International Journal of Heat and Mass Transfer 55 (2012) 68736880

Contents lists available at SciVerse ScienceDirect

International Journal of Heat and Mass Transfer

journal homepage: www.elsevier.com/locate/ijhmt

Dynamic modeling and control of plate heat exchanger

Salam K. Al-Dawery a,, Ayham M. Alrahawi a, Khalid M. Al-Zobai b
a
Department of Chemical Engineering, University of Nizwa, Birkat Al Mouz, PC 616, Sultanate of Oman, Oman
b
Department of Chemical Engineering, University of Baghdad, Al-Jadiryah, Baghdad, Iraq

a r t i c l e i n f o a b s t r a c t

Article history: A general model has been developed to suggest the transient responses of a plate heat exchanger. The
Received 22 April 2012 predicted and experimental step responses of the system have been analyzed using the frequency
Received in revised form 27 June 2012 response analysis. The results indicate that the system is represented by a rst order lag and dead time.
Accepted 29 June 2012
A closed t between the simulated and experimental data has been obtained.
Available online 23 July 2012
To verify the presented model, temperature control has been applied on the plate heat exchanger using
both conventional and fuzzy logic controllers. Results show that the performance of the fuzzy logic con-
Keywords:
troller produces transient responses with less settling time and less oscillatory behavior compared to that
Plate heat exchanger
Dynamic modeling
under the conventional controller. Comparisons between simulated and experimental responses indicate
Transient response that the developed model is capable of predicting the transient responses of the plate heat exchanger,
Fuzzy controller satisfactory.
2012 Elsevier Ltd. All rights reserved.

1. Introduction It is very important to know the behavior of plate heat exchang-

Plate heat exchangers are important components of chemical pro-
cesses and power industries; they are being increasingly used in pet-
rochemical processes due to their highly efcient heat recovery and
their compactness [1]. Plate heat exchangers are probably the most
common type of heat exchangers applicable for a wide range of oper-
ating temperatures and pressures. Aminian et al. [2] stated that plate
heat exchanger weighs 95% less than comparable conventional shell-
and-tube exchangers and provide 10001500 square meters of heat
transfer per cubic meters of exchanger volume.
Due to the importance of the plate heat exchangers, many
researchers have investigated the modeling and the optimum de-
sign parameters for the system. For example; Lalot et al. [3] have
suggested a non-linear model using an extended Kalman lter
and it has been applied to on-line detection of fouling in the heat
exchanger. Ling and Peng [4] have developed articial neural net-
works in order to predict the pressure drop and heat transfer char-
acteristics in the plate heat exchanger. Mishra et al. [5] have used a
generic technique of algorithm-based optimization for plate heat
exchangers. Their optimization program is aimed to minimize
the number of entropy generations for a specic heat duty. Yang
et al. [6] have developed a 3-dimensionally distributed-parameter
model for plate heat exchangers. The model allows for varying the
local uid thermo-dynamical properties inside the ow path.
Corresponding author. Tel.: +968 2544536; fax: +968 25446245.
E-mail address: salam@Unizwa.edu.om (S.K. Al-Dawery).
ers when they are subjected to transient ow and to know the ow
variation required to control them when temperature changes take
place. Dynamic analysis of plate heat exchangers provide informa-
tion about transient responses subjected to various disturbances.
However, there is a limited number of studies that have been re-
ported on the transient analysis and control of plate heat exchang-
ers. Burns [7] has developed a model for a heat exchanger system
based on the analysis of the real poles and zeros of the system
transfer function. His results indicate that the transfer function of
the heat exchanger can be approximated by a lead/lag function.
Ghanim [8] has studied the dynamics of plate heat exchangers
using a step change technique applied to the cold water ow rate.
His analysis of the results shows that the system is represented by
a rst order system with a negligible time delay. In his work, the
time constant has been measured for various ow rates and he
concluded that the time constant is inversely proportional to the
ow rate. Also, another theoretical and experimental analysis of
the dynamic characteristics of plate heat exchangers has been pre-
sented by Khan et al. [9]. Their investigation has shown that the
system is modeled by rst and second order transfer function with
dead time. Recently, Das and Dwivedi [10] have used a predictive
model to suggest transient responses for different variable
changes, however, no clear model or transfer function has been
presented. Mahdi et al. [11] have developed a two-dimensional
dynamic model for milk fouling in a plate heat exchanger. Their
conclusion stated that the aggregation rate of unfolded protein is
found to increase exponentially with wall temperatures rising
and can be accompanied by a substantial reduction in the heat-

