract: Tuning of proportional integral and derivative (PID) controllers for stable SOPTDZ
(second order plus time delay with a zero) systems is proposed. The PID controller was
designed based on internal model control and stability analysis principles. The proposed
controllers were applied to the various stable transfer function models of jacketed CSTR,
isothermal CSTR carrying out Van de Vusse reaction and a bioreactor. Simulation results on
various transfer function models and on non-linear model equations of jacketed CSTR carrying
first-order exothermic reaction, isothermal constant volume CSTR carrying out Van de Vusse
reaction and a bioreactor are given to show the effectiveness of the proposed PID controllers.
The performance under model uncertainty is also studied considering perturbation in one
parameter at a time. The performance of the proposed controllers is compared with the
controller designed by direct synthesis method.
Keywords: Internal model controller, Stability analysis method, Direct synthesis method,
CSTR, Bioreactor, PID controller, Second order plus time delay system with a zero (SOPTDZ)
Introduction
Proportional integral and derivative (PID) controllers give satisfactory performance for
many control processes. Owing to their simplicity and usefulness, the PID controller has
become a powerful solution to control a large number of industrial processes.
Performance of the control systems is complicated by the numerator dynamics
(presence of a zero) of the process. Several processes exhibit the second-order plus
*Author for Correspondence. E-mail: padvan@rediffmail.com

158 D. KRISHNA ET AL.
time delay system with a zero (SOPTDZ) transfer function model. Examples of such processes are the
transfer function model relating concentration of the product to the feed flow rate in isothermal
Continuous Stirred Tank Reactor (CSTR) carrying out the Van de Vusse reaction [1], the transfer
function model relating the reactor temperature to jacket temperature in jacketed CSTR [1], the
transfer function model relating concentration of the cells to the dilution rate in bioreactors [2] and the
transfer function relating the temperature of an incinerator to the inlet load rate in a municipal waste
incinerator [3]. Many recycle processes where energy and mass recycle take place are represented by
the SOPTDZ transfer function model [4].
Many methods of designing PID controllers are available in the literature. There are methods
based on stability analysis [510], extension of the Halmann method [11], modified gain and phase
margin method [12], optimization method using artificial neural networks [13], gain and phase
margin methods [14], extension of smith predictor using modern H control theory [10], simple
method [15], etc.
Chien et al. [16] have proposed a direct synthesis method of designing PID controllers for
SOPTDZ systems. They have equated the integral time and derivative time of the series type of
controller to two time constants of the system and the resulting closed loop characteristic equation is
compared with the desired characteristic equation to calculate the controller gain. In the present
work, the performance of the proposed controllers is compared with the direct synthesis method
[16].
In the present work, a simple method for designing PID controllers based on internal model
control principles and stability analysis is proposed. The PID tuning parameters are given as a
function of process model parameters. Simulation results on various transfer function models and
non-linear models are given to show the efficiency of the proposed controllers.
Proposed Methods
IMC Method
The process transfer function for SOPTD with a zero is given by
Kp ð1 psÞeLs
Gp ðsÞ ¼ (1)
a1 s2 þ a2 s þ 1
In Equation (1), the sign represents negative zero and the sign represents positive zero. The
process transfer function is split into two parts
Gp ðsÞ ¼ Gpþ ðsÞGp ðsÞ (2)
where Gpþ ðsÞ is the invertible part and Gp ðsÞ is the non-invertible part.
The IMC controller for this system is given as:
1
Q ¼ ½Gpþ ðsÞ :f (3)
where f is IMC filter and is given by

1
f ¼ n
ð1 þ ksÞ
INDIAN CHEMICAL ENGINEER Vol. 54 No. 3 September 2012

Tuning of PID Controllers 159
The order (n) of the denominator is selected in such a way to make controller realizable
(the controller numerator order should be less than or equal to that of the order of denominator).
The variable l is the Internal Model Control (IMC) filter time constant, which is a tuning parameter,
and l is selected by trial and error procedure in such a way that the maximum magnitude of
complementary sensitivity function is between 1 and 1.5 [17]. The equivalent PID controller is
obtained using the following equation:
Q
Gc ðsÞ ¼ (4)
ð1 Gp QÞ
Equation (4) can be rearranged into the following form:
1 ð1 þ a0 sÞ
Gc ðsÞ ¼ Kc ½1 þ þ sd s (5)
sI s ða1 s2 þ a2 s þ 1Þ
which is the PID controller with a lead lag filter.

Using the above procedure, the PID controller parameters for the SOPTD transfer function
model with a positive zero and a negative zero are derived and given in Table 1.
Table 1. PID parameters for SOPTDZ systems by IMC method
Process transfer function Kp ð1 þ psÞeLs Kp ð1 psÞeLs

