Heat Transfer Engineering

ISSN: 0145-7632 (Print) 1521-0537 (Online) Journal homepage: http://www.tandfonline.com/loi/uhte20

The Effect of a Number of Baffles on the

Performance of Shell-and-Tube Heat Exchangers

Bassel A. Abdelkader & Syed M. Zubair

To cite this article: Bassel A. Abdelkader & Syed M. Zubair (2017): The Effect of a Number of
Baffles on the Performance of Shell-and-Tube Heat Exchangers, Heat Transfer Engineering, DOI:
10.1080/01457632.2017.1404806

To link to this article: https://doi.org/10.1080/01457632.2017.1404806

Published online: 22 Dec 2017.

Submit your article to this journal

Article views: 23

View related articles

View Crossmark data

Full Terms & Conditions of access and use can be found at

http://www.tandfonline.com/action/journalInformation?journalCode=uhte20

Download by: [UNIVERSITY OF ADELAIDE LIBRARIES] Date: 28 December 2017, At: 08:37
 HEAT TRANSFER ENGINEERING
, VOL. , NO. , –
https://doi.org/./..

The Effect of a Number of Baffles on the Performance of Shell-and-Tube Heat

Exchangers
Bassel A. Abdelkader and Syed M. Zubair
Department of Mechanical Engineering, King Fahd University of Petroleum & Minerals, Dhahran, Saudi Arabia
Downloaded by [UNIVERSITY OF ADELAIDE LIBRARIES] at 08:37 28 December 2017

ABSTRACT
The number of baffles has an impact on the thermal-hydraulic performance of a shell-and-tube heat
exchanger (STHX), thus a model was developed using Engineering Equations Solver software to
solve the governing equations. The program uses Kern, Bell-Delaware, and flow-stream analysis (Wills
Johnston) methods to predict both the heat-transfer coefficient and pressure drop on the shell side
of an STHX. It was found that Bell-Delaware method is the most accurate method when compared
with the experimental results. The effect of a number of baffles, mass flow rate, tube layout, fluid
properties and baffle cut were investigated. The analysis revealed that an increase in the number of
baffles increases both the heat-transfer coefficient and pressure drop on the shell-side. Increasing
the mass flow rate, the heat transfer coefficient increases; however, the pressure drop increases at a
higher rate. For a large number of baffles, the pressure drop decreases with an increase in the baffle
cut. It also shows that the heat transfer coefficient increases at a higher rate with the square tube
layout, whereas the rotated square and triangular layouts have approximately the same behavior.

Introduction
be between 0.4 and 0.6 of the shell diameter. Shah and
Shell-and-tube heat exchangers (STHXs) are commonly Pignotti [5] concluded that through each baffle spacing,
used in several industrial applications. This is primarily the temperature change is very small compared to the
due to the feasible construction features that comply with total temperature difference of the fluid on the shell side
the international standards, high heat transfer capability, of an STHX. Therefore, the fluid on the shell side can
ease of maintenance, and possible upgrades [1]. These be considered as uniform at every cross-section. Than
type of heat exchangers have been developed over the et al. [6] developed a model to calculate the highest heat
years and several improved configurations were proposed transfer capability of an STHX without surpassing the
for enhancing the heat transfer capability and maintain- allowable pressure drop. It was concluded that increasing
ing the allowable pressure [2]. The baffles in STHX are the baffle spacing decreases the pressure drop on the shell
primarily used to force the shell-side fluid to move in a side. However, on the tube side, if the pressure drop is not
crossflow arrangement, thus enhancing the shell-side heat within the allowable limits, increasing the total number
transfer coefficient. The secondary purpose of baffles is to of tubes or decreasing the number of passes, results in a
support the tube bundles. Therefore, the overall impact desired pressure drop on the tube side.
of the baffles is to improve the heat transfer capability but Mohammadi et al. [7] studied numerically the heat
with an increase in pressure drop [3]. transfer and pressure drop with changing baffle orien-
Chit et al. [4] studied the effect of baffle spacing on the tation in an STHX. The results showed that the baffle
shell-side flow using theoretical and numerical methods; orientation has a significant effect on the heat transfer
it was concluded that the pressure drop increases at a coefficient and pressure drop on the shell side. They
higher rate than heat transfer coefficient with a decreas- found that the vertical baffle orientation is more effec-
ing baffle spacing. They found that there is an optimal tive than the horizontal orientation with the leakage
value for baffle spacing to shell diameter ratio which flow paths. However, without leakage, the simulation
results in the highest heat transfer for a specified pressure results showed that the horizontal baffle orientation is
drop on the shell side. The optimum ratio is found to more effective than the vertical one, mainly in the inlet

CONTACT Professor Syed M. Zubair [email protected] Department of Mechanical Engineering, KFUPM Box # , King Fahd University of Petroleum
& Minerals, Dhahran , Saudi Arabia.
Color versions of one or more of the figures in this paper can be found online at www.tandfonline.com/uhte.
©  Taylor & Francis Group, LLC
 2 B. A. ABDELKADER AND S. M. ZUBAIR

and outlet regions. Jozaei et al. [8] investigated the heat tube layout. Therefore, the objective of this paper is to
transfer and pressure drop with a number of baffles in a investigate the performance of an STHX using three dif-
segmental baffled STHX for a specified heat duty. It was ferent design approaches. In this regard, first, the results
concluded that at low baffle spacing, pressure drop and are compared with the experimental data available in the
heat transfer coefficient are high; however, U/P ratio is literature. This is followed by investigating in detail the
low. This is primarily due to the fact that effect of baffle effect of mass flow rate, a number of baffles, baffle cut,
spacing has a higher impact on pressure drop compared various hydrocarbon fluid properties and tube layout on
to the shell-side heat transfer coefficient. As baffle spacing the heat transfer coefficient, pressure drop as well as the
increases, the pressure drop decreases at a higher rate heat transfer coefficient to the pressure drop ratio.
than the overall heat transfer coefficient; therefore, U/P
ratio increases. Jozaei et al. [9] in another study inves-
Design approach
tigated the effect of baffle spacing on the heat transfer,
pressure drop, and cost of STHX. They concluded that It is worth mentioning that while designing STHXs, the
Downloaded by [UNIVERSITY OF ADELAIDE LIBRARIES] at 08:37 28 December 2017