0017-9310/$ - see front matter 2012 Elsevier Ltd. All rights reserved.
http://dx.doi.org/10.1016/j.ijheatmasstransfer.2012.06.094
6874 S.K. Al-Dawery et al. / International Journal of Heat and Mass Transfer 55 (2012) 68736880

Nomenclature

a constant in Eq. (8) ss steady-state

A area of heat transfer, m2
b constant in Eq. (8) Greek
c constant in Eq. (8)
f damping coefcient
Cp specic heat, J/kg K s time constant
D derivative control x frequency
e error, K
I integral control
G transfer function Abbreviation and nomenclature on the schematic diagram
m_ mass ow rate, kg/s ADC analog to digital converter
ix imaginary Amp volt amplier
K gain, K/kg/s AP air-compressor
M mass capacity, kg APR air-pressure lter and regulator
P proportional control DAC digital to analog converter
P(t) controller output pressure signal, psi FCV ow control valve
S Laplace term I/P current to pressure converter
t time, s P1 cold water pump
T Temperature, K P2 hot water pump
U over all heat transfer coefcient, W/m2 K PC personal computer
PHE plate heat exchanger
R1 hot water rotameter
Subscript
R2 cold water rotameter
C cold water & controller T100 cold water tank
D derivation
T200 hot water tank
H hot water TE temperature element
I inlet TIC temperature indicator control
I integral TT temperature transmitter
o outlet
V/I volt to current converter
P process

transfer coefcient. Tan et al. [12] have studied the use of articial
neural networks in order to simulate the thermal performance of a
plate heat exchanger with no suggested model or transfer function.
Only a few researchers have focused on the control of heat
transfer systems such as boiler and plate heat exchangers. Assilian
and Mamdani [13] have realized that the fuzzy logic controller
could be used not only in the treatment of complex heuristic sys-
tems, but it could also be applied to hard systems such as industrial
plant controllers. In their report, small boiler steam engine control-
lers specifying heuristic fuzzy control rules for two feedback loops
have been implemented and the results have been compared with
those of the Direct Digital Control for different set points. The re-
sults show that the quality of control with the fuzzy controller is
better than that obtained by the xed controller. Kickert and Van
Naute Lemke [14] have applied three types of fuzzy logic control-
lers to the temperature control of a warm water plant and com-
pared the results with that under the conventional PI controller.
They used a continuous type membership functions to describe
the fuzzy sets. The process has properties that are difcult to con-
trol. These difculties arise from non symmetric behavior for heat-
ing and cooling, noise, dead time and the inuence of the ambient
temperature on behavior of the process. King and Mamdani [15]
have applied a temperature control using a fuzzy logic controller
on the boiler and stirred tank reactor. The formulating design rules
of the fuzzy logic controller require some knowledge of the process
such as process delays, and speed and magnitude of responses.
They concluded that processes could be controlled effectively using
heuristic rules based on fuzzy statements and the results indicate
that the fuzzy control system is much less sensitive to process
parameter changes and can give good control at all operating
points. Morison [16] has proposed and tested a control algorithm
that calculates the set point of the hot water temperature from
the set point of pasteurization temperature and an average
temperature based on the steady-state analysis and the effects of
disturbances on the process temperatures. His results indicate that
the control based on a steady temperature is more effective and
could be very good for disturbance rejection and set point tracking.
It can be seen that the necessity for developing a general math-
ematical model for plate heat exchangers is an important task for
the purpose of their control and their transient behavior. Thus,
the present work has been conducted to nd a general dynamic
model and a suitable control system for the plate heat exchanger
in order to enhance the research studies of the transient dynamics
of such systems. Theoretical and experimental works will be car-
ried out; transient responses of the plate heat exchanger will be
investigated using frequency response analysis. In addition, one
of the goals is to maintain the outlet temperature of the cooling
water in the plate heat exchanger using conventional and ad-
vanced control systems. The control systems applied are: PI, PID
and fuzzy logic controllers. The nal results of the performance
will be compared theoretically and experimentally.
2. Mathematical model formulation
2.1. Mathematical model of plate heat exchanger
Unsteady-state energy balances have been used as the basis for
the derivation of the mathematical model for the plate heat ex-
changer. Two approaches are possible to employ overall balances.
In rst approach, the overall heat transfer coefcient (U) is consid-
ered to be constant and in the second approach, the overall heat
transfer coefcient is considered as a function of the hot stream
mass ow rate. Hence, in the latter instance, U is also considered
as a function of time (i.e. U(t)).
Assuming U to be constant, the unsteady-state energy balance
around the cold plate is given by:
 S.K. Al-Dawery et al. / International Journal of Heat and Mass Transfer 55 (2012) 68736880 6875

