Journal of Food Engineering 69 (2005) 235–244

www.elsevier.com/locate/jfoodeng

Milk fouling simulation in helical triple tube heat exchanger

P.K. Sahoo a, I.A. Ansari b, A.K. Datta b,*

a
Department of Food Engineering, Faculty of Agricultural Engineering, Bidhan Chandra Krishi Viswavidyalaya,
Mohanpur 741 252, Nadia, West Bengal, India
b
Department of Agricultural and Food Engineering, Indian Institute of Technology, Kharagpur 721302, West Bengal, India

Received 3 November 2003; accepted 6 August 2004

Abstract

Heat exchanger fouling is a common phenomenon during high temperature processing of milk. The temperature of milk raises
from 90
C to 150
C during Ultra-High-Temperature (UHT) sterilization process. At such high temperature, the minerals and dena-
tured proteins deposit on the heat exchanger surface, also known as fouling. Fouling acts as resistance to heat transfer, hence the
performance of the heat exchanger is reduced. Using hydrodynamics and heat balance concept, a mathematical model was formu-
lated. Simulation was performed with the model to predict the fouling behavior as a function of time and position within the helical
triple tube heat exchanger (HTTHE). At an early period of operation, the uniform fouling deposit occurs throughout the length of
the heat exchanger due to constant heat exchanger wall temperature. With progress of time, the fouling deposit and also Biot num-
ber (i.e., local fouling factor) increases towards the outlet of the heat exchanger since the interface temperature between fouling
deposit and milk approaches towards the bulk milk temperature, that increases towards the heat exchanger outlet. The fouling
deposit stabilizes after 105 min since no net deposit occurs after that time.
 2004 Elsevier Ltd. All rights reserved.

Keywords: Milk; Fouling; Simulation; Helical triple tube; Heat exchanger; UHT

1. Introduction tion on the heat exchanger surface is a serious problem

UHT sterilization of milk means heating milk by a
continuous process to 135–150
C and holding it for a
few seconds. The sterilized milk is not completely free
of organisms. It is free of sporulating, toxicogenic and
pathogenic organisms to a level so that it remains safe
for consumption for several weeks at room temperature
(Pien, 1955). The electrochemical deposition of milk sol-
ids such as denatured proteins and mineral salts, is most
severe in the temperature range of 90–120
C (Burton,
1988). Since a tubular heat exchanger is likely to main-
tain a constant wall temperature of 160
C, the deposi-
*
Corresponding author. Fax: +91 3222 82244.
E-mail addresses: pks_1973@lycos.com (P.K. Sahoo), ansari@
mailcity.com (I.A. Ansari), akd@agfe.iitkgp.ernet.in (A.K. Datta).
for UHT sterilizers. Fouling layers increase the heat
transfer resistance, which leads to reduction of milk out-
let temperature. Fouling also reduces the flow area of
milk so that pressure drop increases. Since fouling re-
duces the continuous run time of a heat exchanger, hence
frequent cleaning is needed. Although many researchers
and scientists all over the world have already worked and
developed lots of fouling models, the actual mechanism
of fouling is yet to be known. After surveying numerous
fouling mechanisms (Changani, Belmar-Beiny, & Fryer,
1997; Taborek, Aoki, Ritter, Palen, & Knudsen, 1972a)
it is suggested to develop a fouling model considering
flow pattern and temperature of the process plant.
According to Visser and Jeurnink (1997) the very high
flow rates of fluid can be able to prevent its solids depo-
sition and subsequent sticking. A series of possible scale
up strategies (i.e., constant Reynolds number, constant

0260-8774/$ - see front matter  2004 Elsevier Ltd. All rights reserved.
doi:10.1016/j.jfoodeng.2004.08.014
236 P.K. Sahoo et al. / Journal of Food Engineering 69 (2005) 235–244