model, Gp a1 s2 þ a2 s þ 1 a1 s2 þ a2 s þ 1
Gp , invertible part Kp ð1 þ psÞ Kp
a1 s2 þ a2 s þ 1 ða1 s2 þ a2 s þ 1Þ
IMC filter, f 1 1
ðks þ 1Þ ðks þ 1Þ2
ða1 s2 þ a2 s þ 1Þ ða1 s2 þ a2 s þ 1Þ
IMC controller,
Q ¼ ½Gpþ ðsÞ1 :f
Kp ð1 þ psÞðks þ 1Þ Kp ðks þ 1Þ2
Time delay 1 Ls 1 0:5Ls
1 þ 0:5Ls
approximation, e Ls
Q ða1 s2 þ a2 s þ 1Þ ð1 þ 0:5LsÞða1 s2 þ a2 s þ 1Þ
Gc ðsÞ ¼ ð1G p QÞ Kp ð1 þ psÞ½ðks þ 1Þ ð1 LsÞ Kp ½ðks þ 1Þ2 ð1 þ 0:5LsÞ ð1 psÞð1 0:5LsÞ
ð1 þ a0 sÞ
Equivalent PID Kc ½1 þ s1s þ sD s 1 Kc ½1 þ s1s þ sd s
I ða2 s þ 1Þ I ða1 s2 þ a2 s þ 1Þ
controller, Gc(s)
a2 a2
Kc
Kp ðk þ LÞ Kp ð2k þ L þ pÞ
tI a2 a2
a1 a1
tD a2 a2
a0 * 0.5L
a1 * 0:5Lk2
ð2k þ L þ pÞ
k2 þ Lk 0:5Lp
a2 p
ð2k þ L þ pÞ

Stability Analysis Method

The transfer function model of SOPTDZ system in Equation (1) is considered. The transfer function
of the PID controller is given by:
!
1 1
Gc ðsÞ ¼ Kc 1 þ þ sD s (6)
sI s as þ 1
The combined transfer function model of the process and the controller is given by:
!
Kp ð1 psÞeLs 1 1
Gc ðsÞGp ðsÞ ¼ Kc 1 þ þ sD s (7)
a1 s2 þ a2 s þ 1 sI s as þ 1
tI, tD and a parameters are selected in such a way that the numerator of PID controller is equal to
the denominator of the process transfer function and a is set equal to p if the transfer function is
SOPTD with negative zero, otherwise it is made equal to zero.
The phase angle criterion for GpGc is written and solved for vc using fsolve of Matlab. Using
amplitude criteria, Kc,max is obtained. From Kc,max, the design value of Kc is calculated as:
Kc;max
Kc ¼ (8)
Gm
Table 2. PID parameters for SOPTDZ systems by stability analysis method
Process transfer function, Gp Kp ð1 þ psÞeLs Kp ð1 psÞeLs

a1 s2 þ a2 s þ 1 a1 s2 þ a2 s þ 1

PID controller, Gc Kc 1 þ s1s þ sD s as 1þ 1 Kc 1 þ s1s þ sD s as 1þ 1
I I
tI a2 a2
a1 a1
tD a2 a2
a p 0
Kc Kp eLs Kc Kp ð1 psÞeLs
Gc(s)Gp(s) sI s sI s
Phase angle criterion p 2 Lxc þ p ¼ 0 p 2 Lx 1
c tan ðpxc Þ þ p ¼ 0
xc sI px c sI
ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
Kc,max Kp Kp 1 þ p2 x2c
Kc;max Kc;max
Kc Gm Gm
Gm is selected between 1.5 to 4 Gm is selected between 1.5 to 4
tI a2 a2
a1 a1
tD a2 a2
a p 0

Table 3. PID settings for different methods;

ð1 þ a0 sÞ
GC ¼ Kc 1 þ s1s þ sD s
I ð1 þ a2 s þ a1 s2 Þ
Transfer function model Controller Kc tI tD a0 a1 a2
Case study 1 IMC-PID 7.2502 0.9626 0.2902 * * 0.1252

SA-PID 5.9393 0.9626 0.2902 * * 0.1252
DS-PID 6.6379 0.9626 0.2902 * * 0.1251

Case study 2 IMC-PID 6.6833 0.2437 0.3738 * * 0.1000
SA-PID 4.4992 0.2437 0.3738 * * 0.1000
DS-PID 4.1809 0.2438 0.3741 * * 0.1000
Case study 3 IMC-PID 0.5882 2 0.5 0.1 0.0188 0.1882
SA-PID 0.5879 2 0.5 * * *
DS-PID 0.29 1.0 1.0 * * *
Case study 4 IMC-PID 1.4685 0.8627 0.2154 0.05 0.0038 0.0850
SA-PID 1.4685 0.8627 0.2154 * * *-
DS-PID 1.4588 0.8627 0.2154 * * 0.0415
SA-PID 3.1944 5.7382 0.5514 * * *
DS-PID 3.3725 5.7382 0.5514 *- * 0.0618
SA-PID 8.386 9.0155 0.2443 * * *
DS-PID 8.8075 9.0155 0.2443 * * 0.0251
where Gm is gain margin, which is the tuning parameter. For this transfer function model, Chien
et al. [16] selected tI and tD in a similar manner, but the selection of Kc was based on the desired
closed loop characteristic equation and this leads to a value of Kc which is a function of the closed
loop time constant. The closed loop time constant is derived as the function of process parameters.
In the present method, Gm is used as tuning parameter which is selected between 1.5 and 4 in such
way that a good performance is obtained.
Using the above procedure, the PID controller parameters for the SOPTD transfer function
model with a positive zero and a negative zero are derived and given in Table 2.
Simulation Results
An IMC-PID controller (PID controller designed by the proposed IMC method) and an SA-PID
controller (PID controller designed by the proposed stability analysis method) were applied to
various CSTR transfer function models and non-linear models to show the efficiency of the
proposed controllers. The PID controller parameters designed by the proposed methods and direct
synthesis method for all the case studies are reported in Table 3. The performance of the
proposed controllers in terms of integral time weighted absolute error (ITAE) was compared with
the DS-PID controller (PID controller designed by direct synthesis method [16]), and is reported
in Table 4.