the optimal baffle spacing ranges between 43 to 63% of design specifications must be identified as completely as
inside shell diameter for the required heat duty. possible. Besides, the fluid flow rates, inlet and outlet
Li and Kottke [10] investigated the effect of baffle temperatures, and pressure constraints, the exact require-
spacing on the heat transfer and pressure drop in an ments of the process engineer as well as the past experi-
STHX for different tube arrangements. It was shown that ence needed for the design engineer must be discussed
heat transfer and pressure drop increase with the baffle in detail. For example, predicting the overall heat trans-
spacing for a fixed Reynolds number since the leakage is fer coefficient requires calculation of the tube-side and
reduced through the baffle-shell clearance. Gaddis and shell-side heat transfer coefficients. For tube-side flow in
Gnielinski [11] developed a model for calculating the an STHX, Nusselt number and friction factor correlations
pressure drop on the shell side with the segmental baffles. from the available literature can be used depending on the
The results showed that the expected deviations between flow conditions, as is typically done in a thermal design
the experimental and theoretical predictions are within problem. The shell-side analysis for both heat transfer and
35%. Petinrin and Dare [12] studied the effect of tube pressure drop analysis is normally studied using three dif-
layout on the performance of STHX using continuity, ferent approaches.
momentum, and energy equations. The results showed
that the bulk of heat transfer and pressure drop occur
Kern method
during the crossflow of shell-fluid fluid over the tube bun-
dle. For a given shell-side pressure drop, the triangular Kern method [1] is mostly used for a preliminary design
layout (30°) is found to be the most suitable followed by for an STHX. It typically provides conservative results. If
the combined (mix between 30° and 60°), and the rotated there are no baffles, the flow will be along the tubes that
triangular (60°), respectively, Deshpande and Hinge [13] are inside the shell. Thus, the heat transfer coefficient can
developed a model to design a single segmental STHX be calculated from the equivalent diameter. However, the
with a vertical cut arrangement and compared it with a baffles increase the shell-side heat transfer coefficient sig-
horizontal cut orientation. Using Kern method [1], they nificantly due to an increase in the turbulence. In a baffled
concluded that changing the flow pattern to a vertical cut STHX, the velocity of the fluid varies because of the con-
does not improve the shell-side heat transfer coefficient. fined area between the tubes across the tube bundle. The
Vukić et al. [14] investigated experimentally the effect correlations obtained for flow in tubes are not applicable
of a number of segmental baffles on the performance for flow over the tube bundles with segmental baffles.
of STHXs in a laboratory experimental unit. The results McAdams [15] suggested the following correlations for
showed that the heat transfer depends on the shell-side heat transfer coefficient on the shell side:
geometry (number of segmental baffles, baffle cut size,  0.14
baffle distance, the first and last baffle position to the μb
Nu = 0.36Re0.55 Pr1/3 (1)
inlet and outlet nozzle, the size of constructive clear- μw
ances). Increasing the hot fluid flow rate decreases the
temperature drop across the baffle. They also found that for
an increase in the number of baffles has a higher influ-
ence on the STHX effectiveness than the increase in the 2 × 103 < Re < 1 × 106
hot fluid flow rate. Pr > 0.6
It is obvious from the above discussion that the per-
formance of an STHX depends on the mass flow rate, For calculating the pressure drop due to fluid fric-
number of baffles, baffle cut, fluid properties, and the tion, excluding the nozzle losses, is computed using the
 HEAT TRANSFER ENGINEERING 3

following equation [1]

f G2 Ds (Nb + 1)
Ps =   (2)
7.5 × 1012 De μμwb

f = 0.4137Re−0.2585 (3)

Bell-Delaware method
Bell-Delaware method [2] recognizes that there is only a
portion of the fluid on the shell side in a baffled STHX
flows across the tubes in a pure crossflow arrangement;
Downloaded by [UNIVERSITY OF ADELAIDE LIBRARIES] at 08:37 28 December 2017

that is, normal to the axis of the tubes. While the remain-
ing fluid flows through the bypass areas. This is due to Figure . Schematic of tube bundle bypass stream [].
the fact that the fluid seeks a path of least resistance from
the inlet to the outlet. In a typical design, the non-ideal Shell-side heat transfer coefficient
flow streams represent up to 40% of the total flow rate
The heat transfer coefficient on the shell side, hs , is
and hence is important to account for these effects on the
obtained by multiplying ideal crossflow heat transfer coef-
heat transfer and pressure drop [2]. It should be noted that
ficient, hc , by correction factors for the nonideal cross flow
the total flow is divided into several streams, as shown in
in a baffled heat exchanger [2],
Figures 1 to 3. These flow streams include the tube hole
leakage stream, crossflow stream, the tube bundle bypass
hs = hc JC JL JB JR JS Jμ (4)
stream, the shell-to-baffle bypass stream, and the pass par-
tition bypass stream.
Baffle window flow correction factor
The baffle window correction factor accounts for the
nonideal crossflow effects through the window zone
because the velocity of these streams is different. It
mainly depends on the baffle spacing and a baffle cut.
The baffle cut correction factor is calculated from the
following expressions [2],

JC = 0.55 + 0.72(1 − 2Fw ) (5)

θctl sin θctl
Fw = − (6)
360 2π
   
Ds BC
θctl = 2cos−1 1−2 (7)
Dctl 100

where Fw is a fraction of the cross-sectional area of the

window, BC is the baffle cut, Ds is the shell diameter, Dctl
is the centerline tube limit diameter, and θ ctl is the angle
between the centerline and baffle cut.