Nomenclature

Af area of flow of process fluid, m2

Ai2 heat transfer surface area based on inside diameter of middle tube, m2
Ao1 heat transfer surface area based on outside diameter of innermost tube, m2
Bi Biot number, dimensionless
Bimi Biot number at outer surface of innermost tube, dimensionless
Bimo Biot number at inner surface of middle tube, dimensionless
CPf specific heat of process fluid at mean temperature, J/kg K
Dc coil diameter, m
Deq equivalent diameter of tube, m
Di1 inner diameter of innermost tube, m
Di2 inner diameter of middle tube, m
Do1 outer diameter of innermost tube, m
E activation energy, J/mol
f friction factor, dimensionless
g acceleration due to gravity = 9.81 m/s2
hc1 steam/condensate heat transfer coefficient at surface of innermost tube, W/m2 K
hc2 steam/condensate heat transfer coefficient at surface of outermost tube, W/m2 K
hf film heat transfer coefficient with fouling, W/m2 K
hf0
film heat transfer coefficient without fouling, W/m2 K
kc thermal conductivity of condensate, W/m K
kd deposition rate constant, s
1
kr removal rate constant, s
1
L length of tube, m
N number of nodes
NDN dean number, dimensionless
(NDN)critical critical dean number, dimensionless
NNu Nusselt number, dimensionless
NPr Prandtl number, dimensionless
NRe Reynolds number, dimensionless
Q flow rate of process fluid, m3/s
q heat transfer rate, W
R universal gas constant = 8.314 J/mol K
rd1 outer radius of innermost tube including fouling layer on its outer surface, m
rd2 inner radius of middle tube including fouling layer on its inner surface, m
ri1 inner radius of innermost tube, m
ro1 outer radius of innermost tube, m
ri2 inner radius of middle tube, m
ro2 outer radius of middle tube, m
t time, s
Tf temperature of process fluid,
C
T IN
f initial temperature of process fluid,
C
Tfi temperature at the interface of fouling deposit and process fluid,
C
Ti1wall temperature of wall at inner side of innermost tube,
C
Ti2wall temperature of wall at outer side of middle tube,
C
Ts temperature of steam,
C
Ui2 overall heat transfer coefficient for the inner surface of middle tube in helical triple tube heat exchanger
with fouling, W/m2 K
0
U i2 overall heat transfer coefficient for the inner surface of middle tube in helical triple tube heat exchanger
without fouling, W/m2 K
Uo1 overall heat transfer coefficient for the outer surface of innermost tube in helical triple tube heat exchan-
ger with fouling, W/m2 K
0
U o1 overall heat transfer coefficient for the outer surface of innermost tube in helical triple tube heat exchan-
ger without fouling, W/m2 K
 P.K. Sahoo et al. / Journal of Food Engineering 69 (2005) 235–244 237

vf average velocity of process fluid, m/s

x distance along heat exchanger length, m
qc density of condensate, kg/m3
qf density of process fluid at mean temperature, kg/m3
k latent heat of condensation, J/kg
kd thermal conductivity of fouled layer, W/m K
kw thermal conductivity of tube wall, W/m K
/mi film coefficient and thermal conductivity ratio for outer surface of innermost tube, dimensionless
/mo film coefficient and thermal conductivity ratio for inner surface of middle tube, dimensionless
sw shear stress at the tube wall, N/m2
lc viscosity of condensate, Pa s
lf viscosity of process fluid at mean temperature, Pa s
lwall viscosity of process fluid at wall temperature, Pa s
d thickness of fouling layer, m
DP pressure drop, N/m2
Da temporal separation of nodes along-characteristic, s

shear stress and constant temperature change for fouling in the innermost tube and also in the outer annulus at a
fluid) is tested using a computational model of a simple constant temperature of 160
C. Milk inlet temperature
heat exchanger (Schreier & Fryer, 1995). Air bubble at is maintained 90
C, which is taken from preheating sec-
the surface accelerates the fouling process and causes a tion of the UHT sterilizer setup. Milk outlet tempera-
shift in the protein composition from serum proteins to- ture is recorded at regular time intervals. Milk gains
wards casein. At high shear rates, the flow condition may enthalpy from steam through tube walls (i.e., walls of
prevent this deposition of casein micelles (Jeurnink, middle and innermost tubes). The enthalpy balance on
1995). The model of Swartzel (1983) predicts the run time the outer surface of the inner tube and also on the inner
and changes in thermal effects throughout the processing surface of the middle tube of the concentric HTTHE, as
of milk. The mathematical model developed by Sandu shown in Fig. 1 is given by:
and Singh (1991) estimates the increased energy costs

dT f
associated with fouling of heat exchangers. Delplace, Af qf vf C pf T f dt
Af qf vf C pf T f þ dx dt
Leuliet, and Levieur (1997) and Yoon and Lund (1994) dx
developed temperature profile models to predict the þ U o1 Ao1 ðT s
T f Þdt þ U i2 Ai2 T s
T f dt
change in product temperature during fouling. The
dT f
model of Reitzer (1964) relates the linear temperature ¼ Af qf C pf dxdt
change with second order chemical reaction for crystal dt
growth. A temperature dependent milk fouling model or;
is formulated and simulated in a double concentric tube

