Aleta
Aleta
Nomenclature
ra , rb – inner and outer radius, respectively
t – thickness of the fin
ha – fin thickness at base
h – convective heat transfer coefficient
C1 – profile function, ha ra
k(T ) – variable thermal conductivity
ka – thermal conductivity at ambient temperature
κ – parameter describing variation of thermal conductivity
β – dimensionless parameter describing variation of thermal conductivity,
κ(Ta − T∞ )
T, Ta , T∞ – fin, base and ambient temperature, respectively
r, φ – polar coordinates
α – linear coefficient of thermal expansion
E – Young’s modulus
σr , σφ , εr , εφ – radial and tangential stress and strains, respectively
C, D1 , D2 , η, A, B – constants p
ψ – thermo-geometric parameter, 2hra2 /(k∞ ha
ξ, ξ1 – dimensionless radius of fin, ξ = (r − ra )/ra , ξ1 = ξ + 1
R – dimensionless outer radius, R = rb /ra
θ – dimensionless temperature, θ = (T − T∞ )/(Ta − T∞ )
σr , σ φ – dimensionless radial and tangential stress, σr /E and σφ /E
χ – dimensionless coefficient of thermal expansion, α(Ta − T∞ )
ν – Poisson’s ratio
Ac , dAs – cross sectional area of fin amd elemental surface area of fin, respectively
n – profile parameter
Ap , Bp – Adomian polynomials
438 A. Mallick et al.
1. Introduction
The rapid heat dissipation from a surface to the surroundings is required in many engineering
and industrial applications such as heat exchangers, semiconductors, transformers, motors and
many other electrical, electronical and mechanical components. A fin is an extended surface
frequently used for direct heat dissipation from a hot surface to its surroundings (Kern and
Kraus, 1972; Kraus et al., 2001).
Thermal analyses of the fin with various geometries are presented by many researchers (Yeh,
1997; Mokheimer, 2002). However, most of the available works focus on straight fins for the-
ir ease in manufacturing and mathematical formulation. Several mathematical techniques like
Adomian’s Decomposition Method (ADM), Homotopy Perturbation Method (HPM), Variatio-
nal Iteration Method (VIM), and Deferential Transformation Method (DTM) can be found in
use by many researchers to obtain semi-analytical solutions to nonlinear heat equations for fins
(Miansari et al., 2008; Moradi, 2011). The improvement in heat transfer is shown mainly to be
associated with their surface modification. So, an enhancement in the heat transfer rate criti-
cally depends on the selection of correct fin shape. Radial shape is observed to outperform the
straight fin in respect of heat transfer ability (Behnia et al., 1998). A hyperbolic profile contains
even larger surface area to yield more heat dissipation from the surface to the surroundings.
Thus, a radial fin with a hyperbolic profile would be the most preferred shape for a better heat
transfer rate with less material involved.
Recently, the heat transfer in an annular fin with a hyperbolic profile was presented by Aksoy
(2013). Non-uniform temperature distributions during the heat transfer process induce thermal
stresses in the fin material. Thermal stresses are responsible for various mechanical failures, i.e.,
crack propagation, creep and fatigue failure that can reduce the fin life. The study of thermal
stresses in them is, therefore, important to prevent their early damage. So far, only few authors
(Chiu and Chen, 2002; Mallick et al., 2015) have reported the analysis of stresses developed due
to variation of the temperature gradient in an annular fin, and these papers mainly focus on
uniform thickness of the fins. A careful review of the published literature reveal that till date
no work reports theoretical studies for thermal stresses in an annular radial fin of a hyperbolic
profile with the consideration of variable thermal conductivity.
In this paper, a novel analytical approach is used to predict a near closed form solution
for thermal stresses in an isotropic homogeneous annular fin with a hyperbolic profile. A va-
riable thermal conductivity parameter is considered in the analysis. The temperature field is
obtained by solving a non-linear steady sate heat conduction-convection equation of a fin using
the Adominan Decomposition Method. The solution for the temperature field is expressed in a
polynomial form. The plane stress condition in a rotational symmetric geometry with respect
to its axis is considered in this study. A classical thermo-elasticity relation coupled with the
solution of the temperature field is employed to obtain the stress field. The study includes the
affect of various non-dimensional parameters such as thermal conductivity parameter (β), the
thermo-geometric fin parameter (ψ), the coefficient of thermal expansion (χ) and Poisson’s ra-
tio (ν) on the stress field. The results are compared with those given by Chiu and Chen (2002)
to resolve the accuracy of the present method. The analyses presented envisage an improved fin
design process.