dT co t K7 K8
m _
_ c C p T ci  T co t mhtC p T hi  T ho t M C C p 1 T ho s T co s _ h s
m 12
dt 1 sh s 1 sh s
and the unsteady-state energy balance around the hot plate is given where:
by: M C Q
sc _ c p ; K 5 _ and
dT ho t mc C p Q mc C p Q
_ h C p T hi  T ho t m
m _ c tC p T ci  T co t Mh C p 2 RT hi T ho  T ci  T co Mh C Q
dt K6 sh ; K7 and
_ c Cp Q
m mh C  Q _ hCp Q
m
Linearization of the two equations above and transforming to _ bh
C p T hi  T ho  RT hi T ho  T ci  T co Zm
Laplace domain gives: K8 Q and
_ hCp Q
m _ b
amh c
K1 K2 Zc m _ b1
h
T co s _ h s 
m T ho s 3 R
1 sc s 1 sc s amh c2
_ b

By substituting Eq. (11) into Eq. (12), the following transfer

K3 K4 function is obtained:
T ho s _ h s 
m T co s 4
1 sh s 1 sh s
T hi T ho _ T co s K 8 sc s K 8 K 7 K 6
where sc M _ c ; K1
c
and K 2 m h Mh
_ c ; sh m
_ h ; K3
T hi T ho
and Gs
m_c
m m_c m m _h
mh s sc sh s2 sc sh s 1  K 5 K 7
K 4 m_
h
By substituting Eq. (4) into Eq. (3), the following equation is Hsa s 1
13
obtained: s2p s2 2fsp s 1
T co s K 1 sh K 1  K 2 K 3 n o1=2
K 8 sc sc sh K 8 K 7 K 6

_ h s sc sh s2 sc sh s 1  K 2 K 4
5 where sa H1K ; sp ; H 1K and
m 5 K7 1K 5 K 7 5 K7

sc sh
_ c mh
m
As 1  K 2 K 4 1  m
_ T hi T ho
m
_ T hi T ho f
_ c 0, and K 1  K 2 K 3 0
h
_h m m_c _c
m m _h 2sc sh 1  K 5 K 7 1=2
Thus, Eq. (5) becomes: From the above mathematical derivation of the dynamic model
T co s K 1 sh K for the plate heat exchanger it can be seen that the system is rep-
Gs 6 resented by rst order lead and second order lag expression. Thus,
mh s sc sh s sc sh 1 sp s
the overall behavior of the plate heat exchangers would follow the
where: K sKc 1ssh and sp ssccssh responses of a rst order lag system.
h h
For the case when U is a function of time (i.e. U(t)), the unstea-
dy-state energy balance around the cold plate is given by:
2.2. Control system for plate heat exchanger
 
T  T co t T ho t  T ci
_ c C p T ci  T co t AUt hi
m The conventional PID and fuzzy logic controllers have been
2 2
implemented for temperature control on the plate heat exchanger.
dT co t The mathematical representation of the PID controller is given by:
Mc C p 7
dt
 Z 
The resistance to heat transfer is considered by following formula 1 det
Pt Pss K C et etdt T D 13
[8]; TI dt

