Professional Documents
Culture Documents
ScienceDirect
j o u r n a l h o m e p a g e : w w w. e l s e v i e r. c o m / l o c a t e / i j r e f r i g
A R T I C L E
I N F O
A B S T R A C T
Article history:
The thermal performance of hyperbolic profile annular fins subjected to dehumidifying op-
erating conditions is studied. An analytical solution for completely wet fin is derived using
November 2015
psychrometric relationship for completely and partially wet operating conditions is then
under both partially and fully wet operating conditions. An excellent agreement is observed between analytical and numerical solutions for completely wet fin. The fin optimization
Keywords:
is presented based on the analytical solution of completely wet fin. Finally, a finite element
Annular fin
formulation is used for studying the two-dimensional effects of orthotropic thermal con-
Hyperbolic profile
ductivity on the thermal performance of fin under partially and fully wet operating conditions.
Analytical solution
2016 Elsevier Ltd and International Institute of Refrigeration. All rights reserved.
Fin optimization
Finite element analysis
Orthotropic thermal conductivity
Mass transfer
Partially wet fin
* Corresponding author. Mechanical Engineering Department, KFUPM Box # 1474, King Fahd University of Petroleum & Minerals, Dhahran
31261, Saudi Arabia. Tel.: +966 13 860 3135; Fax: +966 13 860 2949.
E-mail address: smzubair@kfupm.edu.sa (S.M. Zubair).
http://dx.doi.org/10.1016/j.ijrefrig.2016.01.006
0140-7007/ 2016 Elsevier Ltd and International Institute of Refrigeration. All rights reserved.
Nomenclature
B
Bi
Co
cp
h
hD
h
ifg
k
L
Le
M
M0
m0
Patm
Q
Q*
q
qb
RH
r*
r
R
T
Ta
1.
t*
t
Introduction
43
Greek symbols
efficiency
dimensionless temperature
Q*
* ax
Qm
(1)
where Q is heat transfer rate through the fin and Q max is the
maximum possible heat transfer rate from the fin if the entire
fin surface is at the prime surface temperature and humidity
ratio.
High efficiency fins are desirable for effective heat transfer. For the case of dry operating conditions, the rate of sensible
heat loss can be increased by using force convection e.g. by
using a fan. However, for a given fin material and geometry,
the maximum value of convective heat transfer under dry operating conditions is governed by the dimensionless parameter
called Biot number. For a circular cross-section pin fin it is given
by:
Bi =
hr
k
(2)
44
Matrix
Parallel
to fiber
(W/m K)
Normal
to fiber
(W/m K)
Polymer
Polymer
Epoxy
330
10100
370
310
310
6.5
0.8
= 0.005
kk*
rz
kk*
rz = 0.01
0.7
= 0.03
kk*
rz
0.6
= 0.1
kk*
rz
0.5
= 1
kk*
rz
L/r =10
0.9
0.4
0.3
2.
0.2
0.1
0
Bi r
10
It is very common to make certain assumptions for simplifying one-dimensional analysis of the fins. The assumptions
considered in the present study are:
(a) The thermal conductivity of the fin, heat transfer coefficient and latent heat of condensation of the water vapor
are constant;
(b) The heat and mass transfer are under steady-state
condition;
2.1.
We consider an annular fin of hyperbolic profile with base thickness of t b * and temperature Tb . The temperature at tip is Tt ,
which for fully wet situation, should be less than the dew point
temperature of the ambient air at temperature Ta and humidity ratio a .The length of the fin is L . The fin geometry and
operating conditions are depicted in Fig. 2.
Applying the energy balance on the infinitesimal circular
ring of width dr* , which has an average thickness of t* , at a
radius of r* from the center of the tube
(3)
where qr* is the conduction heat transfer rate in the radial direction. It can be expressed as,
qr* = 2 r *t (r * ) k
dT
dr *
(4)
Note that t (r* ) shows the variable thickness of the fin along
its length, given by:
r*
t (r * ) = t * b b
r*
qr *
dr *
r *
(6)
therefore;
d2 T
dT
dt (r * ) dT
qr * + dr * = qr * + k 2 r *t (r * )
+ 2 t (r * )
+ 2 r *
dr *
dr *2
dr *
dr * dr *
(7)
qr * = qconvection + qcondensation + qr * + dr *
qr * + dr * = qr * +
45
(5)
also:
2
h
= c pLe 3
hD
(8)
where
hDi fg = hB
(9)
and
B=
i fg
(10)
c pLe 3
d2 T 1 dT
1 dt (r * ) dT
2h
2hD
+
+
+
(Ta T ) +
i fg ( a ) = 0
dr *2 r * dr * t (r * ) dr * dr * kt (r * )
kt (r * )
(11)
subject to the following boundary conditions:
at
r = rb,
T = Tb
and = b
(12)
at
r = rt ,
dT
=0
dr
(13)
or
at r = rt , k
2.1.1.