An axisymmetric undeformed annular fin of a hyperbolic profile (Fig. 1) with variable thermal
conductivity (k(T)) is considered. The fin is exposed to a conductive-convective environment
at a constant ambient temperature and the tip of the fin is maintained to be well-insulated
Effect of heat transfer on thermal stresses... 439
with a traction free condition. As thickness of the fin is relatively small compared to its radial
dimension, the temperature distribution and the stress field can be assumed to vary in the radial
direction only.
The profile function for fin tapering from the base to tip is given by
r n
t = ha with n 1 (2.1)
ra
where n is the profile parameter.
The steady state energy balance equation for the axisymmetric annular fin is expressed
(Mokheimer, 2002) as
d dT hdAs
kAc − (T − T∞ ) = 0 (2.2)
dr dr dr
where k = ka [1 + κ(T − T∞ )]. The notations used in Eq. (2.1) and Eq. (2.2) are defined in the
nomenclature.
The fin profile becomes hyperbolic when the profile parameter n = −1. Equation (2.2) can
be expressed in the following dimensionless form
d2 θ d2 θ dθ 2
+ βθ + β − ψ 2 (1 + ξ)θ = 0 with 0¬ξ ¬R−1 (2.3)
dξ 2 dξ 2 dξ
where
s
T − T∞ r − ra rb 2hra2
θ= β = κ(Ta − T∞ ) ξ= R= ψ=
Ta − T∞ ra ra ka ha
are non-dimensional parameters.
In order to evaluate the temperature distribution, the following non-dimensional boundary
conditions are employed
0
that θ=1
ξ= dθ (2.4)
R−1
that =0
dξ
where
∞ ∞
′ 2
X X
′′
N A = θθ = Ap N B = (θ ) = Bp
p=0 p=0
are the nonlinear terms. The Adomian polynomials An and Bn are estimated as follows
A0 θ0 ′′
θ 0
′′
A θ1 θ0 θ
1
1′′
A θ2 θ1 θ0 θ
2
2′′
A3 =
θ3 θ2 θ1 θ0 θ
3
A θ θ3 θ2 θ1 θ0 ′′
θ 4
4
4
· · · · · · · ·
·
· · · · · · · ·
(3.3)
B0
θ0′ ′
θ 0
′ ′
B θ1 θ0′ θ
1
1′
′
B θ2 θ1′ θ0′ θ
2
′ 2′
B3 =
θ3 θ2′ θ1′ θ0′ θ
3
B θ ′ θ3′ θ2′ θ1′ θ0′ ′
θ 4
4
4
· · · · · · · ·
·
· · · · · · · ·
where (·)′ and (·)′′ are d/dξ and d2 /dξ 2 , respectively. An inverse operator L−1 can conveniently
be used as a two-fold identifying integral in both sides of Eq. (3.2). Applying the Maclaurin
series, yields
∞ ∞
! ∞ ∞
! ! !
dθ(0) 2 −1
X
2 −1
X
−1
X
−1
X
θ = θ(0) + ξ +ψ L θp + ψ L ξθp − β L Ap − β L Bp
dξ p=0 p=0 p=0 p=0
(3.4)
The first two terms of the right-hand side of Eq. (3.4) can be defined by
θ0 = 1 + Cξ (3.5)
with p 0.
