International Journal of Latest Technology in Engineering, Management & Applied Science (IJLTEMAS)

Volume VI, Issue IV, April 2017 | ISSN 2278-2540

Integral Transform Methods for Inverse Problem of

Heat Conduction with Known Boundary of Semi-
Infinite Hollow Cylinder and Its Stresses
S. S. Singru, N. W. Khobragade
Shri Dnyanesh Mahavidyalaya Nawargaon, Department of Mathematics, MJP Educational Campus,
RTM Nagpur University Nagpur 440 033, India.

Abstract: Three dimensional inverse transient thermoelastic use of FEM harmonic basic functions. The essence of the
problem of a semi-infinite hollow cylinder is considered within problem is aimed at presenting an approximate solution based
the context of the theory of generalized thermoelasticity. The on possibly large finite element. Introduction of harmonic
lower surface, upper surface and inner surface of the semi- functions allows reducing the order of numerical integration
infinite hollow cylinder occupying the space
as compared to a classical Finite Element Method. Numerical
D {( x, y, z ) R : a ( x y ) b, 0 z }
3 2 2 1/ 2
are calculations confirm good efficiency of the use of basic
known boundary conditions. Finite Marchi-Zgrablich transform harmonic functions for resolving direct and inverse problems
and Fourier sine transform techniques are used to determine the of stationary heat conduction. Gao-Lian Liu [4] studied the
unknown temperature gradient, temperature distribution, inverse heat conduction problem with free boundary and
displacement and thermal stresses on outer curved surface of a
cylinder. The distribution of the considered physical variables
transformed into one with completely known boundary, which
are obtained and represented graphically. is much simpler to handle. As a by-product, the classical
Kirchhoffs transformation for accounting for variable
Keywords: Thermoelastic problem, semi-infinite hollow cylinder, conductivity is rederived and an invariance property of the
Thermal Stresses, inverse problem, Marchi-Zgrablich transform inverse problem solution with respect to variable conductivity
and Fourier sine transform.
is indicated. Then a pair of complementary extremum
I. INTRODUCTION principles is established on the image plane, providing a
sound theoretical foundation for the Ritzs method and finite
K hobragade et al. [1, 5-11] have investigated
temperature distribution, displacement function, and
stresses of a thin as well as thick hollow cylinder and
element method (FEM). An example solved by FEM is also
given.
Khobragade et al. [2] have established displacement function Michael J. Cialkowski [13] presented the application of heat
, temperature distribution and stresses of a semi-infinite polynomials for solving an inverse problem. The heat
cylinder. polynomials form the Treffetz Method for non-stationary heat
conduction problem. They have been used as base functions
Yoon Hwan Choi et. al. [16] discussed the temperature in Finite Element Method. Application of heat polynomials
distributions of the heated plate investigated with the permits to reduce the order of numerical integration as
condition that the line heating process was automatic. The compared to the classical Finite Element Method with
temperature variations were also investigated with the formulation of the matrix of system of equations. Gao-Lian
changes of those three variables. The numerical results Liu and Dao- Fang Zhang [3] discussed two methods of
showed that the peak temperature decreased as the moving solution generalized Ritz method and variable-domain
velocity of the heating source increased. It also revealed that FEM both capable of handling problems with unknown
the peak temperatures changed linearly with the changes of boundaries are suggested. Then, three sample numerical
the heating source. Xijing Li, Hongtan Wu, Jingwei Zhou examples have been tested. The computational process is
and Qun He [15] studied one-dimensional linear inverse heat quite stable, and the results are encouraging. This variational
problem. This ill-posed problem is replaced by the perturbed approach can be extended straightforwardly to 3-D inverse
problem with a non localized boundary condition. After the problems as well as to other problems in mathematical
derivation of its closed-form analytical solution, the physics. In the present problem, an attempt is made to study
calculation error can be determined by the comparison the three dimensional inverse transient thermoelastic
between the numerical and exact solutions. problems to determine the unknown temperature, temperature
Michael J. Cialkowski and Andrzej Frckowiak [12] distribution, displacement function and thermal stresses on
presented analysis of a solution of Laplace equation with the upper plane surface of a thin rectangular object occupying the