1 c Controller tuning involves the selection of the best values of KC, TI

a b
8 and TD. This is often a subjective tuning procedure and is certainly
Ut _
mh t
process dependent. A number of methods have been proposed over
where a, b and c are constants the past 50 years; however, the CohenCoon method has been se-
By substituting the overall heat transfer coefcient from the lected due to its reliability for predicting the optimum controller
above formula into Eq. (7) and putting A/2 = Z, the following settings.
equation is obtained: Advanced control techniques are preferable nowadays for con-
trolling many systems in different engineering disciplines. One of
Zm _ bh t these advanced control techniques is the fuzzy logic controller.
_ c C p T ci  T co
m T T ho t  T ci  T co t
_ h t c hi
am b
One of the aims of fuzzy sets theory is the development of a meth-
dT co t odology for the formulation and solution of problems that are too
Mc C p 9 complex or too ill-dened to be analyzed by conventional tech-
dt
niques. Hence, the fuzzy sets theory is likely to be recognized as
The unsteady-state energy balance around the hot plate leads to:
a natural development and evaluation of scientic thinking [15].
  The general block diagram of the fuzzy logic control system is
T  T co t T ho t  T ci
_ h C p T hi  T ho t AUt hi
m shown in Fig. 1 [17]. The complete procedures and details of the
2 2
fuzzy controller design can be found in the work of Passino and
dT ho t Yurkovich [18].
Mc C p 10
dt
By substituting the overall heat transfer coefcient from Eq. (8) into
3. Experimental studies
Eq. (10) and putting A/2 = Z in the same manner to that of Eq. (7),
the resultant equation and Eq. (9) are linearized and transformed.
Many experimental tests have been performed on the plate heat
The following equations are obtained:
exchanger for the purpose of its modeling, its dynamic behavior
K5 K6 and its control system. Different operating conditions and two con-
T co s T ho s _ h s
m 11
1 sc s 1 sc s trol theories have been applied. The analysis of the results has, also,
been carried out.
6876 S.K. Al-Dawery et al. / International Journal of Heat and Mass Transfer 55 (2012) 68736880

In P
Calculate Convert Evaluate Compute Out
change of to fuzzy fuzzy control
Process
error set control action
definition rules
form

Fig. 1. Block diagram of fuzzy logic controller.

TE AMP

TT PC
P
H
To the drain
E

TIC V/I

I/P APR
R1 R2

FCV
T100 T200

AP
P2
P1

Fig. 2. Schematic diagram of the experimental rig.

Table 1
Description of the experimental rig.

Code Component Description

AP Air-compressor JUN-air compressor, 340 W, 50 cycles, 2900 rpm, maximum air pressure 150 psig
APR Air-pressure lter and regulator Honeywell air lter regulator, inlet pressure (maximum) 226 psi, regulator range 1.586 psi
I/P Transducer Platon type 100, 420 mA input and 0.21 bar output
FCV Flow control valve Size 00 , platon type Mvc/A, pneumatic action, 315 psig, air to close, valve ow coefcient Cvmax = 3.1
PHE Heat exchanger a-level plate heat exchanger, type Pol.
P1 Cold water pump Centrifugal pump, 3-phases, 210 W, Struart Turner Ltd., 210 W, 11.454.4 l/min, pomes guila 3-D.
P2 Hot water pump 00 , 400 W, 1.6 m3/h centrifugal pump
PC Personal computer 386 IBM
R1 Hot water rotameters G.E.C. Elliott rotameters (variable area) type, series 2000, range 020 l/min
TIC Temperature indicator and controller 3-term (PID), Honeywell versapak type-7450 auto/manual station, 420 mA input and 420 mA output
T100 Cold water tank Galvanized cubic tank, capacity 0.5 m3
T200 Hot water tank Rectangular tank 60  100  58 cm3, with three immersion heaters each with 3 kW
TT Temperature transmitter Honeywell resistance bulb temperature transmitter model Y-785309
V1 Cold water valve 3= 00
4 global valve
V2 Hot water valve 3= 00
4 global valve
TE Temperature element Thermocouple type T
Amp Volt amplier
 S.K. Al-Dawery et al. / International Journal of Heat and Mass Transfer 55 (2012) 68736880 6877

Table 2 control valve opening is adjusted to maintain the hot water ow

Plate heat exchanger specications. rate. The process must be left under operating without any distur-
Equivalent diameter of channel, mm 5.15 bances until the steady state is reached. After reaching the steady
Volume of channel, m3 85  106 state condition, than, the control systems are applied. For each test,
Length of channel, mm 33.125 a step change is made and then the outlet temperature of cold
Heat transfer area of PHE, m2 0.627
Minimum Reynolds number (Re) 300
water is recorded until new steady state is reached.
Maximum operating temperature 130 oC
Maximum operating pressure 16 bar
4. Results and discussion