Baffle leakage correction factor

The fraction of flow through the gaps between the baffle
and shell, and the baffle and tube diameter, which impact
the heat transfer coefficient and pressure drop, can be
Figure . Shell-side flow paths in a baffled heat exchanger []. calculated by [2],
 4 B. A. ABDELKADER AND S. M. ZUBAIR

Figure . Diagram of leakage streams; (a) hole leakage, and (b) shell-to-baffle bypass [].
Downloaded by [UNIVERSITY OF ADELAIDE LIBRARIES] at 08:37 28 December 2017

    
Ssb Ssb the distance between shell wall and tube bundle, and Nss
JL = 0.44 1 − + 1 − 0.44 1 −
Ssb + Stb Ssb + Stb is the number of pairs of sealing strips.
 
Ssb + Stb
× exp −2.2 (8)
Sm
Unequal baffle spacing correction factor
Ssb = 0.00436Ds Lsb (360 − θds ) (9)

π 
Stb = [(Dt + Ltb )2 ] Nt (1 − Fw ) (10) The effect of inlet and outlet baffle spacing larger than
4  the central baffle spacing is accounted in the unequal baf-
Dctl fle spacing correction factor. When the inlet and outlet
Sm = Lbc Lbb + (Lt p − Dt ) (11)
Lt p,e f f spacing is larger than central baffle spacing, the velocity is
  
BC lower which has an adverse effect on heat transfer. When
θds = 2cos−1 1 − 2 (12)
100 the inlet and outlet spacing is equal to central spacing
JS = 1. Otherwise, it is calculated from [2],
where Ssb is the shell-to-baffle leakage area, Stb is the
tube-to-baffle leakage area, Sm is the crossflow area at the (Nb − 1) + (Lbi /Lbc )1−n + (Lbo/Lbc )1−n
center line, θ ds is the baffle cut angle in degrees, Lsb is JS = (17)
(Nb − 1) + (Lbi /Lbc ) + (Lbo/Lbc )
the shell-to-baffle clearance, Lbc is the baffle spacing, and
Ltp,eff is the tube pitch for 30° and 90° layouts while for For laminar flow n = 1/3 while for transition and tur-
45° is equal to 0.707 of the tube pitch. bulent flow n = 0.6.

Tube bundle bypass correction factor

Laminar flow correction factor
As stated earlier, the fluid seeks the flow of least resis-
The laminar flow correction factor accounts for the
tance from the inlet to the outlet. Therefore, a portion of
adverse temperature gradient formed in the boundary
the fluid flows through bypass areas, the bypass correc-
layer.
tion factor accounts for this effect. This correction factor
can be minimized by reducing the space between the shell  0.18    0.18 
10 20 − Re 10
wall and tube bundle or by placing sealing strips. It is cal- JR = + −1 (18)
Nc 80 Nc
culated from the following expressions [2],
 √  where
JB = exp − Cbh Fsbp 1 − 3 2rss (13)
Sb Nc = (Ntcc + Ntcw )(Nb + 1) (19)
Fsbp = (14)
Sm
Sb = Lbc [(Ds − Dotl ) + L pl ] (15) and
   
Nss 0.8 BC (Ds − Dctl )
rss = Ds  BC  (16) Ntcw = Ds − (20)
L pp
1 − 2 100 L pp 100 2

For laminar flow, Cbh is equal to 1.35 while for tran- Here Nc is the total number of tube rows crossed by the
sition and turbulent, Cbh is equal to 1.25. In the above flow, Ntcw is the number of tube rows in the window area,
equations, Sb represents the bypass area, Lpl represents and Ntcc is the number of tube rows between the baffle tips.
 HEAT TRANSFER ENGINEERING 5

Table . Empirical coefficients for ji and fi [].

Layout Re a1 a2 a3 a4 d1 d2 d3 d4

°  – . − . . . . − . . .
 – . − . . − .
 – . − . . − .
 – . − . . − .
< . − . . − .
°  – . − . . . . − . . .
 – . − . . − .
 – . − . . − .
 – . − . . − .
< . − . . − .
°  – . − . . . . − . . .
 – . − . . +.
 – . − . . − .
 – . − . . − .
Downloaded by [UNIVERSITY OF ADELAIDE LIBRARIES] at 08:37 28 December 2017

< . − . . − .

Wall viscosity correction factor Ps = Pcb + Pw + Pe (25)

Heat transfer correlations are typically estimated with where Pcb is pressure drop in central baffle spaces, Pw
properties at the mean temperature; however, the tem- is pressure drop in baffle windows, and Pe is pressure
perature changes from the bulk to the wall. The effect of drop in the entrance and exit baffle spaces.
variation in properties between the bulk and wall fluid
temperatures is corrected by the viscosity ratio [2],
 m
μb
Jμ = (21)
μw

The exponent m for heating and cooling of liquid is

usually 0.14, while for gases no correction is required.

Ideal crossflow heat transfer coefficient

The ideal crossflow heat transfer coefficient is calculated
from the correlation [2],
−2/3
hc = ji c pG Pr (22)

where
 a
1.33
ji = a1 Rea2 (23)
Lt p/Dt
a3
a= (24)
1 + 0.14Rea4
where Reynolds number is based on tube diameter and
the values of constants a1 , a2 , a3 , and a4 are listed in
Table 1.