Table 4. Performance comparison in terms of ITAE for linear models
Performance, ITAE
Transfer function model Controller Servo problem Regulatory problem
Case study 1 IMC-PID 0.3662 0.1085

SA-PID 0.3697 0.1314
DS-PID 0.3611 0.1323

SA-PID 0.4141 0.0725
DS-PID 0.4165 0.0785
SA-PID 6.2447 19.552
DS-PID 24.658 39.803
SA-PID 0.7933 1.1267
DS-PID 0.8967 1.1136
SA-PID 20.969 22.817
DS-PID 20.751 21.058
SA-PID 290.72 53.683
DS-PID 219.72 52.494
Case Study 1
Consider a jacketed CSTR carrying a first-order irreversible exothermic reaction. The heat of
reaction is removed by a coolant in the jacket to maintain the temperature of reaction. The model
equations of jacketed CSTR without jacket dynamics [1] are:
Table 5. Parameter values of Jacketed CSTR for case study 1 [1]
Feed concentration, CAf 0.132 lbmol/ft3

Feed temperature, Tf 519.678R
Reactor volume (nominal) 100 ft3
Heat transfer area 88 ft2
Diameter 4 ft
Operating volume 85 ft3
Operating flow rate 340 ft3/h
Overall heat transfer coefficient, U 75 Btu/h ft2 8F
Activation energy, Ea 32400 Btu/lb mol
Frequency factor, K0 16.96 1012/h
Heat of reaction, DH 39000 Btu/lb mol
Product of density and specific heat, r Cp 53.25Btu/ft3 8F
Ideal gas constant, R 1.987 Btu/lbmol 8R

1.4
1.2
Response 1
0.8
IMC-PID
0.6 SA-PID
DS-PID
0.4
0.2
0
0 1 2 3 4 5 6 7 8 9 10
Time
Fig. 1. Servo and regulatory performance of jacketed CSTR for linear model (case study 1).

dCA F Ea
¼ ðCAf CA Þ K0 exp CA (9)
dt V RT
!
dT F DH Ea UA
¼ ðTf T Þ þ K0 exp CA ðT Tj Þ (10)
dt V qCp RT V qCp
The above non-linear model equations are linearized around a stable operating point, CA 0.0644
lb mol/ft3, T560.778R and Tj 539.678R. The values of the process parameters are given in
Table 5. The process transfer function relating the reactor temperature to the jacket temperature
along with a measurement delay of 0.02 h is given by
T ðsÞ 1:458s þ 11:65

¼ e0:02s (11)
TJ ðsÞ s2 þ 3:446s þ 3:58
The servo and regulatory response with these controllers is shown in Fig. 1. The servo performance
of the three controllers was almost the same, but the regulatory response of the proposed controllers
was better than that of the DS-PID controller [16]. The simulation of non-linear model equations
(Equations (9) and (10)) of the jacketed CSTR with the proposed IMC-PID, SA-PID and DS-PID
controllers [16] for servo response (change in reactor temperature 560.77562.778R) is shown in
Fig. 2. The regulatory response of jacketed CSTR for the change in feed temperature from 519.67
to 521.678R, change in feed concentration from 0.132 to 0.134 lb mol/ft3, change in jacket
temperature from 539.67 to 541.678R and change in ratio F/V from 4 to 4.2/h are shown in

563.5
IMC-PID
SA-PID
DS-PID
563
Response 562.5
562
561.5
561
560.5
0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5
Time
Fig. 2. Servo performance of jacketed CSTR (T560.77562.778R) for non-linear model (case study 1).
Figs 36, respectively. The proposed controllers performed better than the DS-PID controller [16]
(refer to Table 6).
The robustness of the proposed IMC-PID, SA-PID and DS-PID controllers [16] was studied for
the model uncertainties considering 920% perturbation in process parameters (DH, K0, U)
561.4
IMC-PID
561.3
SA-PID
DS-PID
561.2
Response
561.1
561
560.9
560.8
560.7
0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5
Time
Fig. 3. Regulatory performance of jacketed CSTR for change in feed temperature (Tf 519.67521.678R) for
non-linear model (case study 1).

560.9
DS-PID
560.85 IMC-PID
SA-PID
560.8
560.75
Response
560.7
560.65
560.6
560.55
560.5
0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5
Time
Fig. 4. Regulatory performance of jacketed CSTR for change in feed concentration (CAf 0.1320.134 lbmol/
ft3) for non-linear model (case study 1).
considering one parameter at a time for regulatory problem (TJ). The performance is reported in
terms of ITAE values in Table 7. The controllers were designed based on nominal model parameters
and applied to the process with perturbed parameters. The robustness of the proposed IMC-PID
controller was better for model uncertainty considering one parameter at a time when compared
with SA-PID and DS-PID controllers [16].
Case Study 2
Consider a jacketed CSTR carrying a first-order irreversible exothermic reaction. The heat of
reaction is removed by a coolant in the jacket to maintain the temperature of reaction. The model
equations of jacketed CSTR without jacket dynamics [18] are given by:

dCA q Ea
¼ ðCA0 CA Þ K0 CA exp (12)
dt V RT
! " !#
dT q ðDHÞK0 CA Ea qC CPC qC hA
¼ ðT0 T Þ þ exp þ 1 exp ðTC0 T Þ
dt V qCp RT qCP V qC qCp
(13)
The above non-linear model equations were linearized around stable operating point CA 0.1 mol/l
and T 438.5417 K. The values of the parameters are given in Table 8. The transfer function