4.1. Model of plate heat exchanger

3.1. Experimental setup

The best method of obtaining an empirical linear process model

The layout of the experimental set-up used in this work is
is to dene its parameters such as dead-time, time constant and
shown schematically in Fig. 2. All experimental items are listed
damping coefcient. Two different step changes have been applied,
in Table 1. The main part of the experimental apparatus is the
100% as a positive step in the hot water ow rate and 50% as a
plate heat exchanger (type ALFA-LAVAL, PO1 containing 21
negative step in the cooling water ow rate. The results of the
stainless steel plates, assembled in counter-current conguration,
transient response of the outlet temperature of cooling water
single pass/single pass, and 10 channels per pass). The specica-
expressed in perturbations are shown in Figs. 4 and 5.
tions of the plate heat exchanger are given in Table 2.
The analysis of these step responses indicate that the model for
The sump tank capacity for the hot water supply is 0.07 m3. The
the plate heat exchanger can be represented by rst order lag with
tank is designed and coupled with three immersed heaters supply-
dead time. The determined values for the time constant and dead
ing the energy required and minimizing the variation of tempera-
time are as follows: In the rst test, the time constant and dead
ture of the hot water stream. Each heater has power 3 kW. The
time are determined to be 12.5 and 5 s respectively, while in the
sump tank and all pipe connection are insulated with ber glass.
second test, the time constant and dead time are 13.5 and 5 s
Tap water has been used as the cooling medium due to its reason-
respectively. The variation in the time constant is due to the
able heat capacity.
non-linearity of the system. However, the mismatch between
The Analog temperature controller is a PID controller (Honey-
experimental and simulated responses and the saturation of that
well type-7450 auto/manual and 4-20 mA output) and it is tted
appears in the nal stage at each response is due to the control
with the resistance temperature transmitter (Honeywell type Y-
valve hysteresis and stickiness (as demonstrated in Fig. 3). Such
78530g). The ow control valve is (platoon, type MVc/A, air
deection in the responses can be eliminated by using a valve posi-
to close pneumatic action of 315 psig input). The valve is con-
tioner [19]. The valve stickiness resists the valve stem movement
nected to a current-to-pressure converter (Platon type 100, with
especially when the input signal to the valve is too small. The com-
420 mA input and 0.11 bar output).
puter could be another source of error due to batch buffering status
For the purpose of computer control application, a 12 bit inter-
of data that occurs within DAC.
face unit is used for ADC and DAC converters; 8 channels for ADC
One useful and practical method for process analysis is the
and 3 channels for DAC. The computer used is a 386-IBM with
process pulse responses. These pulse responses will be a basis for
RAM of 4 M byte. Volt-to-pressure converter is, also, used to con-
creating the frequency response data and plots, namely Bode plot
vey the computer signal output to the pneumatic control valve.
[20]. In this method, it is only requiring a fraction of time that
can be directed to sinusoidal function and expanded by Fourier
3.2. Experimental procedure
series. The input and output functions are then transformed to give
the system frequency domain G(ix).
A calibration curve has been drawn to determine the valve hys-
An experimental pulse test has been performed using 100% as a
teresis as shown in Fig. 3. For each test, the heating elements are
positive step in the hot water ow rate with duration of 5 s. The
switched on for 30 min before starting in order to reach an optimal
temperature response of outlet cold water is shown in Fig. 6.
temperature and at the same time, the cold water ow rate must
be maintained at the desired ow rate. Then, the pneumatic
10

200

8
180
T of cold water, C

A
6
160
Pressure, kpa

0.632A
140 4
100 % step change
% vs pressure+
% vs pressure- T = 8.37 (1-exp(-(t+5)/12.53)
120
2

100

0
80 0 20 40 60
0 20 40 60 80 100 Time, s
Stem position, %
Fig. 4. Temperature response for a 100% positive step in hot water feed ow rate,
Fig. 3. Input pressure versus valve opening (%). step introduced at time = 0.0.
6878 S.K. Al-Dawery et al. / International Journal of Heat and Mass Transfer 55 (2012) 68736880

10 200

8 100
of cold water, oC

Angle, rad
0

-
150% step change
4 -100
T = 8.46 (1-exp(-(t+5)/13.47) Frequency vs Angle

2 -200

0 -300
0 20 40 60 0.0001 0.001 0.01 0.1
Time, s

Fig. 5. Temperature response for a 50% negative step in cooling water feed ow Fig. 8. Bode plot of phase angle.
rate, step introduced at time = 0.0.