Effect of heat transfer on thermal stresses... 441
In the present analysis, the estimation of the first four significant terms, i.e, n = 0 to 3, of
the temperature field is now expressed in the following form
The total temperature field can now be estimated from Eqs. (3.5) and (3.7) to yield a
polynomial form
m
X
θ= Ki ξ i with 0¬ξ ¬R−1 (3.8)
i=0
d2 u 1 + n du (νn − 1)u dT T
+ + − (1 + ν)α + n =0 (4.3)
dr 2 r dr r2 dr r
Introducing a new non-dimensional radius, ξ1 = r/ra , the following termo-elastic equation
of equilibrium is obtained
d2 u 1 + n du (νn − 1)u h dθ θi
+ + − (1 + ν)ra α (Ta − T∞ ) + (Ta − T∞ )n =0 (4.4)
dξ12 ξ1 dξ1 ξ12 dξ1 ξ1
The temperature field in Eq. (3.8) is now modified by the new non-dimensional radius ξ1
and can be expressed in an analogous form
m
X
θ= Li ξ1i 1 ¬ ξ1 ¬ R (4.5)
i=0
Equations (4.4) and (4.5) give together a new form of the equilibrium equation
m
" #
d2 u 1 + n du (νn − 1)u
(i + n)Li ξ1i−1
X
+ + = (1 + ν)ra χ (4.6)
dξ12 ξ1 dξ1 ξ12 i=0
The complete solution to Eq. (4.6) consists of a homogeneous and a particular solution, and
is given as
m
Li (i + n)
u = D1 ξ1η1 + D2 ξ1η2 + ξ i+1
X
(1 + ν)χra (4.7)
i=0
n(1 + i + ν) + i(i + 2) 1
p
where D1 , D2 , and η1,2 = −(n/2) ± 1 − νn + n2 /4 are constants. The stress field Eq. (4.1) can
now be expressed in the non-dimensional form
h du u1 i 1
σr = +ν − χ(1 + ν)θ
dξ1 ξ1 ra 1 − ν2
h u (4.8)
du 1
i 1
σφ = +ν − χ(1 + ν)θ
ξ1 dξ1 ra 1 − ν2
where σ r = σr /E and σ φ = σφ /E are non-dimensional radial and tangential stresses, respecti-
vely. The near closed form solution for the stress field is obtained by combining Eq. (4.7) and
Eq. (4.8)
m
iLi
σ r = Aξ1η1 −1 + Bξ1η2 −1 − χ
X
ξi
i=0
n(1 + i + ν) + i(i + 2) 1
m (4.9)
i(n + i + 1)Li
−η2 Aξ1η1 −1 η1 Bξ1η2 −1
X
σφ = − −χ ξi
i=0
n(1 + i + ν) + i(i + 2) 1
where
D1 (η1 + ν) D2 (η2 + ν)
A= B=
ra (1 − ν 2 ) ra (1 − ν 2 )
are constants estimated from the boundary conditions σ r = 0 at ξ1 = 1 and R.
A near closed form solution for thermal stresses in the isotropic annular fin with a hyperbolic
profile is derived. ADM is employed to obtain the non-dimensional temperature field represented
in a polynomial form. The integral constants C are evaluated using the minimum decomposition
error J = θi+1 −θi , (J ¬ 10−4 ) approach. Figure 2 represents the values of the integral constant C
for different values of the variable thermal conductivity parameter. The accepted C values are
the x-values corresponding to the minimum decomposition error in each case. These constants
directly influence the variation of local temperature distribution. The determined values of C
for ψ = 0.2 and β = 0.3, 0 and −0.3 are −0.1737, −0.2167 and −0.2758, respectively.
With the best of literature search, this work is the first attempt to estimate thermal stres-
ses in an annular fin with a hyperbolic profile along with a variable thermal conductivity. Due
to this limitation, stress fields for the hyperbolic fin profile obtained from the present formu-
lation could not be compared. However, the present formulation for stress fields can be used
to obtain results for an annular fin of uniform thickness by setting n = 0 in Eq. (4.9). The-
se stress field results are compared with those by Chiu and Chen (2002) shown in Fig. 3.
The results in Fig. 3 take into account the same parameters, κ = ±0.00018, ξ1 = 1 to 3,
h = 50 W/(m2 K), k∞ = 186 W/(mK) and t(= ha ) = 0.004 m and boundary conditions for cla-
rity of the comparison. Both temperature and stress fields in uniform thickness obtained from
Eqs. (4.5) and (4.9) visibly reveal close agreement.