560.9
IMC-PID
SA-PID
560.85 DS-PID
560.8
Response
560.75
560.7
560.65
560.6
0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5
Time
Fig. 5. Regulatory performance of jacketed CSTR for change in jacket temperature (Tj 539.6342
541.63428R) for non-linear model (case study 1).
relating the reactor temperature to the coolant flow rate along with a measurement time delay of
0.0302 h is given by
T ðsÞ 0:8775s 8:774 0:0302s

¼ e (14)
qC ðsÞ s2 þ 2:674s þ 10:97
The servo and regulatory responses of IMC-PID, SA-PID and DS-PID controllers [16] for a linear
model are shown in Fig. 7. The performance of all controllers was almost the same for both servo
and regulatory problems (refer to Table 4). The simulation of non-linear model equations
(Equations (12) and (13)) of jacketed CSTR with IMC-PID, SA-PID and DS-PID controllers [16]
for the servo problem (change in reactor temperature from 438.541 to 440 K) is shown in Fig. 8.
The regulatory problem of the jacketed CSTR for a change in feed temperature from 350 to 355 K is
shown in Fig. 9. The regulatory problem of the jacketed CSTR for a change in initial concentration
from 1 to 1.1 mol/l, change in feed flow rate from 100 to 110 l/mol, change in coolant temperature
from 350 to 355 K and change in coolant flow rate from 103.41 to 105 l/min was also studied. For
both servo and regulatory problems, the IMC-PID controller gave a better response compared with

560.9
DS-PID
560.8 IMC-PID
SA-PID
560.7
560.6
Response
560.5
560.4
560.3
560.2
560.1
0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5
Time
Fig. 6. Regulatory performance of jacketed CSTR for change in feed rate to volume of reactor (F/V 44.2/
min) for non-linear model (case study 1).
the SA-PID and DS-PID controllers [16] (refer to Table 6). The proposed controllers are robust
(refer to Table 7) for 920% uncertainty in DH, K0 and hA for the regulatory problem (qC).
Case Study 3
The following transfer function model was considered [12]:
ð1 1:6sÞe0:2s
Gp ðsÞ ¼ (15)
s2 þ 2s þ 1
The servo and regulatory response of the process transfer function model with IMC-PID, SA-PID
and DS-PID controllers [16] is shown in Fig. 10. The performances of the servo and regulatory
problem of IMC-PID and SA-PID controllers were better than those of the DS-PID controller [16]
(refer to Table 4).
Case Study 4
Consider a continuous stirred tank reactor where in the following isothermal seriesparallel (Van de
Vusse) reactions takes place [1]. Here the product B is the desired one:
K1 K2
A ! B ! C
K3
2A ! D

INDIAN CHEMICAL ENGINEER
168
D. KRISHNA
Table 6. Performance comparison in terms of ITAE for non-linear models
Controller Servo problem Regulatory problem
ET AL.
Case study 1 T560.77562.770 R Tf 519.67521.670 R CAf 0.1320.134 Tj 539.63541.630 R F/V 44.2 min 1
lbmole/ft3
IMC-PID 0.1630 0.1817 0.0372 0.0304 0.1759
SA-PID 0.1659 0.1981 0.0404 0.0331 0.1917
DS-PID 0.1743 0.2167 0.0452 0.0366 0.2092
Case study 2 T438.54440 K Tf 350355 K CAf 11.1 mol/l q 100110 l/min Tc0 350355 K
IMC-PID 0.5899 0.2722 1.0097 0.4973 0.2985
SA-PID 0.6122 0.4095 1.4850 0.7428 0.4493
DS-PID 0.6161 0.4431 1.5996 0.8022 0.4858
Case study 4 CB from 1.117 F/V from 0.5714 CAf from 10
to 1.2 mol/l) to 0.6/min to 12 mol/l
IMC-PID 0.087 0.0320 0.1274
SA-PID 0.0922 0.0325 0.1309
DS-PID 0.0881 0.0322 0.1280
Case study 5 X from 0.3018 D from 0.15
to 0.308 g/l to 0.16/h
IMC-PID 0.3715 0.2888
Vol. 54 No. 3 September 2012
SA-PID 0.5306 0.2438

DS-PID Unstable 0.2264
Case study 6 X from 0.379 D from 0.11
to 0.36 g/l to 0.115/h)
IMC-PID 0.9251 0.4560
SA-PID 0.8956 0.4511
DS-PID 0.8441 Unstable
Table 7. Regulatory performance comparison under model uncertainty in terms of ITAE for non-linear model
Case study 1 Controller 20% DH 20% DH 20% K0 20% K0 20% U 20% U