Pressure drop calculation

The pressure drop across the shell is composed of three
parts; the entrance and exit baffle spaces, all central baffle
spaces, and all baffle windows, as shown in Figure 4. Figure . Pressure drop regions in shell-side flow [].
 6 B. A. ABDELKADER AND S. M. ZUBAIR

Pressure drop in central baffle spaces Pressure drop in the entrance and exit baffle spaces
The pressure drop across tube bundle between the baffles The pressure drop in the entrance and exit baffle is calcu-
is based on ideal tube bundle pressure drop (PbI ) for one lated from Eq. 40, while the pressure drop correction fac-
baffle section [2], tor for unequal baffle spacing is calculated from Eq. 41.
These are [2],
Pcb = PbI (Nb − 1)RB RL (26)  
Ntcw
G2 Pe = PbI 1 + RB RS (40)
PbI = 0.002 fi Ntcc Rμ (27) Ntcc
ρ  2−nc  2−nc
 d Lbc Lbc
1.33 RS = + (41)
fi = d1 Red2 (28) Lbo Lbi
Lt p/Dt
d3
d= (29) Flow stream analysis
1 + 0.14Red4
Downloaded by [UNIVERSITY OF ADELAIDE LIBRARIES] at 08:37 28 December 2017

The flow stream analysis presented in this section is based

where PbI is the ideal bundle pressure drop for one baf- on the analysis proposed by Wills and Johnston [1], which
fle compartment, fi is the friction factor, and the empiri- is basically a simplified version of Palen and Taborek [16]
cal constants d1 , d2 , d3 , and d4 are listed in Table 1. The work. In this method, flow rates and pressure drop is
viscosity correction factor (Rµ ), leakage correction fac- calculated by using a basic hydraulic model. The flow in a
tor (RL ), and bypass correction factor (RB ) was calculated typical STHX is divided into the cross flow, leakage, and
from the following correlations [2], bypass streams, as shown in Figure 5 The heat transfer
coefficient on the shell-side is calculated from actual


μb m crossflow rate on the shell-side [1]. It should be noted
Rμ = (30) that the pressure drop between any two points is same
μw
 √  regardless of paths connecting these points. The fluid
RB = exp −Cbp Fsbp (1 − 3 2rss ) (31)
 p 
flows from A to B through various routes, each designated
RL = exp −1.33(1 + rs )rlm (32) by a subscript. For example, leakage flows occur between
p = −0.15(1 + rs ) + 0.8 (33) the tubes and the baffles (t) and between the baffle and
the shell (s). Some of the flow passes over the tubes in a
cross flow (c), and part bypasses the tube bundle (b). The
Pressure drop in baffle windows crossflow and bypasses streams combine to form a further
stream (w), which passes through the window zone. For
For turbulent flow, the pressure drop in the baffle window
each of these streams, a flow coefficient ni is defined as
is calculated from Eq. 34, while for laminar flow pressure
drop is calculated from Eq. 35. The first term refers to the Pi = ni ṁ2i (42)
cross flow and the second term is the longitudinal flow.
where Pi is the pressure drop for stream “i” and ṁi is
The hydraulic diameter and window area are calculated
the mass flow rate of the stream. Combined coefficients
from Eqs. 37 and 39. All these equations are summarized
na , np , and ncb can be defined such that
as [2],
  P = na ṁ2w (43)
0.001G2w
Pw = Nb (2 + 0.6Ntcw ) RL Rμ (34)
2ρ
   
Gw μ Ntcw Lbc
Pw = Nb 26 + 2
ρ Lt p − Dt Dw
 2 
0.002Gw
+ RL Rμ (35)
2ρ
ṁ
Gw = √ (36)
Sm Sw
4Sw
Dw = (37)
πDt Ntw + (πDs θds /360)
Ntw = Ntt Fw (38)
  π 
πD2s θds sin θds
Sw = − − Ntw D2 (39) Figure . Equivalent shell side flow.
4 360 2π 4 t
 HEAT TRANSFER ENGINEERING 7

P = n p ṁtot
2
(44) co(Ds − 2Lc )/PT P + Nss
nb = (60)
Pc = Pb = ncb ṁ2w (45) 2ρSb2

Therefore, from series–parallel flow configurations where

(refer to Figure 5), one can easily write as, Sb = (2δby Ds + δ pp )Lbc (61)
Pb = Pc (46) The crossflow resistance coefficient varies with the flow
and rate. It is expressed as
Nc K f
PAB = Ps = Pt = Pw + Pb (47) nc = (62)
2ρSm 2
The conservation of mass requires the sum of flow rates
The parameter Kf, for in-line square arrays, is given as
of all the streams to be the total mass flow rate of the fluid
on the shell side, which is fixed. It gives, 0.207 × 103 0.102 × 103
K f = 0.272 + +
Downloaded by [UNIVERSITY OF ADELAIDE LIBRARIES] at 08:37 28 December 2017