Figure 4 illustrates the effect of various thermo-mechanical parameters on the non-
-dimensional temperature and stress field. Unless mentioned otherwise, the numerical values
Effect of heat transfer on thermal stresses... 443
Fig. 2. Estimation of the integral constant using the minimum decomposition error approach
Fig. 3. Comparison of the results for (a) temperature distribution, (b) radial stress distribution and
(c) tangential stress distribution for an annular fin with uniform thickness. The results heve been
estimated from the present closed form solution for the fin with variable thickness by setting n = 0
of the non-dimensional parameters are taken to be β = 0, ψ = 0.2, χ = 1 and ν = 0.3 for all
the cases. In Fig. 4a, it can be seen that the steeper temperature gradient is associated with a
lower variable thermal conductivity parameter β. As a result, the mean temperature difference
between the base and fin tip is decreased with an increase in β. On the other hand, higher thermo-
geometric parameter ψ induces higher temperature gradient. Lower thermo-geometric parameter
indicates the fin as thermally thin with less thermal resistivity. Thus, the heat conduction inside
the body is much faster than the heat convection away from the surface. Furthermore, it can
be observed that the coefficient of thermal expansion χ and Poisson’s ratio ν do not affect the
temperature field. Nevertheless, the variation of stresses is influenced by all thermo-mechanical
parameters. The stress field in Eq. (4.9) can be seen to vary linearly with the coefficient of ther-
mal expansion. Consequently, the stress field can be changed with modification in the coefficient
of thermal expansion by n-times. Thus, the maximum of non-dimensional radial and tangential
444 A. Mallick et al.
stress magnitudes: −0.0558 and −0.19523 increase to −0.1126 and −0.39046, respectively, by a
change in the coefficient of thermal expansion from 1 to 2 (negative value indicates compressive
stress) shown in Fig. 4b and 4c. The parameters β and ψ significantly influence the variation of
stress fields. The stress magnitude increases with an increment of ψ, and decreases with an in-
crease in β. Interestingly, the stress field is very marginally affected by Poisson’s ratio ν (Fig. 4).
This result is reasonable for axisymmetric plane stress assumptions in the analysis of the annular
fin with a hyperbolic profile.
Fig. 4. Effect of various thermo-mechanical properties on (a) temperature distribution, (b) radial stress
field, and (c) tangential stress field. Unless mentioned otherwise, β = 0, ψ = 0.2, χ = 1 and ν = 0.3
A comparison of the stress field between a fin of uniform thickness and that of a hyperbolic
profile keeping the same material volume condition are depicted in Fig. 5a and 5b. Except for
the profile geometry, all other parameters, i.e. V = 4.0212 · 10−5 m3 , κ = 0, h = 50 W/(m2 K),
k∞ = 186 W/(mK) (Chiu and Chen, 2002) are maintained the same in both profiles. Herein,
the maximum value of σ r and σ φ (compressive) are found to be less in the hyperbolic profile
than in the uniform thickness condition. The σ r variation with the length parameter indicates
a better symmetric distribution over dimensionless radius in the hyperbolic profile than that of
the uniform thickness profile.
Furthermore, the σ φ (compressive) variation in the hyperbolic profile is significantly lower
compared to the uniform thickness case near to the base of the fin. A marginally higher σ φ
(tensile) variation is observed in the hyperbolic profile near the fin tip. These results reveal that
the fin with the hyperbolic profile is much safer from the material failure view point than that of
the uniform thickness profile due to the lower stress level. Conversely, for the same stress field,
the fin with the hyperbolic profile is much more compact and requires less material.
Figure 6 shows the surface plot of the temperature and stress field along the radial direction
for variable β and ψ. The non-dimensional temperature surfaces (Fig. 6a) reveal that the tem-
perature difference (∆θ = θra − θ∞ ) from the base to tip gradually increases with a decrease
in variable thermal conductivity parameters. On the other hand, the reverse nature is observed
with the variation of the thermo-geometric parameter. The results obtained suggest that for a
Effect of heat transfer on thermal stresses... 445
Fig. 5. Comparison of (a) radial stress field and (b) tangential stress field between the uniform and
hyperbolic annular fin. The volume and other properties are the same in both cases
Fig. 6. Surface domain for (a) temperature field, (b) radial stress field, and (c) tangential stress field
along the radial direction ξ1 with the variation of thermal parameters (i) β and (ii) ψ
specific surface geometry the heat transfer can be enhanced either by decreasing the thermo-
geometric parameter or by selecting materials with higher thermal conductivity. In addition to
the heat transfer enhancement, the investigation of material failure is also an important aspect
for the designer of the fin. The surface plots for σ r and σ φ distribution with the variation of
β and ψ are depicted in Fig. 6b and 6c, respectively. From the surface representation of stress
distribution, it is apparent that the maximum σ r (compressive) and σ φ (compressive or tensile)
446 A. Mallick et al.
are achievable either with a lower value of β or by the use of a higher thermo-geometric parame-
ter ψ. The lower value of β induces high thermal resistance. As a consequence, poor heat transfer
through the fin material causes a higher value of the local temperature difference between two
neighbouring material points along the radial direction. Thus, a very low heat transfer can lead
to the excessively induced thermal stress that reduces the fin life. Therefore, a study of the
maximum limit of heat dissipation ability and the maximum limit of thermal stresses induced
is necessary for better design.