6 Table 3
Tuned parameters of PI and PID controller.

Controller parameters
5
Kc TI TD
PI 1.1 9
4
PID 1.8 10.3 1.7
T, oC

7
2
6

1
5
Response, oC

PI simu
0 4
PI real
0 20 40 60
3
Time, s

Fig. 6. Pulse temperature response. 2

0
0
-1
-10 0 20 40 60 80 100 120 140
Log of modulus, BD

Time, s
-20
Fig. 9. Comparison between simulation and experimental responses of PI
controller.
-30

-40 20 dB. This result indicates that the model of the plate heat ex-
changer can be considered as a rst order lag. Since, the phase an-
-50 gle is pulled far away from 90o, it shows a clear embedded dead
time in the process. The analysis of the above result indicate that
-60 the transfer function representing the model of the plate heat ex-
0.01 0.1 1 10 changer can be written as:
, Rad/s
1:97e5s
Fig. 7. Bode plot of log of modulus. Gs
12:5s 1

The resultant frequency responses that have been calculated 4.2. Control of plate heat exchanger
using pulse test are shown in Figs. 7 and 8. These gures contain
Bode plot and phase angle respectively. From Bode plot, it can be Both experimental and simulation studies have been carried out
seen that the resultant slope at a high frequency is approximately to control the outlet temperature of the cooling water in the plate
 S.K. Al-Dawery et al. / International Journal of Heat and Mass Transfer 55 (2012) 68736880
4
PID simu
PID real
Response, oC
-2
-4
0 100 200 300 400 500
Time, s
Fig. 10. Comparison between simulation and experimental responses of PID
controller.
Fig. 11. Comparison between simulated temperature responses under PI, PID and
fuzzy logic controller.
heat exchanger. For simulation purposes, MATLAB Simulink has
been used. For the temperature control application using the con-
ventional PI and PID controllers, the best controller settings have
been determined are given in Table 3. For temperature control pur-
poses, a 100% negative step changes in the cooling water feed ow
rate have been made.
The simulated responses of the temperature control of plate
heat exchanger using PI and PID controller are shown in Figs. 9
and 10. These responses show an acceptable level of validity of
the derived model for the plate heat exchanger. Also, it can be seen
that the PI controller provides a better temperature control than
that achieved using the PID controller in terms of shorter setting
time and smaller size of oscillation. For the purpose of comparison,
experimental studies have been performed for temperature control
of the plate heat exchanger using PI and PID control. The experi-
mental responses of plate heat exchanger under PI and PID control-
ler are shown in Figs. 9 and 10. In respect to both experimental and
simulation results, it can be seen that temperature responses of
plate heat exchanger under PI and PID are within an acceptable
level. Also, it can be seen that the PI controller provides a better
temperature control than that achieved using the PID controller
in terms of shorter setting time and smaller size of oscillation. This
result is corroborated by the theoretical model. However, the
6879
600
Fig. 12. Comparison between experimental temperature responses under PI, PID
and fuzzy logic controller.
deviation between experimental and simulated responses which
causes a sluggish behavior in the experimental results are caused
by the control valve hysteresis and stickiness. Such deection in
the responses can be eliminated using a valve positioner as ex-
plained in previous section.
For the application of the advanced control techniques such as
the fuzzy logic controller, a well designed fuzzy controller has been
constructed and used for the temperature control of the plate heat
exchanger. Firstly, a simulation study has been performed in order
to show the reliability of the fuzzy controller application. The
result of the simulation for temperature control under the fuzzy
controller is shown in Fig. 11. This response reveals a good and
an acceptable level of control behavior. For the purpose of compar-
ison, the temperature control response under fuzzy logic is com-
pared to that obtained using PI and PID controllers, the response
obtained under fuzzy controller appears to be superior to that ob-
tained under PI and PID controllers as shown in Fig. 11.
The experimental temperature response using fuzzy logic
controller is shown in Fig. 12 along with that obtained using PI
and PID. This gure shows the superiority of the fuzzy logic appli-
cation over that obtained using both the PI and PID controller in
term of fast settling time and less oscillation.
5. Conclusions
The mathematical model developed for the plate heat exchan-
ger shows a considerable level of non-linearity. After applying lin-
earization to initial mathematical equations and transforming to
the Laplace domain, the developed model appears to be described
by a transfer function with rst order lead and second order lag.
However, the analysis of the experimental data using frequency re-
sponse analysis technique shows that the plate heat exchanger is
represented by a rst order lag with dead time. An acceptable level
t has been obtained between theoretical and experimental re-
sponses; such a t would prove the validity of the proposed model
for the plate heat exchanger.
From both experimental and theoretical applications of the con-
ventional and fuzzy logic controller, it appears that the fuzzy logic
controller shows a superior performance over that of the conven-
tional controllers. Finally, based on the results obtained from the
developed model for the plate heat exchanger, it can be concluded
such model can be used as a tool to enhance the operation of the
plate heat exchanger and hence, can be suggested as a general
guideline for future studies.
6880 S.K. Al-Dawery et al. / International Journal of Heat and Mass Transfer 55 (2012) 68736880