dT
= h (Ta T ) + hDi fg ( a ) for convective tip
dr
(14)
d2
r*
= mo 2
[ + B ( a )]
dr *2
rb *
Fig. 2 Schematic of a completely wet hyperbolic annular
fin.
(15)
46
= Ta T
(16)
Now we need a relation between the temperature difference and humidity ratio to solve the above equation. In
this regard, we will use the same relation as that of Sharqawy
and Zubair (2007), expressed as
= a2 + b2 T
Q * b = Ab k
= Abk ( b + p ) M2 3
(17)
23 23
23 23
Ai M rt * Bi M rb * + Ai M rb * Bi M rt *
23 23
23 23
Ai
M
r
Bi
M
*
r
*
Ai
M
r
*
Bi
M
r
*
t
t
b
b
a2 = b
b2 =
dew b
Tb
Tdew Tb
dew b
Tdew Tb
M0
m 2
= 0
rb *
(19)
with
2h
m0 =
kt * b
M = M0 (1 + b2 B)
2
(20)
(21)
d2
M2 r * = M02 r *B ( a )
dr *2
(22)
at
r * = rb *, = b
At
r * = rt *,
d
=0
dr
= b
(23)
(24)
and
2 2
2 2
Ai M 3 rt * Bi M 3 r * Ai M 3 r * Bi M 3 rt *
+ p
=
b + p
23 23
23 23
Ai M rt * Bi M rb* Ai M rb* Bi M rt *
Q *max = m02kAs (1 + b2 B) ( b + p )
2 2
2 2
Ai M 3 rt * Bi M 3 rb * + Ai M 3 rb * Bi M 3 rt *
A
(31)
= b M 4 3
As
23 23
23 23
Bi
M
r
Ai
M
r
Bi
M
r
Ai
M
r
*
*
*
*
t
t
b
b
2.1.2.
Although the assumption of linear relationship (cf. Eq. (17)) simplifies the analytical solution of the differential Eq. (34), it does
not account for the actual non-linear psychrometric correlations of an air-water vapor mixture. Therefore, we will obtain
a numerical solution using the actual non-linear psycrometric
relationship.
Normalizing Eq. (11) using the following relations:
r=
r*
T T
, = a
,
L
Ta Tb
a
a b
(32)
And substituting:
(25)
b
Co = B a
Ta Tb
(33)
We get
B ( a a2 b2 Ta )
(1 + b2B)
(26)
Ai ( X ) =
(30)
where;
p =
(29)
(18)
d
dr* r * = rb *
1 1
2
XK1 3 X3 2
3
d2
r
= (mo L )2 [ + Co ]
dr 2
rb
The boundary conditions are:
at
r = rb, = 1
and
=1
(35)
and
(27)
at
1 2 3 2
2
Bi ( X ) =
X I1 3 X
+ I X3 2
1 3 3
3 3
(34)
(28)
or
r = rt ,
d
=0
dr
(36)
r = rt ,
d
= (mo L )2 ( + Co )
dr
47
(37)
Q=
(38)
Q max =
Q max
*
= (m0 L )2 ( R + 1) (1 + Co )
2 rb *k (Ta Tb )
(39)
2 ( R 1) ( + Co ) rdr
rb
(40)
(R + 1) (1 + Co )
2.2.
A partially wet fin condition exists when the fin base temperature is lower, but the fin tip temperature is higher than
the dew point of air. Under such situation, there is a radius,
r = r , where the surface temperature equals the dew point of
the air, i.e. T ( r ) = Tdew . The fin is then divided into two regions:
a wet region for rb r r , with the surface temperature lower
than Tdew and a dry region from r r rt , with the surface temperature higher than Tdew (see Fig. 3). In this regard, separate
governing differential equations must be written for each region.