IMC-PID 2.8989 3.611 1.1273 1.7152 0.5239 0.4889
SA-PID 3.1493 3.975 1.2318 1.8778 0.5739 0.5328
DS-PID 3.2106 3.911 1.2142 1.8797 0.5651 0.5422
Case study 2 Controller 20% DH 20% DH 20% K0 20% K0 20% hA 20% hA
IMC-PID 1.7479 2.312 0.4968 0.8601 0.0043 0.0172
SA-PID 2.4933 3.767 0.7217 1.4252 0.0065 0.0261

DS-PID 2.6716 4.147 0.7735 1.5825 0.0071 0.0283
Case study 5 Controller mmax 10% Y10%
IMC-PID 0.0635 0.765
SA-PID 0.0514 0.661
DS-PID 0.0485 0.609
Case study 6 Controller mmax 10% Y10%
IMC-PID 0.1156 0.705
SA-PID 0.1103 0.720
DS-PID 0.1000 0.680
Table 8. Parameters values of jacketed CSTR for case study 2 [18]
Feed concentration (CA0) 1.0 mol/l

Feed temperature (T0) 350 K
Coolant temperature (TC0) 350 K
Steady-state concentration (CAS) 0.1 mol/l
Steady-state temperature (TS) 438.5417 K
Feed flow rate (q) 100 l/min
Coolant flow rate (qc) 103.41 l/min
Volume (V) 100 l
Heat transfer coefficient (hA) 7 105 cal/min K
Frequency factor (K0) 7.2 1010/min
Activation energy (E/R) 104 K
Heat of reaction ( DH) 2 105 cal/mol
Density, r 103 g/l
Specific heat, CP 1 cal/g K
The non-linear model equations for the species A and B are:
dCA F
¼ ðCAf CA Þ K1 CA K3 CA2 (16)
dt V
dCB F
¼ CB þ K1 CA K2 CB (17)
dt V
The above non-linear equations were linearized around stable operating point CA 3 g mol/l and
CB 1.1170 g mol/l. The process parameters are given in Table 9. The transfer function relating the

1.4
IMC-PID
1.2 SA-PID
DS-PID
Response 1
0.8
0.6
0.4
0.2
0
0 5 10 15 20
Time
Fig. 7. Servo and regulatory performance of jacketed CSTR for linear model (case study 2).
desired concentration (CB) to the ratio flow rate and volume of the reactor along with a
measurement time delay of 0.1 min is given by
CB ðsÞ 1:117s þ 3:1472

¼ e0:1s (18)
F=V ðsÞ s2 þ 4:6429s þ 5:3821
The servo and regulatory problem of the process transfer function model with IMC-PID,
SA-PID and DS-PID controller is shown in Fig. 11. The performances of servo and regulatory
problems were almost the same for all the controllers (refer to Table 4). The simulation of non-
linear model equations with the IMC-PID, SA-PID and DS-PID controllers [16] for the servo
problem (change in product concentration CB from 1.117 to 1.2 g mol/l) is shown in Fig. 12. The
regulatory problem of constant volume CSTR for the change in flow rate (F/V) from 0.5714 to
0.6/min is shown in Fig. 13. For the servo problem, all the controllers gave similar responses, but
for the regulatory problem the SA-PID controller was superior compared with the IMC-PID and
DS-PID controller (refer to Table 6).
Case Study 5
The mathematical model of a fermenter is given by the following equations [2] where the microbial
growth is assumed to follow Monod kinetics:
dX
¼ ðl Kd ÞX Xu (19)
dt

dS l
¼ m X ðS Sf Þu (20)
dt Y

441.5
IMC-PID
SA-PID
441 DS-PID
Reactor temperature
(Degress Rankine) 440.5
440
439.5
439
438.5
0 1 2 3 4 5 6 7
Time (minutes)
Fig. 8. Servo performance of jacketed CSTR for change in reactor temperature (T438.54440 K) for non-
linear model (case study 2).
439.4
IMC-PID
SA-PID
439.2 DS-PID
Reactor Temperature
(Degrees Rankine)
439
438.8
438.6
438.4
438.2
0 1 2 3 4 5 6 7
Time (Minutes)
Fig. 9. Regulatory performance of jacketed CSTR for change in feed temperature (T0 350355 K) for non-
lmax S
l¼ (21)
km þ S

2.5
SA-PID
2 IMC-PID
DS-PID
1.5
Response
1
0.5
-0.5
0 5 10 15 20 25 30 35 40 45 50
Time
Fig. 10. Servo and regulatory performance (case study 3).
Table 9. Parameter values of CSTR for case study 4 [1]
Parameter Values of parameter

F
V
0.5714/min
5
K1 6
/min
5
K2 3
/min
K3 1
6
/min1
CAf 10 gmol/l
where X is cell concentration and S is substrate concentration. The above non-linear model
equations are linearized around stable operating point X 0.3018 g/l and S0.0452 g/l. The values
of parameters are given in Table 10. The transfer function model relating cell concentration and
dilution rate along with a measurement time delay of 0.3 h is given by
X ðsÞ 0:2803ð1 3:4063sÞe0:3s
¼ (22)
uðsÞ ð3:1638s2 þ 5:7382s þ 1Þ
The servo and regulatory problems of the process transfer function model with IMC-PID, SA-PID
and DS-PID controllers [16] are shown in Fig. 14. The servo and regulatory performances of IMC-
PID and SA-PID controllers were better than that of the DS-PID controller [16] (refer to Table 4).
The simulation of non-linear model equations with IMC-PID, SA-PID and DS-PID controllers [16]
for the servo problem (change in cell concentration X from 0.3018 to 0.308 g/l) is shown in Fig. 15.
For the servo problem, the IMC-PID and SA-PID controllers stabilized the bioreactor but the
DS-PID controller [16] could not stabilize it (refer to Table 6). The regulatory problem for the