dT f dT f
heat exchanger to predict fouling thickness on the heat Af qf C pf dx vf þ ¼ ðU o1 Ao1 þ U i2 Ai2 ÞðT s
T f Þ
dx dt
exchanger surface (Ranjan & Datta, 1999). Biot number
is a dimensionless local fouling factor that can be used or;
for the design of heat exchanger. In this paper we have
dT f dT f ðU o1 Ao1 þ U i2 Ai2 ÞðT s
T f Þ
formulated and then simulated a local milk fouling vf þ ¼
dx dt Af qf C pf dx
model to predict the fouling thickness, Biot number
and bulk milk temperature as a function of time and po- or;
sition along the length of the helical triple tube heat
dT f dT f ðU o1 At1 þ U i2 At2 ÞðT s
T f Þ
exchanger. vf þ ¼
dx dt qf C pf
where;
2. Development of fouling model
Ao1 2pro1 dx 2r A
At1 ¼ ¼   ¼  2 o1 2  At2 ¼ i2
2.1. Enthalpy balance in control element Af dx p r2i2
r2o1 dx ri2
ro1 Af dx
2pri2 dx 2r
During thermal processing, milk flows in the middle ¼  2  ¼  2 i2 2 
p ri2
r2o1 dx ri2
ro1
annular portion of the Helical triple tube heat exchanger
(HTTHE) at a constant flow rate of 135 l/h. Steam flows ð1Þ
238 P.K. Sahoo et al. / Journal of Food Engineering 69 (2005) 235–244

Ui2,Ai2

Af.ρf.νf.CpεTf.δt

Uol,Aol δTf
Af .ρf .νf .Cpf Tf + δx δt
δx

δx Tube wall
Passage for steam
x x + δx
Passage for milk

Fig. 1. Enthalpy balance on a differential volume of a triple tube heat exchanger.

Eq. (1) is a constitutive expression in partial differential The condensate heat transfer coefficient (hc1) is esti-
form for fluid temperature distribution in tubular heat mated (Kern, 1997) by the following expression:
exchanger.
0:25
k3c q2c gk
hc1 ¼ 0:5754 ð4Þ
2.2. Heat transfer equations for HTTHE lc Di1 ðT s
T i1wall Þ

The distribution of fluid, wall and deposit tempera- The fluid side heat transfer coefficients (h0f and hf) are
tures in a concentric triple tube heat exchanger is shown estimated based on an iterative technique developed by
in Fig. 2. The temperature distribution was used to de- Sahoo (2001). The basic equation involved in the itera-
rive the heat transfer expressions as given below. tive method is as follows:
(i) Without fouling, outer surface of the inner tube  0:53
 
2   3 0:8 0:33 Di2 Deq
0 h 0
r ln ro1 N Nu ¼ 0:027ðN Re Þ ðN Pr Þ 1 þ 3:5
q 4 hf f i1 ri1 ri1 Do1 Dc
Ts
Tf ¼ 0 þ þ 5
hf Ao1 hc1 kw ro1 ð5Þ
Putting, The properties of the process fluid such as qf, lf, Cpf etc.
 
are estimated at bulk mean temperature of the fluid for
h0f h0f ri1
ln ro1
ri1
/mi ¼ þ ð2Þ each iteration step.
hc1 kw Eq. (5) is derived from Dittus–Boelter equation
and q ¼ U 0o1 Ao1 ðT s
T f Þ. (Geankoplis, 1997) replacing the term ðllwall f
Þ
0:14
by
We get, 0:53
ðDDo1i2 Þ (Zuritz, 1990) for flow in annular region of tube.

 D
1 1 ro1 Furthermore the term ½1 þ 3:5ð Deqc Þ is multiplied on the
¼ /mi þ1 ð3Þ Dittus–Boelter equation (Chopey & Hicks, 1984) for
U 0o1 h0f ri1
the helical configuration.
(ii) With fouling, outer surface of the inner tube
Steam in 2     3
outermost tube 0 ro1 0 rd1
Steam in 0
q 4 hf h f r i1 ln ri1
h f r i1 ln ro1
0
ri1 hf 5
innermost tube
Ts
Tf ¼ 0 þ þ þ
hf Ao1 hc1 kw kd rd1 hf
Ts
Ts
Tsw Putting,
Twd Deposit Tsw
0 ro1
Twd h0f hf ri1 lnð ri1 Þ
/mi ¼ þ ;
Tf hc1 kw
Tfd
Tfd Wall of
Innermost tube q ¼ U o1 Ao1 ðT s
T f Þ
and

h0f ro1 ln rrd1

o1
Milk Bimi ¼ ð6Þ
kd
Fig. 2. Distribution of fluid, wall and deposit temperatures in a
concentric triple tube heat exchanger. In Eq. (6), rd1 = ro1 + d.
 P.K. Sahoo et al. / Journal of Food Engineering 69 (2005) 235–244 239