6. Conclusions
A method of analysis for the determination of thermal stresses in an annular fin with a hyperbolic
profile using variable thermal conductivity is presented. The ADM coupled with thermo-elasticity
approach is chosen to derive an approximate analytical solution for thermal stresses. In order
to validate the present analytical solution, the results are compared with the results available
in literature and found to be in very good agreement. The effects of various non-dimensional
parameters on the temperature field and stress fields are also investigated. Based on the present
study, the remarkable outcomes are:
i. ADM is useful and efficient to obtain a general closed form solution for the stress field in fin
of a hyperbolic profile as well as of a uniform thickness with variable thermal conductivity.
ii. The temperature field is influenced by the parameters β and ψ only, while, the stress
fields are governed by all non-dimensional parameters β, ψ, χ and ν. Notably, the effect
of Poisson’s ratio ν on the stress field is observable only in hyperbolic fin profile. And this
effect is very insignificant compared to the contribution of other parameters.
iii. The variation of radial stress with the length parameter exhibits better symmetric distri-
bution in the hyperbolic profile then that of the uniform thickness.
iv. The surface plot for the temperature field and stress field with different values of β and ψ
discloses the nature of thermo-parametric dependence on the temperature and stress field.
v. Improvement in heat transfer can be attained either by increasing the thermal conducti-
vity parameter or setting a lower thermo-geometric parameter. The lower value of local
temperature difference may induce higher thermal stresses in the part. So, an appropriate
combination of the heat dissipation limit and induced stress field can only be achieved
with an efficient design of the fin profile. The hyperbolic section of the fin appears to be a
better choice in all respects.
vi. The present method of analysis is expected to help understanding of the heat transfer
phenomena and thermal stress development in the hyperbolic fin.
Appendix
The Ki estimation (14 terms considered in this study) are given below
K0 = 1 K1 = C
1
K2 = (−C 2 β + ψ 2 + C 2 β 2 − ψ 2 β − C 2 β 3 + ψ 2 β 2 + C 2 β 4 − ψ 2 β 3 )
2
1
K3 = (Cψ 2 + ψ 2 + 3C 3 β 2 − 4Cβψ 2 − βψ 2 − 6C 3 β 3 + β 2 ψ 2 + 7Cβ 2 ψ 2 + 7C 3 β 4
6
− 8Cψ 2 β 3 − ψ 2 β 3 )
Effect of heat transfer on thermal stresses... 447
1
K4 = (2Cψ 2 − 5C 2 βψ 2 − 6Cβψ 2 + ψ 4 − 15C 4 β 3 + 28C 2 β 2 ψ 2 + 10Cβ 2 ψ 2 − 5βψ 4
24
+ 32C 4 β 4 + 4C 3 β 4 + C 2 β 3 − 53C 2 β 3 ψ 2 − 16Cβ 3 ψ 2 − β 2 ψ 2 + 7β 2 ψ 4 )
1
K5 = (−13C 2 βψ 2 + Cψ 4 + 4ψ 4 + 38C 3 β 2 ψ 2 + 56C 2 β 2 ψ 2 − 20Cβψ 4 − 18βψ 4
120
+ 78C 5 β 4 + 12C 4 β 4 − 205C 2 β 3 ψ 2 − 120C 2 β 3 ψ 2 − 2Cβ 3 ψ 2 + 96Cβ 2 ψ 4 + 50ψ 4 )
1
K6 = (6Cψ 4 + 4ψ 4 + 120C 3 β 2 ψ 2 − 21C 2 βψ 4 − 86Cβψ 4 − 18βψ 4 + ψ 6 − 116C 4 β 3