1.6
IMC- PID
1.4
SA- PID
1.2 DS- PID
0.8
Response
0.6
0.4
0.2
-0.2
-0.4
0 2 4 6 8 10 12 14 16 18 20
Time
Fig. 11. Servo and regulatory performance of constant volume CSTR carrying out Van de Vusse reaction for
1.22
1.2
IMC -PID
1.18
SA -PID
DS -PID
Response
1.16
1.14
1.12
1.1
0 5 10 15
Time
Fig. 12. Servo performance of constant volume CSTR carrying out Van de Vusse reaction for change in
product concentration CB from 1.117 to 1.2 gmol/l for non-linear model (case study 4).
change in dilution rate D from 0.15 to 0.16/h is shown in Fig. 16. The performance and the
robustness of the proposed controllers for 10% uncertainty in mmax and Y (regulatory problem) in
terms of ITAE are given in Table 7.

1.132
1.13
1.128
IMC -PID
1.126 SA -PID
DS -PID
1.124
Response
1.122
1.12
1.118
1.116
1.114
1.112
0 1 2 3 4 5 6 7 8 9 10
Time
Fig. 13. Servo and regulatory performance of constant volume CSTR carrying out Van de Vusse reaction for
change in ratio of feed flow rate to volume of reactor (F/V0.57140.6/min) for non-linear model
(case study 4).
Table 10. Process parameters of the bioreactor [2]
Parameter Case study-5 Case study-6
Specific growth rate, mmax 0.4/h 0.4/hr

Yield coefficient, Y 0.4 g/g 0.4 g/g
Maintenance, m 0 0.01/h
Monod constant, km 0.05 g/l 0.05 g/l
Substrate feed concentration, Sf 1.001 g/l 1.001 g/l
Steady state cell concentration, Xs 0.3018 g/l 0.379 g/l
Steady-state substrate concentration, Ss 0.0452 g/l 0.019 g/l
Dilution rate, u 0.15/h 0.11/h
Decay rate, kd 0.04/h 0
Case Study 6
The non-linear model equations (Equations (19) and (20)) are linearized around stable operating
point X 0.379 g/l and S 0.019 g/l. The values of parameters are given in Table 10. The transfer
function model relating cell concentration and dilution rate along with a measurement time delay of
0.3 h is given by
X ðsÞ 0:0291ð1 28:7121sÞe0:3s

¼ (23)
uðsÞ ð2:2024s2 þ 9:0155s þ 1Þ

1.5
0.5 IMC-PID
Response
SA-PID
DS-PID
0
-0.5
-1
0 5 10 15 20 25 30 35 40 45 50
Time
Fig. 14. Servo and regulatory performance of fermenter for linear model (case study).
0.314
IMC -PID
0.312
SA-PID
0.31
0.308
0.306
Response
DS-PID (Unstable) (not shown in Fig)

0.304
0.302
0.3
0.298
0.296
0.294
0 5 10 15 20 25 30 35 40 45 50
Time
Fig. 15. Servo performance of bioreactor for change in cell concentration from X0.30180.308 g/l for non-
The servo and regulatory problems of the process transfer function model with IMC-PID, SA-PID
and DS-PID controllers [16] are shown in Fig. 17. The servo and regulatory performances of IMC-
PID and SA-PID controllers were better than that of the DS-PID controller [16] (refer to Table 4).

0.305
SA-PID
IMC-PID
0.304 DS-PID
0.303
Response
0.302
0.301
0.3
0.299
0 5 10 15 20 25 30 35 40 45 50
Time
Fig. 16. Regulatory performance of bioreactor for change in dilution rate (D0.150.16 h) for non-linear
model (case study 5).
1.5
0.5
0
IMC-PID
-0.5 SA-PID
Response
DS-PID
-1
-1.5
-2
-2.5
-3
-3.5
0 10 20 30 40 50 60 70 80 90 100
Time
Fig. 17. Servo and regulatory performance of bioreactor for linear model (case study 6).
The simulations of non-linear model equations with IMC-PID, SA-PID and DS-PID controllers [16]
for the servo problem (change in cell concentration X from 0.379 to 0.36 g/l) and for the regulatory
problem (change in dilution rate D from 0.11 to 0.115/hr) are shown in Figs 18 and 19, respectively.