We get, L v2
DP ¼ 4f qf ð13Þ

 Deq 2
1 1 1 ro1 ri1 ri1 h0f
¼ þ ðBimi
1Þ þ ð7Þ The expression for Deq with the assumption of uni-
U o1 U 0o1 h0f ri1 ro1 hf rd1
form fouling thickness (d) both in the outer surface of
(iii) Without fouling, inside surface of the middle tube inner tube and the inner surface of the middle tube is
2   3 given by:
0 ro2
q h 0 h f r o2 ln ri2 ro2
Ts
Tf ¼ 0 4 f þ þ 5 Deq ¼ Di2
Do1
4d ð14Þ
hf Ai2 hc2 kw ri2
Substituting the value of Deq in equation (13), we get
where, 32fLQ2 qf
DP ¼ 2 3
ð15Þ

0:25 p2 ðDi2 þ Do1 Þ ðDi2
Do1
4dÞ
k3c q2c gk
hc2 ¼ 0:725 ð8Þ The friction factor can be calculated for turbulent flow
lc Do2 ðT s
T i2wall Þ
in helical tube by the following expression (Srinivasan,
Putting, Nandapurkar, & Holland, 1970).
  0:042
0 ro2
h0 hf ro2 ln ri2
f ¼
 0:5 0:2 ð16Þ
/mo ¼ f þ ð9Þ N Re DDeqc
hc2 kw
and q ¼ U 0i2 Ai2 ðT s
T f Þ: (for (NDN)critical < NDN < 14,000 and 0:0097 < Deqc <
D

Then, 0:135).

 NRe can be derived for triple tube heat exchanger as
1 1 ri2
¼ /mo þ 1 ð10Þ follows.
U 0i2 h0f ro2  
Q
(iv) With fouling, inside surface of the middle tube qf vf Deq qf Af Deq
    N Re ¼ ¼
2 3 lf lf
0 ro2 0 ri2
q 4 hf 0 h f r o2 ln ri2
h f r o2 ln rd2 ro2 h0f 5 qf QðDi2
Do1
4dÞ
Ts
Tf ¼ 0 þ þ þ ¼ h i
hf Ai2 hc2 kw kd rd2 hf p
ðDi2
2dÞ2
ðDo1 þ 2dÞ2 lf
4

4qf Q
Using, ¼
pðDi2 þ Do1 Þlf
0 ro2
h0f hf ro2 lnð ri2 Þ
/mo ¼ þ So,
hc2 kw
4qf Q
N Re ¼ ð17Þ
q ¼ U i2 Ai2 ðT s
T f Þ pðDi2 þ Do1 Þlf
and The inlet pressure of the helical triple tube heat exchan-
h0f ri2 ln rrd2i2 ger was maintained at 5.1 bar when running the heat
Bimo ¼ ð11Þ exchanger with milk at constant flow rate of 135 l/h.
kd The outlet temperature and pressure of the HTTHE
In Eq. (11), rd2 ¼ ri2
d. was recorded at 10 min intervals. The outlet tempera-
We get, ture and pressure both decreases with time due to in-

 crease of fouling deposit. The properties (i.e. qf, lf
1 1 1 ri2 ro2 h0f ro2
¼ þ ðBimo
1Þ þ ð12Þ etc.) of fluid were calculated at bulk mean temperature
U i2 U 0i2 h0f ro2 ri2 hf rd2 of fluid. From the recorded pressure drop data, mean
fouling thickness was calculated using the Eq. (15),
2.3. Experimental fouling which is shown in Table 1. From mean fouling deposit
data, average Biot numbers were calculated using Eqs.
During UHT processing of milk in the tubular heat (6) and (11).
exchanger, considerable amount of fouling deposit on
the heat exchanger surface. The fouling deposits increase 2.4. Simplification of heat transfer equations
the pressure drop across the heat exchanger. The pres-
sure drop for the turbulent flow in a tubular heat ex- The simplification of heat transfer equations means
changer can be computed with a standard fluid flow the expression of overall heat transfer coefficients (i.e.
equation (Geankoplis, 1997) as given below. Uo1 and Ui2) as the function of their respective Biot
240 P.K. Sahoo et al. / Journal of Food Engineering 69 (2005) 235–244