Re Re2
ṁtot = ṁb + ṁc + ṁs + ṁt (48)
0.286 × 10 3
− for 3 < Re < 2 × 103 (63)
In this method, it is assumed that the flow coefficients Re3
ns , nt, nw , and nb are constants independent of flow rate 0.249 × 104 0.927 × 107
and are dependent only on geometry, thus Fcr and nc are K f = 0.267 + −
Re Re2
calculated from Eqs. 47 and 54. Here ns is shell-to-baffle 0.10 × 10 11
leakage resistance coefficient, nt is tube-to-baffle resis- + for 2 × 103 < Re < 2 × 106
Re3
tance coefficient, nw is window flow resistance coefficient, (64)
and nb is bypass flow resistance coefficient. These are cal-
culated by the following expressions [1], The flow-stream analysis method does not specifically
deal with the heat transfer; however, a reasonable estimate
ṁc Dt
Re = (49) of the shell-side heat transfer coefficient can be made by
μbSm assuming that it corresponds to that calculated by using
ṁc = Fc ṁs (50) ṁc rather than ṁs in a standard crossflow correlation,
 0.5
np  0.14
na 0.651 0.34 μb
Fc =   0.5  (51) Nu = 0.211Re Pr (65)
μw
1 + nnbc
for
na = nw + ncb (52)
2 3 × 102 < Re < 2 × 105 , Pr < 600
ncb = 1/ n−0.5
c + n−0.5
b (53)
2 while the pressure drop across the shell by neglecting
n p = 1/ n−0.5
a + n−0.5
s + nt−0.5 (54)
the inlet and exit nozzles is calculated by,
The values of flow coefficients ns , nt , nw , and nb are
P = n pṁtot
2
(Nb + 1) (66)
considered constant for a given flow configuration and are
given by [1],
Calculation procedure
0.036(tb/δsb ) + 2.3(tb/δsb )−0.177
ns = (55)
2ρSs 2 To study the performance of STHXs, the fluid flow rates,
inlet and outlet temperatures, and geometry must be
where
known. With this information, the performance of STHX
Ss = π (Ds − δsb )δsb (56) can be investigated by three different methods discussed
0.036(tb/δtb ) + 2.3(tb/δsb )−0.177 in the previous section (Kern – Eqs. 1 to 3, Bell-Delaware
nt = (57) – Eqs. 4 to 41, and flow-stream analysis – Eqs. 42 to 66).
2ρSt 2
Thus, the following results can be calculated from these
where
methods:
St = Nt π (Ds + δtb )δtb (58) 1. heat transfer coefficients on the shell side and tube
side
The window and bypass flow coefficients are, respec-
2. shell side and tube side pressure drop
tively, given as
An Engineering Equation Solver (EES) based model
1.9 exp(0.6856Sw /Sm ) was developed for flow through a segmental baffled STHX
nw = (59)
2ρSw 2 to investigate the effect of a number of baffles on shell-side
 8 B. A. ABDELKADER AND S. M. ZUBAIR

heat-transfer coefficient and pressure drop using the three

methods discussed above. The EES program was divided
into five main steps.
a) calculation of fluid properties as well as heat trans-
fer and fluid flow areas,
b) calculation of heat transfer coefficient on shell-side
and tube-side,
c) calculation of overall heat-transfer coefficient,
d) calculation of heat duty and outlet fluid tempera-
tures, and
e) calculation of pressure drop for shell-side.
Downloaded by [UNIVERSITY OF ADELAIDE LIBRARIES] at 08:37 28 December 2017

Results and discussion

The goal of this study is to investigate the effect of a
number of baffles, baffle cut, mass flow rate, and tube
layout on the pressure drop and heat transfer coefficient
in an STHX. To study these parameters an EES model was
developed using Kern, Bell-Delaware, and flow stream
analysis methods.

Comparison between different methods

To establish the credibility of the results presented in this
paper, the experimental data of STHX, type 1–2U, inves- Figure . Heating fluid outlet temperature for BC = % and inlet
tigated by Vukić et al. [14] is used. The heat exchanger tube temperature = °C; (a) Nb = , and (b) Nb = .
data is presented in Table 2. They reported the accu-
racy of the water-flow measuring system in the range of to the other two methods. It is important to emphasize
0.029–0.200%. For measuring the pressure drop, ori- that all possible effects are considered in this method;
fice accuracy was in the range of 0.704–0.819%, and particularly unequal baffle spacing is only considered in
the accuracy of thermocouple measurements was in the Bell-Delaware approach.
the range of 0.143–0.769%. Figure 6 presents the tube
outlet temperatures for two different number of baffles Effect of number of baffles
(1 and 5) calculated from the heat duty and overall
heat transfer coefficient. Figure 7 shows the total shell It is obvious that increasing the number of baffles
side pressure drop comparison with a single baffle case. increases the heat transfer coefficient and pressure drop
The experimental results are compared with the results on the shell side. However, the increase in heat-transfer
obtained from the developed EES model. It can be seen
that Bell-Delaware method provides very close results
(within 5%) of the experimental data when compared

Table . Heat exchanger data [].

Parameter Shell Side Tube Side

Fluid Water Water

Material Carbon Steel Copper
Inlet Diameter (mm) . 
Outlet Diameter (mm) . 
Inlet Temperature °C °C
Volume Flow Rate (m /h)  
Tube Layout 
Number of Tubes 
Number of Passes 
Baffle Cut % Figure . Total shell side pressure drop increase for BC = % and
Tube Pitch  mm Nb = .
 HEAT TRANSFER ENGINEERING 9

spacing (small number of baffles), the three methods

are in a good agreement; however, as the baffle spacing
decreases the Kern method deviates from the other two
approaches.

Effect of mass flow rate

The shell-side mass flow rate has a direct impact on
Reynolds number and, thus on convective heat transfer
coefficient as well as pressure drop. Figure 10 illustrates
Downloaded by [UNIVERSITY OF ADELAIDE LIBRARIES] at 08:37 28 December 2017

Figure . The impact of a number of baffles using different meth-

ods; (a) shell-side heat transfer coefficient, and (b) shell-side pres-
sure drop.

coefficient using the Bell-Delaware method is less than

the one calculated using Kern and flow-stream analysis
approaches, as shown in Figure 8(a). This is primarily due
to the effect of unequal baffle spacing. As the number of
baffles increases, the baffle spacing to inlet spacing ratio
decreases, thus decreasing the unequal baffle spacing
correction factor (Js ) and baffle leakage correction factor
(JL ). However, the ideal crossflow heat transfer coefficient
increases as shown in Figure 9. Figure 8(b) represents the
effect of a number of baffles on the shell-side pressure
drop. Increasing the number of baffles increases the pres-
sure drop on the shell side. For example, at large baffle