References [10] K.S. Das, A.K. Dwivedi, Dynamics of plate heat exchangers subject to ow
variations, Int. J. Heat Mass Transfer 50 (2007) 27332743.
[11] Y. Mahdi, A. Mouheb, L. Oufer, A dynamic model for milk fouling in a plate heat
[1] B.R. Lamb, Plate heat exchangers, a low cost route to heat recovery, J. Heat
exchanger, Appl. Math. Model. 33 (2009) 648662.
Recovery Syst. 2 (3) (1982) 247255.
[12] C.K. Tan, J. Ward, S.J. Wilcox, R. Payne, Articial neural network modeling of
[2] K. Aminian, L. English, D. Patchen, H. Siriwardane, Thermoacoustic liquefaction
the thermal performance of a compact heat exchanger, Appl. Thermal Eng. 29
of coal mine methane to produce LNG for heavy vehicle applications, U.S.
(2009) 36093617.
Department of Energy Report, Morgantown energy technology center, 1999.
[13] S. Assilian, E.H. Mamdani, An experiment in linguistic synthesis with a fuzzy
[3] S. Lalot, G.R. Jonsson, O.P. Palsson, B. Desmet, Use of extended Kalman ltering
logic controller, Int. J. Man-Mach. Stud. 7 (1974) 113.
in detecting fouling in heat exchangers, Int. J. Heat Mass Transfer 50 (2007)
[14] W.J. Kickert, H.R. Van Naute Lemke, Application of a fuzzy controller in a
26432655.
warmwater plant, Automatica 12 (4) (1976) 301308.
[4] X. Ling, H. Peng, Optimal design approach for the plate-n heat exchangers
[15] P.J. King, E.H. Mamdani, The application of fuzzy control systems to industrial
using neural networks cooperated with genetic algorithms, Appl. Thermal Eng.
processes, Automatica 13 (3) (1977) 235245.
28 (2008) 642650.
[16] K.R. Morison, Steady state control of plate pasteurisers, Food Control 16 (1)
[5] M. Mishra, P.K. Das, S. Sarangi, Second law based optimisation of crossow
(2005) 2330.
plate-n heat exchanger design using genetic algorithm, Appl. Thermal Eng. 29
[17] D.F. Ahmed, S.K.H. Al-Dawery, H.A.O. Al-Anbari, On-line control of
(2009) 29832989.
neutralization process based on fuzzy logic, Emirates J. Eng. Res. 12 (2)
[6] C. Yang, L. Zhang, J. Zhou, A distributed parameter model and its application in
(2007) 8997.
optimizing the plate-n heat exchanger based on the minimum entropy
[18] K.M. Passino, S. Yurkovich, Fuzzy Control, Addison-Wesley, 1998.
generation, Int. J. Thermal Sci. 49 (2010) 14271436.
[19] S.K. Al-Dawery, Implantation of advanced control techniques on small
[7] A.J. Burns, Dynamic system, measurement and control, ASME 3 (1981).
computer, M.Phil Thesis, Nottingham University, UK, 1988.
[8] M. Ghanim, Dynamics of plate heat exchanger, M.Sc. Thesis, University of
[20] M.L. Luyben, W.L. Luyben, Essentials of Process Control, McGraw Hill, 1997.
Baghdad, Iraq, 1982.
[9] A.R. Khan, N.S. Baker, A.P. Wardle, The dynamic characteristics of a counter-
current plate heat exchanger, Int. J. Heat Mass Transfer 31 (6) (1988)
12691278.