Table 1
Experimental fouling data during UHT processing of milk in HTTHE
Time, min Milk outlet temperature (Tout),
C Pressure drop (DP), bar Mean fouling layer thickness (d), mm
0 142 0.00 0.00
10 141 0.07 0.25
20 140 0.12 0.41
30 139 0.17 0.47
40 135 0.22 0.51
50 134 0.28 0.56
60 133 0.39 0.61
70 130 0.66 0.68
80 128 0.93 0.72
90 126 1.03 0.73
100 124 1.25 0.75
110 123 1.54 0.77
120 123 1.54 0.77
130 123 1.54 0.77

The regression equation of Fig. 4 is:

ro2 h0 ro2
ðBimo
1Þ þ f
ri2 hf rdx2
2
¼ 0:4761Bimo þ 0:5801Bimo þ 0:0407 ð19Þ
Applying Eq. (19) into Eq. (12), the following is
obtained:
ri1 h0
Fig. 3. Plot of ro1
ðBimi
1Þ þ hff  rrd1i1 vs. Bimi. 1 1 0:4761Bi2mo þ 0:5801Bimo þ 0:0407 ri2
¼ 0 þ
U i2 U i2 h0f ro2
numbers. Uo1 can be expressed as the function of Bimi
and Ui2 can be expressed as the function of Bimo. h0f U 0i2
h0
Plotting a graph of rro1i1 ðBimi
1Þ þ hff rrd1i1 vs. Bimi, a U i2 ¼  
h0f þ ð0:4761Bi2mo þ 0:5801Bimo þ 0:0407Þ ri2
ro2
U 0i2
polynomial line of second order (R2 = 0.99) is obtained
as shown in Fig. 3. ð20Þ
h0
The term ½rro1i1 ðBimi
1Þ þ hff rrd1i1  is taken from Eq. (7) Similarly putting Eq. (18) in Eq. (7), the following is
and Bimi data are taken from experimental fouling. obtained:
The regression equation of Fig. 3 is:
h0f U 0o1
ri1 h0 ri1 U o1 ¼  
ðBimi
1Þ þ f ¼ 0:3043Bi2mi þ 0:261Bimi þ 0:0155 h0f þ ð0:3043Bi2mi þ 0:261Bimi þ 0:0155Þ ro1
U 0o1
ro1 hf rd1 ri1
ð18Þ ð21Þ
h0
Plotting a graph of rro2i2 ðBimo
1Þ þ hff rro2
d2
vs. Bi mo, a poly-
nomial line of second order (R2 = 0.99) is obtained as 2.5. Simplification of fluid temperature distribution
shown in Fig. 4. expression
h0
The term ½rro2i2 ðBimo
1Þ þ hff rro2
d2
 is taken from Eq. (12)
and Bimo data are taken from experimental fouling. The characteristic technique described by Acrivos
(1956) is used to convert Eq. (1) in partial differential
form to ordinary differential form. A characteristic mesh
for constant wall temperature heat exchanger is shown
in Fig. 5.
Along the characteristic line:
dx
¼ vf ða-characteristicsÞ ð22Þ
dt
x ¼ constant ðb-characteristicsÞ ð23Þ
Employing the Eqs. (20)–(23) into Eq. (1), this equation
h0
Fig. 4. Plot of ro2
ri2
ðBimo
1Þ þ hff  rro2
d2
vs. Bimo. is simplified to:
 P.K. Sahoo et al. / Journal of Food Engineering 69 (2005) 235–244 241

2 3
ro1 U 0o1 h0f ri2 U 0i2 h0f
24   þ   5
ro1 ri2
h0f þð0:3043Bi2mi þ0:261Bimi þ0:0155Þ U 0o1 h0f þð0:4761Bi2mo þ0:5801Bimo þ0:0407Þ U 0i2
dT f ri1 ro2
¼   ðT s
T f Þ ð24Þ
dt r2i2
r2o1 qf C pf

2.6. Formulation of local fouling rate model The deposit rate constant (kd) is the function of shear
stress at the wall (sw) of the tube, which is shown as
The expression for the local fouling model was derived (Ranjan & Datta, 1999):
based on the concept that the net rate of solids accumu-
lation is the difference between the rate of solids deposit 1011
kd ¼ ð28Þ
and the rate of reentrainment due to fluid shear forces sw
(Fryer & Slater, 1985). The same approach has been fol-
The shear stress at the wall under turbulent flow condi-
lowed by Taborek et al. (1972a) and Taborek, Aoki, Rit-
tion Geankoplis (1997) is expressed as:
ter, Palen, and Knudsen (1972b). A generalized equation
as suggested by Fryer and Slater (1985) can be written for 22:45lf Q
local overall rate of solids accumulation as: sw ¼ ð29Þ
D3eq