For rb r r
d2
r
= (mo L )2 [ + Co ]
dr 2
rb
at
r = rt ,
d
= (mo L )2
dr
(44)
(45)
2 ( R 1) ( + Co ) rdr + rdr
rb
r
=
(R + 1) [1 + Co ]
(46)
(41)
3.
and for r r rt
d2
r
= (mo L )2 [ ]
dr 2
rb
(42)
at
r = r , = dew
(43)
48
0.8
Number of nodes
Fin efficiency
51
61
71
81
91
101
111
0.8257
0.8253
0.8251
0.8249
0.8247
0.8246
0.8245
A
B
C
D
E
F
G
m 0 L=0.2
m 0 L=0.6
m 0 L=1
0.6
0.4
Tb = 7 C
Ta = 27 C
RH = 0.6
R=4
0.2
Analytical
Numerical
0.8
0.6
0.4
Tb = 7 C
Ta = 27 C
RH = 0.6
R=4
0
0.5
1.5
m 0L
2.5
3.5
0.2
0.4
0.6
0.8
is presented in Fig. 5; that shows that the maximum temperature difference is approximately 0.7C (at fin tip) that
corresponds to m0 L = 1 . For practicality, the fin will be operating in the range m0 L 0.6 , and the maximum temperature
difference is 0.2C. So this shows that the approximate linear
relation given by Sharqawy and Zubair (2007) has provided a
reasonable accurate non-iterative analytical solution for the
case of hyperbolic fin under considered operating conditions.
Now the fin performance is studied using a numerical approach because it covers both fully wet and partially wet
operating conditions. The results for different radii ratio are
presented in Fig. 6. The fin base and the ambient temperature values are the same for all cases whereas the RH values
have been varied from 0.2 to 1. The fin is fully dry for RH = 0.2
over the considered range of fin parameter m0 L .The fins are
fully wet for high RH and low m0 L values. Moreover, the fin
efficiency decreases with increasing wet portion of the fin, thus
the completely dry fin condition has the highest efficiency and
the highest relative humidity has always the lowest efficiency.
The effect of radius ratio on the fin efficiency for RH = 0.6
is shown in Fig. 7. It shows that the lower radius ratio has higher
fin efficiency. This can be explained by the fact that a low radius
ratio fin can be obtained by truncating tip portion of a high
radius ratio fin, since the tip portion is obviously less effective than the fin base portion due to low temperature difference
between the fin surface and ambient air.
4.
0.2
0
Fin optimization
The optimum fin dimensions can be obtained by using an optimization techniques, for example in a recent paper Huang
and Chung (2014) developed an inverse design algorithm using
the conjugate gradient method (CGM) to estimate the optimum
shape for fully wet annular fins. Optimization of fins can be
described as a process through which we find the optimum
dimensions of a fin for a required amount of heat transfer, or
determining the maximum possible heat transfer if the
49
1.0
0.8
(a)
0.6
0.6
(b)
0.4
0.2
0.0
0.5
0.4
0.2
Tb = 7 C
Ta = 27 C
R=2
0
0.8
1.5
2.5
3.5
Tb = 7 C
Ta = 27 C
R=3
m 0L
0.5
1.5
3.5
0.8
0.8
0.6
(d)
0.4
0.2
2.5
(c)
m 0L
0.4
0.2
Tb = 7 C
Ta = 27 C
R=4
0
0.5
1.5
m 0L
2.5
3.5
Tb = 7 C
Ta = 27 C
R=5
0
0.5
1.5
m 0L
2.5
3.5
Fig. 6 Effect of relative humidity on fin efficiency. (a) R = 2, (b) R = 3, (c) R = 4, and (d) R = 5.
0.8
0.4
0.2
Tb = 7 C
Ta = 27 C
RH = 0.6
0
0.5
1.5
2.5
3.5
m 0L
Fig. 7 Effect of radius ratio on the fin efficiency for a
hyperbolic annular fin.