720
+ 30C 5 β 4 − 272C 3 β 3 − 91C 4 β 3 ψ 2 − 156C 3 β 3 ψ 2 − 20C 2 β 3 ψ 2 + 193Cβ 2 ψ 2
− 48C 3 β 2 ψ 2 + 124C 2 β 2 ψ 4 + 172Cβ 2 ψ 4 + 16β 2 ψ 4 + 22β 2 ψ 2 − 8βψ 4 − 12βψ 6 )
1
K7 = (10Cψ 4 − 150C 2 βψ 4 − 128Cβψ 4 + Cψ 6 + 9ψ 6 − 1234C 4 β 3 ψ 2 + 329C 3 β 2 ψ 4
5040
+ 1848C 2 β 2 ψ 4 + 584Cβ 2 ψ 4 − 82Cβψ 6 − 112C 3 β 3 ψ 2 − 151βψ 6 )
1
K8 = (−298C 2 βψ 4 + 12Cψ 6 + 28ψ 6 + 5644C 3 β 2 ψ 4 + 5822C 2 β 2 ψ 4 − 243C 2 βψ 6
40320
− 2020Cβψ 6 − 1310βψ 6 − 240C 4 β 3 ψ 6 + 42C 3 β 2 ψ 8 + 172C 2 β 2 ψ 8 + 36Cβ 2 ψ 8
− 2Cβψ 10 + ψ 8 )
1
K9 = (52Cψ 6 + 28ψ 6 + 6792C 3 β 2 ψ 4 + 256C 2 β 2 ψ 4 − 1055C 2 βψ 6 − 3022Cβψ 6
362880
− 686βψ 6 + Cψ 8 + 16ψ 8 )
1
K10 = (40Cψ 6 + 298C 3 β 2 ψ 4 − 3529C 2 βψ 6 − 2692Cβψ 6 + 10Cψ 8 + 50ψ 8 )
1814400
1
K11 = (−5053C 2 βψ 6 − 28Cβψ 6 + 80Cψ 8 + 140ψ 8 )
19958400
1
K12 = (−4C 2 βψ 6 + 15Cψ 8 + 7ψ 8 )
11975040
1
K13 = Cψ 8
7076160
References
1. Adomian G., 1988, Non-Linear Stochastic System Theory and Application to Physics, Kluwer
Academic Publisher, Dordrecht
2. Aksoy I.G., 2013, Adomian decomposition method for heat conduction in an annular fin of hy-
perbolic profile with temperature dependent thermal conductivity, Journal of Thermal Science and
Technology, 33, 1-8
3. Behnia M., Copeland D., Soodphadakee D., 1998, A comparison of heat sink geometries
for laminar forced convection, Proceedings of the Sixth Intersociety Conference on Thermal and
Thermomechanical Phenomena
4. Chiu C.H., Chen C.K., 2002, Application of the decomposition method to the thermal stresses
in isotropic circular fins with temperature-dependent thermal conductivity, Acta Mechanica, 157,
147-158
5. Kern Q.D., Kraus D.A., 1972, Extended Surface Heat Transfer, McGraw-Hill, New York, USA
6. Kraus A.D., Aziz A., Welty, J.R., 2001, Extended Surface Heat Transfer, John Wiley and
Sons, New York
7. Mallick A., Ghosal S., Sarkar P.K., Ranjan R., 2015, Homotopy perturbation method for
thermal stresses in an annular fin with variable thermal conductivity, Journal of Thermal Stresses,
38, 110-132
448 A. Mallick et al.
8. Miansari M.O., Ganji D.D., Miansari M.E., 2008, Application of He’s variational iteration
method to nonlinear heat transfer equations, Physics Letters A, 372, 770-785
9. Mokheimer E.M.A., 2002, Performance of annular fins with different profiles subject to variable
heat transfer coefficient, International Journal of Heat and Mass Transfer, 45, 3631-3642
10. Moradi A., 2011, Analytical solution for fin with temperature dependent heat transfer coefficient,
International Journal of Applied Science and Engineering, 3, 1-12
11. Timoshenko S.P., Goodier J.N., 1970, Theory of Elasticity, McGraw-Hill, New York
12. Yeh R.H., 1997, An analytical study of the optimum dimensions of rectangular fins and cylindrical
pin fins, International Journal of Heat and Mass Transfer, 40, 3607-3615
Manuscript received May 29, 2015; accepted for print August 25, 2015