dBi E 1
¼ k d exp

k r Bi ð25Þ The values of activation energy (E) and removal
dt R T fi rate constant (kr) were taken from Fryer and Slater
where the term k d exp½
ER T1fi  is a solids deposition rate (1985).
term and (krBi) is a solids reentrainment term.
Activation energy (E) = 92 · 103 kJ/kg mole.
Eq. (25) can be modified based on the configuration
Removal rate constant (kr) = 0.0013 s
1.
of helical triple tube heat exchanger as given below.
The thermal conductivity of milk fouling deposit (kd)


dBimi E ðT s
T f Þ was chosen to 0.6 W/m K (Yoon & Lund, 1994).
¼ k d exp

k r Bimi
dt R ðT s
T f ÞT fi


E ðT s
T f Þ
¼ k d exp

k r Bimi 3. Simulation procedure
R T s ðT fi
T f Þ þ T f ðT s
T fi Þ
2   3
E /mi þ 0:3043Bi2mi þ 0:261Bimi þ 0:0155 þ rro1i1 The number of nodes along a-characteristics (i.e. the
¼ k d exp 4
  5 length of heat exchanger in which fluid flows) were
R T s 0:3043Bi2 þ 0:261Bimi þ 0:0155 þ ri1 þ T f /
mi
mi ro1
selected. The temporal separation of nodes along a-

k r Bimi ð26Þ characteristics (i.e. time taken by fluid to move between
two consecutive nodes) were thus set:
2   3
dBimo E /mo þ 0:4761Bi2mo þ 0:5801Bimo þ 0:0407 þ rro2i2 L 1
4
¼ k d exp
  5 Da ¼ ð30Þ
dt R T s 0:4761Bi2 þ 0:5801Bimo þ 0:0407 þ ro2 þ T f / N
1 vf
mo ri2 mo

k r Bimo ð27Þ The values of Bi, Tf and Ts are known at certain nodes
corresponding to the initial conditions and boundary
values.
(i) Initial conditions
Bimi ði; 0Þ ¼ 0; for i ¼ 1; 2; 3; . . . ; N
(i, j+1)
Tf = TfIN Bimo ði; 0Þ ¼ 0; for i ¼ 1; 2; 3; . . . ; N
(i-1, j) (i, j) (i+1, j) (ii) Boundary conditions
Process fluid temperature Tf (i, j) is defined for each
t (i, j-1) simulation.
Integration of Eqs. (24), (26) and (27) were done
using Modified EulerÕs technique. The numerical steps
x Bi = 0 used for estimation of temperature and Biot numbers
Fig. 5. A characteristic mesh for a constant wall temperature heat with time at each node points of the heat exchanger
exchanger. are given below.
242 P.K. Sahoo et al. / Journal of Food Engineering 69 (2005) 235–244
  
1 dT f  Da

Fouling deposit on the outer

T f i þ ; j ¼ T f ð i ; jÞ þ ð31Þ

surface of inner tube, mm

dt ði;jÞ 2
1
2 0.9 Time = 15 min

   0.8 Time = 30 min

1 dBimo  Da 0.7
Bimo i; j þ ¼ Bimo ði; jÞ þ ð32Þ Time = 60 min
2 dt ði;jÞ 2 0.6
Time = 90 min
0.5
   0.4 Time = 105 min
1 dBimi  Da
Bimi i; j þ ¼ Bimi ði; jÞ þ ð33Þ 0.3
2 dt ði;jÞ 2 0 0.5 1 1.5 2 2.5
Length, m

dT f  Fig. 7. Variation of fouling deposit on the outer surface of inner tube
T f ði þ 1; jÞ ¼ T f ði; jÞ þ Da ð34Þ
dt ðiþ1;jÞ along the length of the heat exchanger with time.
2


dBimo 

inner surface of middle

1

Fouling deposit on the

Bimo ði; j þ 1Þ ¼ Bimo ði; jÞ þ Da ð35Þ
dt ði;jþ1Þ 0.9 Time = 15 min

2 0.8

tube, mm
Time = 30 min
 0.7
dBimi 
Time = 60 min
0.6
Bimi ði; j þ 1Þ ¼ Bimi ði; jÞ þ Da ð36Þ
dt ði;jþ1Þ 0.5 Time = 90 min
2
0.4 Time = 105 min

The fouling thickness in each node point is obtained cor- 0.3

0 0.5 1 1.5 2 2.5
responding to its Biot number using Eqs. (6) and (11). Length, m

Fig. 8. Variation of fouling deposit on the inner surface of the middle

tube along the length of the heat exchanger with time.
4. Simulation results and discussion