2hrb *
k
(47)
v=
2 t b *rb * (rt * rb * )
V
=
rb *3
rb *3
(48)
w=
rb *
tb *
(49)
u=
v=
2 rt *
1
w rb *
(50)
1
2w ( R 1 )
(51)
m0 L = u ( R 1 ) w
(52)
rb =
1
R1
(53)
50
R
R1
(54)
R = vw + 1
(55)
Analytical Solution
u (1 + b2 B)
Qb =
10
23
1
2
2
1
Numerical Solution
0.5
1.0
2.0
0
10
-1
10
0.01
0.1
Numerical Solution
(56)
u =2.0
(57)
u =1.0
Q [-]
u =0.5
-1
C2 = u w (1 + b2 B)
23
10
Analytical Solution
1 + vw
10
23
v [-]
10
where
C1 = u w (1 + b2 B)
u =0.25
10
w [-]
rt =
10
u =0.25
(58)
So by differentiating Eq. (56) with respect to w and equating to zero, for fixed values of u and v we can find out that
particular value of w, which will give us the maximum Q .
The resulting analytical expression is rather complex therefore the outcome is presented in a graphical form in Fig. 8,
where the numerical results are presented in terms of dimensionless fin-base radius ( w) and heat transfer rate ( Q ) as a
function of dimensionless volume ( v ) and fin heat transfer parameter ( u). The results are also compared with the numerical
solution. It can be seen that the numerical results are in good
agreement with the analytical solutions such that the numerical solution gives higher Q than analytical solution but lesser
value of w. It is worth noting that the analytical results are
based on an approximate linear temperaturehumidity ratio
relationship given by Sharqawy and Zubair (2007) whereas the
numerical solution is based on the corresponding actual psychrometric relationship. A salient feature of presenting the
results in a graphical format is: although we started with a
known volume of fin material and found the maximum heat
transfer rate, now we can go in the reverse direction, i.e. for a
given heat transfer rate we can find the minimum required fin
material.
The above optimization results are also presented in terms
of fitted regression equations (within 1% to the actual numerical data) as:
w = c1 vn1
(59)
Q = c2 vn 2
(60)
c1 = 1.78863 u0.599547
(61)
-2
10
0.01
0.1
v [-]
10
c2 = 0.78741 u1.2158
(62)
(63)
(64)
5.
The element should be capable of modeling conduction phenomena with convection and condensation boundary conditions
for a two-dimensional orthotropic fin material. The main advantage of this formulation is that it can be used for any
arbitrary fin shape, specified heat flux boundary condition,
specified temperature boundary condition and composite fin
materials (e.g. a coating layer), without any additional mathematical complexity. Some of the advantages of the finite
element method over finite difference method are (Peir and
Sherwin, 2005):
a. The finite difference (FDM) uses the differential form of the
governing equations whereas the finite element method
(FEM) is based on integral formulation. The use of integral
form in FEM provides a more natural treatment of Neumann
boundary condition (i.e. imposed heat flux in case of fin
problem) than the FDM.
b. FEM is better suited than FDM to deal with the complex geometries in two-dimensional problems because the integral
form does not depend on the special mesh structure.
Based on the above discussion, the finite element formulation can be used to extend the present study for other fin
shapes, boundary conditions and composite fins.
The governing differential equation for steady state heat
conduction with no internal heat generation in an orthotropic
material can be expressed as (Lewis et al., 2004).
kx
2 T
2 T
2 T
+ ky 2 + kz 2 = 0
2
x
y
z
(65)
T = Tb
T
kx
l + ky
x
T
kx
l + ky
x
T
+ kz
m
y
T
+ kz
m
y
on S1
T
n + qb = 0
on S2
z
(66)
T
n + h (Ta T ) + hDi fg ( a ) = 0 on S3
z
where, kx, ky , and kz are the thermal conductivities in orthogo , and n are
nal directions, qb is the specified heat flux and l, m
the surface normals. The boundary conditions are very common
in heat transfer problems and are called boundary conditions of first, second and third kinds, respectively (Lienhard and
Lienhard, 2011). In the context of fin application, if the prime
surface is at constant temperature, the surface S1 will be the
fin base. Alternatively, a fin can receive a constant heat flux
from a prime surface to its base. For such cases, the surface
S2 would be the fin base. It is worth mentioning that both
boundary conditions cannot occur simultaneously at the fin
base. The surface S3 would be the remaining fin surface through
which heat transfer occurs between the fin and the surrounding environment.