Figure . Effect of mass flow rate and a number of baffles; (a)
Figure . Variation of unequal baffle spacing, baffle leakage, and on shell-side heat transfer coefficient, (b) on shell-side pressure
the ideal crossflow heat transfer coefficient with a number of drop, and (c) on the heat transfer coefficient per unit pressure drop
baffles. (h/P).
 10 B. A. ABDELKADER AND S. M. ZUBAIR

Effect of baffle cut

The baffle cut usually varies between 15 to 45% of the
shell inside diameter. It is important to emphasize that
very small and very large baffle cuts are unfavorable to
the shell-side heat transfer coefficient. This is due to the
significant deviation from an ideal cross-flow situation.
The optimum value of baffle cut varies between 20 to
35%. Figure 11 illustrates the effect of baffle cut on the
shell-side heat transfer coefficient and pressure drop. It is
found that heat transfer coefficient as well as the pressure
drop on the shell-side decreases with increasing baffle
cut. For a small number of baffles, the baffle cut has a
Downloaded by [UNIVERSITY OF ADELAIDE LIBRARIES] at 08:37 28 December 2017

small effect on the pressure drop, while for a large num-

ber of baffles, the pressure drop decreases significantly.
Furthermore, the heat-transfer coefficient with baffle cut
has the same trend with the number of baffles. Thus, for
a small number of baffles, the h/P ratio increases with
increasing baffle cut, while for a large number of baffles,
there is a small effect on the h/P.

Effect of baffles with different fluids

Fluid properties have an impact on Reynolds number,
the rate of heat transfer, and pressure drop. Figure 12
illustrates the effect of a number of baffles on heat trans-
fer coefficient and pressure drop on the shell side with
different fluids. Using Bell-Delaware method, it is found
that increasing number of baffles will increase heat trans-
fer coefficient and pressure drop. Six fluids (Methane,
Propane, Ammonia, Butane, Ethylene, and Propylene)
were compared and it was found that propylene has the
least heat transfer coefficient and pressure drop, while
methane has the highest heat transfer coefficient and
Figure . Effect of baffle cut and number of baffles; (a) on shell- pressure drop. However, ammonia has the least h/P
side heat transfer coefficient, (b) on shell-side pressure drop, and ratio.
(c) on the heat transfer coefficient per unit pressure drop (h/P).

Effect of number of baffles with different tube layout

the effect of mass flow rate on heat transfer coefficient,
pressure drop, and heat transfer coefficient per unit There are three typical tube layouts; they are triangular
pressure drop on the shell side. Using the Bell-Delaware (30°), rotated square (45°), and square (90°). A trian-
method, the results showed that both heat transfer coef- gular layout accommodates more tubes than a square
ficient and pressure drop increases with increasing mass and rotated square layout. Moreover, triangular layout
flow rate. At a small number of baffles, the mass flow increases turbulence, thus, higher heat-transfer coef-
rate has a small effect on the pressure drop, while at a ficient is expected. However, the triangular layout is
large number of baffles the pressure drop increases sig- difficult to clean, therefore for high fouling fluid on the
nificantly with mass flow rate; however, the heat transfer shell side, a square layout is normally used, but it pro-
coefficients with mass flow rate have the same trend with duces lower turbulence. Therefore, when the shell-side
a number of baffles. Thus, for a small number of baffles, Reynolds number is low, it is better to use a rotated square
the h/P ratio decreases with increasing mass flow rate, layout because this produces much higher turbulence,
whereas the mass flow rate has a small effect on the h/P which results in a higher efficiency of conversion of
with increasing number of baffles. pressure drop to heat transfer.
 HEAT TRANSFER ENGINEERING 11
Downloaded by [UNIVERSITY OF ADELAIDE LIBRARIES] at 08:37 28 December 2017

Figure . Effect of number of baffles with different fluids; (a) on

shell-side heat transfer coefficient, (b) on shell-side pressure drop,
and (c) on the heat transfer coefficient per unit pressure drop Figure . Effect of a number of baffles with different tube layouts;
(h/P). (a) on shell-side heat transfer coefficient, (b) on shell-side pressure
drop, and (c) on the heat transfer coefficient per unit pressure drop
Figure 13 illustrates the effect of tube layout on the (h/P).
heat transfer coefficient and pressure drop. Using the
Concluding remarks
Bell-Delaware method, the results showed that for a
small number of baffles, the triangular layout has the The current study presents the heat transfer coefficient
highest heat transfer coefficient while the square layout and pressure drop calculations for STHXs using Kern,
has the least; however, for a large number of baffles, the Bell-Delaware, and flow-stream analysis (Wills Johnston)
square layout has the highest heat transfer coefficient. methods. Bell-Delaware method found to be the most
It can be seen from the figure that triangular layout has reliable one when compared with the experimental data.
the highest pressure drop while the rotated square has The results show that the performance of STHX strongly
the least pressure drop. However, rotated square has the depends on the number of baffles, baffle cut size, tube lay-
highest h/P ratio for a small number of baffles, whereas out, fluid properties, and mass flow rate.
when the number increases, the square layout has the It is found that increasing the number of baffles,
highest h/P ratio. increases the heat transfer coefficient and pressure drop
 12 B. A. ABDELKADER AND S. M. ZUBAIR