It is evident that the temperature increases along the
length of the heat exchanger towards the outlet as
shown in Fig. 6. As time progresses, the temperature
drops gradually at the respective nodes along the
length of the heat exchanger due to the occurrence of
fouling. From the simulation, a constant outlet tem-
perature of 118
C was found after 105 min of operation
of the heat exchanger. Hence, the rate of solids depo-
sition is equal to that of removal after 105 min. Also
from the experimental results as shown in Table 1, a
constant outlet temperature of 123
C has been found
after 110 min. So the simulated prediction of the temper-
ature is comparable with that of the experimental
results.
Figs. 7 and 8 display the deposit behavior of the heat
exchanger on the outer surface of the inner tube and the
inner surface of middle tube, respectively, with time of
operation. At the beginning, the occurrence of fouling
150
Time = 0 min
Temperature, °C
deposit is uniform along the length of the heat exchan-
ger. The temperature of the interface (Tfi) between the
fouling deposit and the bulk fluid is responsible for the
occurrence of fouling. In early stage of operation, as
Tfi is equivalent to the constant wall temperature (i.e.
steam temperature) of the heat exchanger, uniform
distribution of the fouling deposit occurs throughout
the length of the heat exchanger. After about 30 min,
Tfi approaches towards the bulk fluid temperature. As
the bulk fluid temperature increases towards the outlet
of the heat exchanger, the fouling deposit increases to-
wards the outlet. The trend of the fouling deposit is
same on both surfaces of the heat exchanger. The foul-
ing deposit on the inner surface of middle tube is greater
than that on the outer surface of inner tube according to
the Eqs. (6) and (11). With progress of time, as the foul-
ing deposit increases, the flow area available for the pro-
cess fluid decreases. Since UHT sterilizer runs at a
constant flow rate of 135 l/h, the velocity of process fluid
increases with decrease in flow area, which leads to in-
creased shear force. So the rate of deposit becomes

140 equal to the rate of removal after a certain time due to

Time = 15 min
130 the fluid shear forces. That means fouling deposit stabi-
Time = 30 min
120 lizes after 105 min since no net deposit occurs after that
Time = 60 min
110 time. Since milk flows at constant flow rate with inlet
Time = 90 min
100 temperature of 90
C and steam flows at constant tem-
Time =105 min
90 perature of 160
C, equilibrium temperature as well as
0 1 2 3 the point of deposit stabilization was reached after
Length, m
105 min.
Fig. 6. Variation of temperature along the length of the HTTHE with Figs. 9 and 10 shows the variation of Biot number on
time of operation. the outer surface of the inner tube (i.e. Bimi) and on the
 P.K. Sahoo et al. / Journal of Food Engineering 69 (2005) 235–244 243

simulation results, it is quite clear that the complex

Biot number on the outer

1.4
surface of inner tube

1.3 Time = 15 min fouling phenomenon is the function of the interface

1.2
1.1 Time = 30 min temperature between the fouling deposit and the bulk
1
0.9 Time = 60 min fluid and also the shear stress of the fluid on the heat
0.8 Time = 90 min exchanger surface. The simulated constant Biot num-
0.7
0.6 Time = 105 min
bers can be used for the optimum design of heat
0.5
0 0.5 1 1.5 2 2.5 exchanger.
Length, m

Fig. 9. Variation of Biot number on the outer surface of the inner tube References
along the length of the HTTHE with time.
Acrivos, A. (1956). Method of characteristic technique- Application to
heat and mass transfer problems. Industrial and Engineering
Biot number on the inner surface

1.4 Chemistry, 48, 703–710.

1.3 Time = 15 min Burton, H. (1988). Ultra high temperature processing of milk and milk
1.2 products. London: Elsevier Applied Science.
of middle tube