www.ijltemas.in Page 5
 International Journal of Latest Technology in Engineering, Management & Applied Science (IJLTEMAS)
Volume VI, Issue IV, April 2017 | ISSN 2278-2540

region D: a x a; b y b; 0 z h with known M r (T ,1, 0, b) H ( z, t ) (unknown) (7)

boundary conditions. Here Marchi-Fasulo transforms and
Laplace transform techniques have been used to find the M z (T ,1, 0, 0) 0) for all a r b , t 0 (8)
solution of the problem. M z (T ,1, 0, ) 0 for all a r b , t 0 (9)
In the present paper, an attempt is made to study the theoretical being:
solution for a thermoelastic problem to determine the
temperature distribution, displacement and stress functions of a M ( f , k , k , s ) (k f k f ) s
hollow cylinder with boundary conditions occupying the space
where the prime ( ^ ) denotes differentiation with respect to
D {( x, y, z ) R 3 : a ( x 2 y 2 )1 / 2 b, 0 z h} , where
, radiation constants are k and k on the curved surfaces
r ( x 2 y 2 )1 2 . A transform defined by Zgrablich et al. [2] of the plate respectively.
is used for investigation which is a generalization of Hankels
The radial and axial displacement U and W satisfy the
double radiation finite transform and used to treat the problem
uncoupled thermoelastic equation as Khobragade [9] are
with radiation type boundaries conditions.
U e 1 T
II.PROBLEM FORMULATION 2U (1 2 ) 1 2 at (10)
r 2 r 1 2 r
Consider a hollow cylinder as shown in the figure 1. The
e 1 T
material of the cylinder is isotropic, homogenous and all 2W (1 2 ) 1 2 at (11)
properties are assumed to be constant. We assume that the z 1 2 z
cylinder is of a small thickness and its boundary surfaces
remain traction free. The initial temperature of the cylinder is where
the same as the temperature of the surrounding medium,
U U W
which is kept constant. e (12)
r r r
The displacement function (r , z , t ) satisfying the

differential equation as Khobragade [9] is U , (13)
r
2 1 2 1
at T (1) W (14)
r 2 r r z 2 1 z

with 0 at ra and rb (2) The stress functions are given by

rz (a, z, t ) 0 , rz (b, z, t ) 0 , rz (r , 0, t ) 0 (15)
where and a t are Poisson ratio and linear coefficient of
thermal expansion of the material of the cylinder respectively r (a, z, t ) pi , r (b, z, t ) po , z (r, 0, t ) 0 (16)
and T (r , z, t ) is the heating temperature of the cylinder at
where pi and p o are the surface pressure assumed to be
time t satisfying the differential equation as Khobragade [9]
is uniform over the boundaries of the cylinder. The stress
functions are expressed in terms of the displacement
1 T 2T g (r , z, t ) 1 T components by the following relations as Khobragade [9] are
r r r r z 2
t
(3)
k
where K / c is the thermal diffusivity of the material of
the cylinder, K is the conductivity of the medium, c is its
specific heat and is its calorific capacity (which is
assumed to be constant) respectively, subject to the initial and
boundary conditions

M t (T ,1, 0, 0) F for all a r b , 0 z (4)

M r (T ,1, k1 , a) f1 ( z, t ) , for all 0 z , t 0 (5)

M r (T ,1, k2 , ) f 2 ( z, t ) for all 0 z , t 0 (6)
Figure 1: Geometry of the problem

www.ijltemas.in Page 6
 International Journal of Latest Technology in Engineering, Management & Applied Science (IJLTEMAS)
Volume VI, Issue IV, April 2017 | ISSN 2278-2540

U U W the Fourier sine transform parameter.