on the shell-side, while it decreases the value of correc- ji heat transfer factor in Eq. 22, given by Eq. 23
tion factor due to unequal baffle spacing. The pressure Kf parameter in Eq. 62
drop increases at a higher rate with the number of baf- Lbb bypass channel diametric gap, m
Lbc central baffle spacing, m
fles. It should be noted that increasing mass flow rate Lbi inlet baffle spacing, m
increases velocity which enhances the heat transfer rate, Lbo outlet baffle spacing, m
but it also increases the pressure drop. However, the shell- Lc baffle cut distance, m
side heat transfer coefficient decreases with baffle cut. For Lpl width of the bypass lane between the tubes, m
a large number of baffles, pressure drop decreases with Lpp horizontal tube pitch, m
Lsb shell-to-baffle clearance, m
the baffle cut at a higher rate than for a small number of
Ltp tube pitch, m
baffles. Ltp,eff effective tube pitch, m
Six fluids were compared; it is found that propylene ṁ flow rate, kg/s
has the least heat transfer coefficient and pressure drop, m exponent in Eqs. 21 and 30, for heating and cooling of
while methane has the highest heat transfer coefficient liquid, m = 0.14
Downloaded by [UNIVERSITY OF ADELAIDE LIBRARIES] at 08:37 28 December 2017

and pressure drop. However, ammonia has the smallest Nb number of baffles
Nc total number of tube rows in cross flow
h/P ratio. For a small number of baffles, the triangular Nss number of sealing strips
layout has the highest heat transfer coefficient while the Nt number of tubes
square layout has the lowest; however, for a large number Ntcc number of tube rows crossed between baffle tips in one
of baffles, the square layout has the highest heat transfer baffle section
coefficient. The triangular layout has the highest pressure Ntcw number of tube rows in the window
Ntt total number of tubes
drop while the rotated square has the least pressure drop.
Ntw number of tubes in the window
Nevertheless, rotated square has the highest h/P ratio Nu Nusselt number
for a small number of baffles, whereas, for a large number n exponent in Eq. 17, for laminar flow n = 1/3 while for
of baffles, the square layout has the highest h/P ratio. transition and turbulent flow n = 0.6.
nc exponent in Eq. 41, for laminar flow n = 1 while for
transition and turbulent flow n = 0.2.
Nomenclature na combined flow coefficient in Eq. 43, given by Eq. 52 in
kg−1 .m−1
A tube hole leakage stream, refer to Figure 1 nb bypass flow resistance coefficient, kg−1 .m−1
a exponent in Eq. 23, given by Eq. 24 nc crossflow resistance coefficient, kg−1 .m−1
B crossflow stream, refer to Figure 1 ncb combined flow coefficient in Eq. 45, given by Eq. (53)
BC baffle cut in kg−1 .m−1
b flow bypasses the tube bundle ni flow coefficient, kg−1 .m−1
C bundle by pass stream, refer to Figure 1 np combined flow coefficient in Eq. 44, given by Eq. 54 in
Cbh empirical factor in Eq. 13 kg−1 .m−1
Cbp empirical factor in Eq. 31 ns shell-to-baffle leakage resistance coefficient, kg−1 .m−1
c flow passes over the tubes in a cross flow nt tube-to-baffle clearance resistance coefficient,
co constant in Eq. 60, co = 0.266 for square tube lay- kg−1 .m−1
out and 0.133 for triangular, rotated triangular, and nw window flow resistance coefficient, kg−1 .m−1
rotated square layouts Pr Prandtl number
cp specific heat capacity, J.kg−1 .K−1 PTP spacing between tube rows in the flow direction
D diameter, m P pressure drop, Pa
Dotl outer tube limit diameter, m PAB pressure drop from point A to point B, Pa
Dt tube diameter, m Pcb pressure drop in central baffle spaces, Pa
Dw hydraulic diameter of the baffle window, m Pe pressure drop in entrance and exit baffle spaces, Pa
d exponent in Eq. 28, given by Eq. 29 Pw pressure drop in baffle windows, Pa
E shell to baffle bypass stream, refer to Figure 1 p parameter in Eq. 33
F pass partition bypass stream, refer to Figure 1 R correction factor for pressure drop
f friction factor Re Reynolds number
fi friction factor for ideal bundle pressure drop rlm shell-and-tube to baffle leakage area to crossflow area
Fc fraction of total flow over the tube bundle in a cross at the bundle center line
flow rs shell-to-baffle leakage area to shell-and-tube to baffle
Fsbp bypass to crossflow area ratio leakage area ratio
Fw fraction of the cross-sectional area occupied by the rss number of sealing strips to a number of tube rows
window crossed between baffle tips
G mass flux, kg.s−1 .m−2 S leakage area, m2
Gw mass flux in the window area, kg.s−1 .m−2 Sb bypass area, m2
h heat transfer coefficient, W.m−2 .K−1 Sm crossflow area at the center line, m2
J correction factor for heat transfer coefficient Ss shell-to-baffle leakage area, m2
 HEAT TRANSFER ENGINEERING 13

St tube-to-baffle leakage area, m2 Notes on contributors

Sw window flow area, m2
STHX shell-and-tube heat exchanger Bassel A. Abdelkader is an M.Sc.
s leakage flow between baffle and shell student in Mechanical Engineer-
T temperature, °C ing Department at King Fahd
t leakage flow between tubes and baffle University of Petroleum & Min-
tb baffle thickness, m erals (KFUPM), Saudi Arabia.
U overall heat transfer coefficient, W.m−2 K−1 He earned his Bachelor degree
V̇ volumetric flow rate, m3 .h−1 from Institute of Aviation Engi-
w crossflow and bypass streams combined neering & Technology, Egypt in
2011.