To solve Eq. (65) with boundary conditions (66), an additional equation for is required. The psychrometric correlation
between and T is non-linear, therefore following linear relationship is used:
The details about the justification of using the linear relationship and corresponding values of coefficient a and b are
presented in Appendix.
Therefore, by using the relationships in Eq. (67), the last
boundary condition in Eq. (66) becomes:
kx
T
T
T
+ kz
l + ky
m
n + h (Ta T ) = 0
x
y
z
(67)
(68)
where h (Ta T ) is the combined heat transfer due to convection and condensation. Therefore h and Ta are the
corresponding equivalent heat transfer coefficient and equivalent ambient temperature given by:
h = h (1 + Bb)
Ta =
(69)
Ta + B ( a a)
1 + Bb
5.1.
(70)
= U + q + h
(71)
The functional has three terms that are associated with internal energy (U ) , heat conduction ( q ) and heat convection
(h ) . For the governing equation (65) with associated boundary conditions (66), the terms are (Logan, 2007):
U=
2
2
T 2
1
T
T
kx + ky + kz dV
y
z
2 V
x
q = qbTdS
(72)
S2
1
h = h (Ta T )2 dS
2
S3
Therefore, for the case of equivalent heat transfer h (Ta T )
(cf. Eq. (68)) the functional becomes:
2
2
T 2
1
T
T
+
k
k
x
y
y + kz z dV qbTdS
V
S2
5.2.
1
h (Ta T )2 dS
2
S3
(73)
T
T
[B] [D][B] dV + hTa [N ] [N ] dS {T}
S3
V
= [ N ] qbdS + hTa [N ] dS
T
= a + bT
51
S2
S3
(74)
52
where,
X = 1 + b2 B
and
1 + ( B Ta ) ( a a2 )
=
1 + b2 B
FEA
(75)
Numerical
0.8
0.6
k {T} = {f }
(76)
0.4
Tb = 7 C
Ta = 27 C
0.2
RH = 0.6
R=4
[k ] = [kq ] + [k h ] with
and
k h = X h [N ]T [N ] dS
(77)
S3
} , with:
and
{fh } = hTa [N ]T dS
(78)
The two dimensional axisymmetric finite element formulation has been implemented using commercial software MATLAB
(2015). Four-node quadrilateral isoperimetric element has been
used for meshing the two-dimensional computational domain
of a hyperbolic profile annular fin. As before, the grid independence was verified by obtaining a series of solutions with
different mesh sizes. The mesh size was then selected after
which the grid independence was observed. It is worth mentioning that according to boundary conditions specified by Eq.
(66), the fin base corresponds to surface S1 (i.e. specified temperature), the fin surface corresponds to surface S3 (i.e.
convection and condensation) and the insulated tip corresponds to surface S2 (i.e. zero heat flux). The results of finite
element model for isotropic fin case are validated against the
numerical solution presented in section 2.1.2. The results are
shown in Fig. 9. A closed agreement is observed between the
two solutions. This shows that the presented finite element
formulation is capable of accounting for the non-linear relationship between temperature and humidity ratio with
acceptable engineering accuracy because the numerical solution of section 2.1.2 is based on actual psychrometric data.
6.
0.5
1.5
2.5
3.5
S3
It is worth mentioning that the factors X and are introduced due to the condensation heat transfer. For the unit values
of both parameters, the formulation will be transformed into
standard formulation (for sensible heat transfer only) available in many textbooks on finite element method. This situation
would occur for B = 0 (cf. Eq. (75)). The physical interpretation is that the latent heat of water condensation ( i fg ) (cf. Eq.
(10)) under dry operating conditions is zero.
5.3.
m 0L
Similarly,
{f } = {f } + {f
The advantage of finite element formulation compared to standard finite difference approach is that it can be used for any
irregular geometry without any additional mathematical complexities. The objective of this section is to study the effects
of orthotropic thermal conductivity on thermal performance
of hyperbolic profile annular fins under dehumidifying operating conditions. The fin analyses presented in preceding
sections are one-dimensional, whereas the heat transfer in an
orthotropic annular fin is certainly a two-dimensional problem.
Thus, the axisymmetric form of the derived finite element formulation will be used in this section.
The study under dry operating conditions (Pashah et al.,
2011) has shown that the thermal conductivity has considerable effects when the thermal conductivity in fin longitudinal
direction is higher than the conductivity in lateral direction.