0.395
IMC-PID
0.39 SA-PID
DS-PID
0.385
0.38
Response
0.375
0.37
0.365
0.36
0.355
0 5 10 15 20 25 30 35 40 45 50
Time
Fig. 18. Servo performance of bioreactor for change in cell concentration (X0.3790.36 g/l) for non-linear
0.3805
IMC-PID
SA-PID
0.38
0.3795
Response
0.379
DS-PID (Unstable)(not shown in Fig)

0.3785
0.378
0.3775
0.377
0 5 10 15 20 25 30 35 40 45 50
Time
Fig. 19. Regulatory performance of bioreactor for change in dilution rate (D 0.110.115/h) for non-linear
For the regulatory response, IMC-PID and SA-PID controllers stabilized the bioreactor but the
DS-PID controller [16] could not stabilize it (refer to Table 6). The performance and the
robustness of the proposed controllers for 10% uncertainty in mmax and Y (regulatory problem)

in terms of ITAE are given in Table 7. The proposed controllers were robust for 10% variation
in mmax and Y.
Conclusions
Two PID controller design methods based on internal model control principles and the stability
analysis method for a stable second-order system with time delay and with a zero is proposed. The
zero may be a positive or a negative. The proposed IMC-PID and SA-PID controllers were applied
to control jacketed CSTR carrying out first-order irreversible exothermic reaction, an isothermal
constant volume reactor carrying out Van de Vusse reaction and a bioreactor. The performance of
the proposed IMC-PID and SA-PID controllers was compared with the DS-PID controller [16] and
the IMC-PID controller gave the best performance for both servo and regulatory problems. The
performance of the proposed IMC-PID and SA-PID controllers for uncertainty in parameters
considering one parameter at a time was studied and compared with the performance of PID
controller designed by the DS-PID controller [16]. The IMC-PID and SA-PID controllers gave
robust performance compared with the DS-PID controller [16]. The servo and regulatory
performance comparison is given in terms of ITAE for both linear and non-linear linear models.
Simulation results on non-linear model equations of jacketed CSTR carrying out irreversible first-
order chemical reaction, constant volume CSTR carrying out Van de Vusse reaction and bioreactor
show that the proposed IMC-PID and SA-PID controllers perform better than the DS-PID controller
[16].
References
1. Bequette, B.W., Process Control, Modeling, Design and Simulation, Prentice Hall India, New Delhi
(2003).
2. Kuhlmann, A. and Bogle, I.D.L., ‘‘Controllability Evaluation for Non-minimum Phase- processes with
Multiplicity’’, AICHE J., 47, pp. 26272632 (2001).
3. Rovaglio, M., Manca, D., Pazzaglia, G. and Serafini, G., ‘‘Inverse Response Compensation for the
Optimal Control of Municipal Incineration Plants; Model Synthesis and Experimental Validation’’, Comp.
Chem. Eng., 20, pp. S1461S1467 (1996).
4. Jacobsen, E.W., ‘‘On the Dynamics of Integrated Plants Non-minimum Phase Behavior’’, J. Process
Control, 9, pp. 439451 (1999).
5. Ziegler, J.G. and Nichols, N.B., ‘‘Optimum Settings for Automatic Controllers’’, Trans. ASME, 64, pp.
759765 (1942).
6. Cohen, G.H. and Coon, G.A., ‘‘Theoretical Consideration of Retarded Control’’, Trans. ASME, 75, pp.
827834 (1953).
7. Suresh Bal, P., ‘‘PID Control of Stable Second Order Plus Time Delay (SOPTD) Systems With Zero’’,
M.Tech Thesis, Andhra University, Visakhapatnam (2004).
8. Suresh Bal, P. and PadmaSree, R., ‘‘Tuning of PID Controllers for SOPTD Systems with Zero’’, Proc.
57th Annual Session of Indian Institute of Chemical Engineers (CHEMCON-2004).
9. Sravanthi, M., ‘‘Design and Performance Improvement of PID Controllers for Second Order Plus Delay
Systems With/Without Zero’’, M.Tech. Thesis, Andhra University, Visakhapatnam (2005).
10. Zhang, W., Xu, X. and Sun, Y., ‘‘Quantitative Performance Design for Inverse Response Processes’’, Ind.
Eng. Chem. Res., 39, pp. 20562061 (2000).
11. Jyothi, S.N., Aravind, S. and Chidambaram, M., ‘‘Design on PI/PID Controllers for Systems with a
Zero’’, Indian Chem. Eng., 43, pp. 288293 (2001).

12. Luyben, W.L., ‘‘Tuning PI Controller for Process with Both Inverse Response and Dead Time’’, Ind. Eng.
Chem. Res, 39, pp. 973976 (2000).
13. Huang, C.T., Lin, M.Y. and Huang, M.C., ‘‘Tuning PID Controllers for Processes with Inverse Response
using Artificial Neural Networks’’, J. Chin. Inst. Chem. Eng, 30, pp. 223232 (1999).
14. Huang, Y.J. and Wan, Y.J., ‘‘Robust PID Controller Design for Non-minimum Phase Time Delay
Systems’’, ISA Trans., 40, pp. 3129 (2001).
15. PadmaSree, R. and Chidambaram, M., ‘‘Simple Method of Tuning PI Controllers for Inverse Response
Systems’’, J. Indian Inst. Sci., 83, pp. 7385 (2003).
16. Chien, I.L., Chung, Y.C., Chen, B.S. and Chuang, C.Y., ‘‘Simple PID Controller Tuning Method for
Processes with Inverse Response Plus Dead Time or Large Overshoot Response Plus Dead Time’’, Ind.
Eng. Chem. Res., 42, pp. 44614477 (2003).
17. Seborg, D.E., Edgar, T.H. and Mellichamp, D.A., Process Dynamics and Control, 2nd edn, John Wiley
and Sons, New York (2004).
18. Pottmann, M. and Seborg, D.E., ‘‘Identification of Non-linear Process using Reciprocal Multi-quadratic
Functions’’, J. Process Control, 2, pp. 189203 (1992).
Nomenclature
Gc Control transfer function
Gp Process transfer function
Gm Gain margin
Kc Controller gain
Kc,max Ultimate controller gain
tI Integral time
Kp Process gain
L Time delay
p Numerator time constant
t Time
t Time constant
tD Derivative time
a0, a1, a2 PID Filter time constants
a1. a2 Denominator constants
vc Phase cross over frequency
l IMC filter time constant
ITAE Integral time absolute error
DS-PID Proportional integral and derivative controller designed by direct synthesis method [16]
SA-PID Proportional integral and derivative controller designed by stability analysis method
IMC-PID Proportional integral and derivative controller designed by internal model control method
View publication stats