Greek symbols
θ angle, in degrees
Downloaded by [UNIVERSITY OF ADELAIDE LIBRARIES] at 08:37 28 December 2017

θds baffle cut angle, in degrees

ρ density, kg.m−3 Syed M. Zubair is a Distin-
μ viscosity, N.s.m−2 guished Professor in Mechanical
δby bundle-to-shell radial clearance, m Engineering Department at King
δ pp radial clearance associated with an in-line pass partition, Fahd University of Petroleum &
m Minerals (KFUPM). He earned
δsb shell-to-baffle radial clearance, m his Ph.D. degree from Georgia
δtb tube-to-baffle radial clearance, m Institute of Technology, Atlanta,
Georgia, USA, in 1985. He has
participated in several externally
and internally funded research
Subscripts projects at KFUPM and has
published over 200 research
b bulk fluid or bypasses the tube bundle papers in internationally ref-
bI ideal tube bundle ereed journals. Due to his various activities in teaching and
B bypass research, he was awarded Distinguished Researcher award by
C baffle cut the university in academic years 1993–1994, 1997–1998, and
c cross flow 2005–2006 as well as Distinguished Teacher award in academic
cb central baffle spaces years 1992–1993 and 2002–2003. In addition, he received best
ctl centerline tube Applied Research award on Electrical and Physical Properties
e equivalent of Soils in Saudi Arabia from GCC-CIGRE group in 1993.
i index in Eq. 42, representing flow stream
L leakage
μ viscosity
R laminar flow References
S unequal baffle spacing
[1] R. W. Serth, and T. Lestina, Process Heat Transfer: Princi-
s shell or between the baffle and shell
ples, Applications and Rules of Thumb. 2nd edition, Ams-
sb shell-to-baffle
terdam: Elsevier Science Publishers, 2014.
tot total
[2] J. R. Thome, Engineering Data Book III. Decatur, AL:
t between the tube and baffle
Wolverine Tube, Division of UOP Inc., 2004.
tb tube-to-baffle
[3] G. B. Eke, and C. A. Ebieto, “Performance analysis of shell
w wall or window
and tube heat exchangers: A case study,” J. Emerg. Trends
Eng. Appl. Sci., vol. 3, no. 5, pp. 899–903, 2012.
[4] S. P. Chit, N. A. San, and M. M. Soe, “Flow analysis in
shell side on the effect of baffle spacing of shell and tube
Acknowledgments heat exchanger,” Int. J. Sci. Technol. Soc., vol. 3, no. 5,
The authors acknowledge the support provided by King Fahd pp. 254–259, 2015. doi:10.11648/j.ijsts.20150305.15.
University of Petroleum & Minerals through the project [5] R. K. Shah, and A. Pignotti, “Influence of a finite num-
IN151001. We also gratefully acknowledge the graphics work ber of baffles on shell-and-tube heat exchanger perfor-
carried out by Mr. Muhammad A. Jamil, Graduate Student in mance,” Heat Transf. Engng., vol. 18, no. 1, pp. 82–94, 1997.
ME Department at KFUPM, in preparing Figures 1 to 5. doi:10.1080/01457639708939891.
[6] S. T. M. Than, K. A. Lin, and M. S. Mon, “Heat
Exchanger Design, WASET,” Intl. J. of Mech., Aero. Indus.
Funding Mecha. and Manuf. Engng., vol 2, no. 10, pp. 1151–1158,
2008.
King Fahd University of Petroleum & Minerals through the [7] K. Mohammadi, W. Heidemann, and H. Müller-
project # 151001. Steinhagen, “Numerical investigation of the effect of
 14 B. A. ABDELKADER AND S. M. ZUBAIR

baffle orientation on heat transfer and pressure drop Chem. Eng. Process., vol. 36, no. 2, pp. 149–159, 1997.
of shell and tube heat exchangers with leakage flows,” doi:10.1016/S0255-2701(96)04194-3.
Heat. Transf. Engng., vol. 30, no. 14, pp. 1123–1135, 2009. [12] M. O. Petinrin, and A. A. Dare, “Performance of shell
doi:10.1080/01457630902972694. and tube heat exchangers with varying tube layouts,”
[8] A. F. Jozaei, A. Ghafouri, and M. M. Navaei, “Effect of Br. J. Appl. Sci. Technol., vol. 12, no. 2, pp. 1–8, 2016.
Number of baffles on pressure drop and heat transfer in doi:10.9734/BJAST/2016/20021.
a shell and tube heat exchanger,” WASET Intl. J. Mech. [13] S. S. Deshpande, and S. A. Hinge, “Design and perfor-
Aero. Indus. Mecha. and Manuf. Engng., vol. 2, no. 1, mance study of shell and tube heat exchanger with sin-
pp. 18303–18312, 2015. gle segmental baffle having perpendicular & parallel-cut
[9] A. F. Jozaei, A. Baheri, M. K. Hafshejani, and A. orientation,” Int. J. of Engng. Res. Tech., vol. 3, no. 11,
Arad, “Optimization of baffle spacing on heat trans- pp. 1366–1370, 2014.
fer, pressure drop and estimated price in a shell-and- [14] V. Vukić, M. A. Tomi, P. M. Zivkovi, and G. S. Ili, “Effect
tube heat exchanger,” World Appl. Sci. J., vol. 18, no. 12, of segmental baffles on the shell-and-tube heat exchanger
pp. 1727–1736, 2012. effectiveness,” Hem. Ind., vol. 68, no. 2, pp. 171–177, 2014.
[10] H. Li, and V. Kottke, “Effect of baffle spacing on pres- doi:10.2298/HEMIND130127041V.
Downloaded by [UNIVERSITY OF ADELAIDE LIBRARIES] at 08:37 28 December 2017

sure drop and local heat transfer in shell-and-tube heat [15] A. P. Fraas, Heat Exchanger Design. New York: John Wiley
exchangers for staggered tube arrangement,” Int. J. Heat & Sons, 1989.
Mass Transf., vol. 41, no. 10, pp. 1303–1311, 1998. [16] J. W. Palen, and J. Taborek, “Solution of shell side
doi:10.1016/S0017-9310(97)00201-9. flow pressure drop and heat transfer by stream analy-
[11] E. S. Gaddis, and V. Gnielinski, “Pressure drop on the sis method,” Chem. Eng. Prog. Symp. Ser., vol. 65, no. 93,
shell side of shell and tube with segmental baffles,” pp. 53–63, 1969.