1.1
1
0.9
0.8
0.7
0.6
0.5
0 0.5 1 1.5 2 2.5
Length, m
Fig. 10. Variation of Biot number on the inner surface of middle tube
along the length of the HTTHE with time.
inner surface of the middle tube (i.e. Bimo), respectively,
along the length of the heat exchanger with time. From
the expressions of Biot numbers [i.e. Eqs. (6) and (11)], it
has been seen that the fouling deposit on the inner sur-
face of the middle tube in cylindrical coordinate (i.e.
ri2 ln rrd2i2 ) is greater than that of the deposit on the outer
surface of inner tube in cylindrical coordinate (i.e.
ro1 ln rrd1 ). So Biot number on the inner surface of middle
o1
tube is greater than that on the outer surface of inner-
most tube. The variation is found uniform along the
length of the heat exchanger during early periods of
operation, since Tfi is the same as the wall temperature.
Beyond 30 min, Tfi approaches towards the bulk fluid
temperature and the bulk fluid temperature gradually
increases towards the outlet of the heat exchanger. So
the Biot number increases towards the outlet of the heat
exchanger. As the fouling deposit stabilizes after
105 min, the Bimi and Bimo attain their constant values
at that time.
5. Conclusion
The performance of the heat exchanger in dairy
industry is often hampered due to severe fouling. Opti-
mum design of heat exchanger for dairy processing
is important to enhance their performance. This is
possible if fouling behavior is predicted accurately. In
this work, the fouling behavior of the heat exchanger
was predicted through a simulation using theory and
principles of fluid dynamics and heat transfer. From
Time = 30 min
Changani, S. D., Belmar-Beiny, M. T., & Fryer, P. J. (1997).
Time = 60 min
Engineering and chemical factors associated with fouling and
Time = 90 min cleaning in milk processing. Experimental Thermal and Fluid
Science, 14, 392–406.
Time = 105 min
Chopey, N. P., & Hicks, T. G. (1984). Handbook of Chemical
Engineering Calculations. New York: McGraw-Hill Book Com-
pany, p. 33, Chapter 7.
Delplace, F., Leuliet, J. C., & Levieur, D. (1997). A reaction
engineering approach to the analysis of fouling by whey proteins
of a six channels per pass plate heat exchanger. Journal of Food
Engineering, 34, 91–108.
Fryer, P. J., & Slater, N. K. H. (1985). A direct simulation procedure
for chemical reaction fouling in heat exchangers. Chemical
Engineering Journal, 31, 97–107.
Geankoplis, C. J. (1997). Transport processes and unit operations (third
ed.). New Delhi: Prentice Hall of India Pvt. Ltd., p. 239.
Jeurnink, T. J. M. (1995). Fouling of heat exchanger by fresh and
reconstituted milk and the influence of air bubbles. Milchwissens-
chaft, 50, 189–193.
Kern, D. Q. (1997). Process heat transfer. New York: McGraw-Hill
International, pp. 263–269.
Pien, J. (1955). Definition of sterilized milk, milk sterilization. FAO
Report (p. 7), 716, Rome.
Ranjan, R., & Datta, A. K. (1999). Milk fouling simulation in tubular
heat exchanger. In Proceedings 26th National Conference on fluid
mechanics and fluid power (pp. 369–374), IIT, Kharagpur.
Reitzer, B. J. (1964). Process design and development. Industrial
Engineering Chemistry, 4, 3.
Sahoo, P. K. (2001). Computer aided development and performance
evaluation of a helical triple tube Ultra-High-Temperature (UHT)
milk sterilizer. PhD thesis, IIT Kharagpur, India.
Sandu, C., & Singh, R. K. (1991). Energy increase in operation and
cleaning due to heat exchanger fouling in milk pasteurization. Food
Technology(December), 84–91.
Schreier, P. J. R., & Fryer, P. J. (1995). Heat exchanger fouling: A
model study of the scale up of laboratory data. Chemical
Engineering Science, 50(8), 1311–1321.
Srinivasan, P. S., Nandapurkar, S. S., & Holland, F. A. (1970).
Friction factors for coils. Transaction of Institution of Chemical
Engineers, 48, 156–161.
Swartzel, K. R. (1983). Tubular heat exchanger fouling by milk during
Ultra-High-Temperature processing. Journal of Food Science, 48,
1507–1511.
Taborek, J., Aoki, T., Ritter, R. B., Palen, J. W., & Knudsen, J. G.
(1972a). Fouling: The major unresolved problem in heat transfer.
Chemical Engineering Progress, 68(2), 59–67.
Taborek, J., Aoki, T., Ritter, R. B., Palen, J. W., & Knudsen, J. G.
(1972b). Predictive methods for fouling behaviour. Chemical
Engineering Progress, 68(7), 69.
244 P.K. Sahoo et al. / Journal of Food Engineering 69 (2005) 235–244

Visser, J., & Jeurnink, Th. J. M. (1997). Fouling of heat exchanger in
dairy industry. Experimental Thermal and Fluid Science, 14,
407–424.
Yoon, J., & Lund, D. B. (1994). Comparison of two operating methods
of a plate heat exchanger under constant heat flux condition and
their effects on the temperature profile during milk fouling. Journal
of Food Process Engineering, 17, 243–262.
Zuritz, C. A. (1990). On the design of triple concentric tube
heat exchangers. Journal of Food Process Engineering, 12, 113–
130.