For the values given in Table 1, the thermal conductivity ratio
krz = kr kz ranges between 1 and 110, therefore the effect of
thermal conductivity ratio has been studied over the range
1 krz 300 , the reason to add the value 300 will be discussed in the following.
The results are presented in Fig. 10. It is obvious that for
practical reasons, the isotropic fin (i.e., krz = 1 ) will be useful in
the range m0 L 0.5 because after that, the fin efficiency drops
exponentially. This implies that for a fixed fin geometry and
material, the h values must be small (cf. Eq. (20)), because for
higher h values only base portion of the fin would be effective and tip portion would be at the same temperature as the
incoming ambient air stream. This means that a slender fin
operating under forced convection would have low efficiency.
However, increasing the thermal conductivity ratio starts increasing the fin efficiency for higher values of m0 L . For example
if the objective is to attain a minimum fin efficiency of 60%
then krz = 50 would give the same efficiency for m0 L = 3.5 (the
corresponding efficiency for an isotropic case ( krz = 1 ) would
be only 9%). Another important observation is that the fin is
partially wet for the isotropic fin case; increase in thermal conductivity ratio decreases the dry portion of the fin and increases
the fin efficiency simultaneously such that the fin becomes
completely wet over the considered range of m0 L for krz 20 .
krz =300
0.8
krz =100
krz =50
krz =30
0.6
0.4
krz =10
krz =5
krz =2
krz =20
PW
Tb = 7 C
Ta = 27 C
0.2
RH = 0.6
R=4
0
0.5
1.5
m 0L
2.5
3.5
It is also interesting to note that the gain in thermal performance is not linearly related to increase in thermal
conductivity ratio. The gain in fin efficiency against krz values
is presented in Fig. 11. It can be seen that with respect to isotropic fin at m0 L = 4 ; the gain in fin efficiency is 47% for krz = 50 ,
compared to 60% and 76% for krz = 100 and krz = 300 , respectively. It means that the gain in efficiency is much higher in
the range 2 krz 50 than it is in the range 50 krz 300 .
7.
1. The closed-form analytical solution for completely wet hyperbolic fin; based on an approximate linear temperature
humidity ratio provided excellent results when compared
to numerical solution. The results show a maximum 2% error
in fin efficiency value with respect to the numerical solution based on actual psychrometric relationship, for the
considered set of geometric and operating parameters.
2. A numerical solution is more convenient for partially wet
hyperbolic fin operating conditions due to non-linear actual
psychrometric relationship between the air temperature and
humidity ratio.
3. For completely wet fin conditions, regression equations
(based on analytical solution) have been developed to calculate the optimum fin geometry when the fin volume or
the heat transfer rate is specified.
4. The orthotropic thermal conductivity effects are studied
using a finite element formulation under fully and partially wet fin conditions. The results showed that the higher
thermal conductivity ratio results in better thermal performance of fins at higher values of fin parameter m0 L and
fully wet condition.
Acknowledgements
The authors would like to acknowledge the financial support
given by the King Fahd University of Petroleum and Minerals
for this research work from Budget Head SB131005. Syed Zubair
also acknowledges support from the King Fahd University of
Petroleum and Minerals through the project IN121042.
Concluding remarks
The approximate cubic polynomial relationship between temperature and humidity ratio is (Liang et al., 2000):
1.0
Tb = 7 C
Ta = 27 C
0.8
53
0 T 30C
(A1)
krz =300
RH = 0.6
R=4
krz =100
krz =50
0.6
krz =30
= a + bT
krz =20
0.4
krz =10
krz =5
0.2
krz =2
0.0
0.5
1.5
2.5
3.5
m 0L
Fig. 11 Gain in fin efficiency with respect to isotropic
material fin; for different values of thermal conductivity
ratios.
(A2)
The coefficients a and b are obtained by a regression analysis on 100 data points, each over 1C temperature interval. The
coefficient values are considered with 20 decimal places. The
difference between temperature values obtained through Eq.
(A1) and Eq. (A2) are within 0.005C.
The piecewise linear polynomial is implemented by obtaining the finite element solution in iterative manner. A first
solution is obtained by assuming dry operating conditions and
then following iterative solution uses appropriate a and b values
based on preceding iteration temperature results. The solution is assumed to be converged when the maximum
54
REFERENCES