Tuning of PID Controllers For Continuous Stirred Tank Reactors

Uploaded by

Copyright:

Available Formats

Tuning of PID Controllers For Continuous Stirred Tank Reactors

Uploaded by

Document Information

Original Description:

Original Title

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

Tuning of PID Controllers For Continuous Stirred Tank Reactors

Uploaded by

Copyright:

Available Formats

See discussions, stats, and author profiles for this publication at: https://www.researchgate.

Tuning of PID Controllers for Continuous Stirred Tank Reactors

Article in Indian Chemical Engineer · September 2012

Darapureddi Krishna Ravi Padma Sree

SEE PROFILE SEE PROFILE

The user has requested enhancement of the downloaded file.

Indian Chemical Engineer

Tuning of PID Controllers for Continuous

To link to this article: http://dx.doi.org/10.1080/00194506.2012.766005

PLEASE SCROLL DOWN FOR ARTICLE

Full terms and conditions of use: http://www.tandfonline.com/page/terms-and-conditions

Tuning of PID Controllers for Continuous

Stirred Tank Reactors

D. Krishna1, K. Suryanarayana2, G. Aparna2 & R. Padma Sree2*

*Author for Correspondence. E-mail: padvan@rediffmail.com

where f is IMC filter and is given by

INDIAN CHEMICAL ENGINEER Vol. 54 No. 3 September 2012

Equation (4) can be rearranged into the following form:

which is the PID controller with a lead lag filter.

Table 1. PID parameters for SOPTDZ systems by IMC method

Process transfer function Kp ð1 þ psÞeLs Kp ð1  psÞeLs

INDIAN CHEMICAL ENGINEER Vol. 54 No. 3 September 2012

Stability Analysis Method

Table 2. PID parameters for SOPTDZ systems by stability analysis method

Process transfer function, Gp Kp ð1 þ psÞeLs Kp ð1  psÞeLs

INDIAN CHEMICAL ENGINEER Vol. 54 No. 3 September 2012

Table 3. PID settings for different methods;

Transfer function model Controller Kc tI tD a0 a1 a2

Case study 1 IMC-PID 7.2502 0.9626 0.2902 * * 0.1252

DS-PID 6.6379 0.9626 0.2902 * * 0.1251

INDIAN CHEMICAL ENGINEER Vol. 54 No. 3 September 2012

Table 4. Performance comparison in terms of ITAE for linear models

Transfer function model Controller Servo problem Regulatory problem

Case study 1 IMC-PID 0.3662 0.1085

Case study 2 IMC-PID 0.4044 0.0481

Table 5. Parameter values of Jacketed CSTR for case study 1 [1]

Feed concentration, CAf 0.132 lbmol/ft3

INDIAN CHEMICAL ENGINEER Vol. 54 No. 3 September 2012

T ðsÞ 1:458s þ 11:65

INDIAN CHEMICAL ENGINEER Vol. 54 No. 3 September 2012

INDIAN CHEMICAL ENGINEER Vol. 54 No. 3 September 2012

INDIAN CHEMICAL ENGINEER Vol. 54 No. 3 September 2012

T ðsÞ 0:8775s  8:774 0:0302s

INDIAN CHEMICAL ENGINEER Vol. 54 No. 3 September 2012

INDIAN CHEMICAL ENGINEER Vol. 54 No. 3 September 2012

Controller Servo problem Regulatory problem

SA-PID 0.5306 0.2438

Case study 1 Controller 20% DH 20% DH 20% K0 20% K0 20% U 20% U

SA-PID 2.4933 3.767 0.7217 1.4252 0.0065 0.0261

Table 8. Parameters values of jacketed CSTR for case study 2 [18]

Feed concentration (CA0) 1.0 mol/l

The non-linear model equations for the species A and B are:

INDIAN CHEMICAL ENGINEER Vol. 54 No. 3 September 2012

CB ðsÞ 1:117s þ 3:1472

INDIAN CHEMICAL ENGINEER Vol. 54 No. 3 September 2012

INDIAN CHEMICAL ENGINEER Vol. 54 No. 3 September 2012

Fig. 10. Servo and regulatory performance (case study 3).

Table 9. Parameter values of CSTR for case study 4 [1]

Process transfer function Kp ð1 þ psÞeLs Kp ð1 psÞeLs

Process transfer function, Gp Kp ð1 þ psÞeLs Kp ð1 psÞeLs

T ðsÞ 0:8775s 8:774 0:0302s

CB ðsÞ 1:117s þ 3:1472

X ðsÞ 0:0291ð1 28:7121sÞe0:3s