r ( 2G) (17)
r r z IV. DISPLACEMENT AND STRESS FUNCTION
W U U Substituting the value of temperature distribution from (22) in
z ( 2G) (18)
z r r equation (1) one obtains the thermoelastic displacement
function (r , z, t ) as
U U W
( 2G) (19)
r r z r 2 at 1 1
sin ( pz) S0 (k1 , k 2 , n r )
(r , z, t )
W U 2 1 n1 Cn m 0 n2
rz G
r z
2
t

e q t F eq t dt
2
(20) (24)
where 2G /(1 2 ) is the Lames constant, G is the shear 0
modulus and U, W are the displacement components. Using (24) in the equations (11) and (12) one obtains
Equations (1) - (20) constitute the mathematical formulation ra 1 1 sin ( pz) S0 (k1 , k2 , n r )
of the problem under consideration U t
1 n1 Cn m0 n2
III. SOLUTION OF THE OF THE PROBLEM
r 2 at 1 1
sin ( pz) S 0 (k1 , k 2 , n r )
Applying transform defined in [9] to the equations (3), (4) and
(6) over the variable r having p 0 with responds to the
2 1 n1 Cn m 0 n
2
t
boundary conditions of type (5) and taking Fourier cosine
e q t F eq t dt
2
transform , one obtains (25)
0
2
t

T * (n, m, t ) e p t F ep t dt r 2a 1 1 p cos( pz) S0 (k1 , k2 , n r )
2

(21)
W t
0
2 1 n1 Cn m0 n2
where constants involved T * (n, z, s) are obtained by using t
F e
q 2 t q 2 t
boundary conditions (6). Finally applying the inversion e dt (26)
theorems of transform defined in [9] and inverse Laplace 0
transform by means of complex contour integration and the Substitution the value of (26), (27) in (17) to (20) one obtains
residue theorem, one obtains the expressions of the the stress functions as
temperature distribution T (r , z, t ) and unknown temperature
a 1 1 sin ( pz) S0 (k1 , k2 , n r )
gradient H(z,t) for heating processes respectively as r 2G t
1 n1 Cn m0 n2
2
sin( pz) S 0 (k1 , k 2 , n r )
T (r , z , t )
n1

m 0 n2 2 n rS0 (k1 , k2 , n r )
r2 2
n S0(k1 , k2 , n r )
2
2
t

e q t F eq t dt
2

2
(22) t

e q t F eq t dt
2
0

sin( pz) S 0 (k1 , k 2 , nb)

0
2
H ( z, t ) a 1 1 sin ( pz) S 0 (k1 , k 2 , n r )
n1 m 0 n2 t
2 1 n1 Cn m0 n2
2
t

e q t F eq t dt
2
(23)
2S 0 (k1 , k 2 , n r ) r n S 0 (k1 , k 2 , n r )
0
r 2 at 1 1
p 2 sin ( pz) S 0 (k1 , k 2 , n r )
Where n are the roots of the transcendental equation
2 1 n1 Cn m 0 n2
2 2 u
q (p n )
n is the transformation parameter as defined in appendix, m is

www.ijltemas.in Page 7
 International Journal of Latest Technology in Engineering, Management & Applied Science (IJLTEMAS)
Volume VI, Issue IV, April 2017 | ISSN 2278-2540

t t
F e F e
q 2 t q 2 t
e q 2 t q 2 t
dt (27) e dt (30)
0 0
r 2 a 1 1 p 2 sin ( pz) S0 (k1 , k2 , n r )
z 2G t
V. SPECIAL CASE

2 1 n1 Cn m0 n2 Set F (r, t ) (1 e ) (r r0 )
t
(31)
2
t

e q t F eq t dt
2
Applying finite Marchi-Zgrablich transform defined in [9] to
0 the equation (31) one obtains

a 1 1 sin ( pz) F (n, t ) (1 e t ) r0 S0 (k1 , k2 , n r0 ) (32)

t
2 1 n1 Cn m0 n2 Substituting the value of (32) in the equations (21) to (31) one
obtains
2
r 2
0 1 2 n
S ( k , k , r ) 2 rS ( k , k , r ) S ( k , k , r )
n 0 1 2 n n 0 1 2 n
2
sin( pz) S 0 (k1 , k 2 , n r )
2 T (r , z , t )
n1

m 0 n2
at 1 1
2S 0 (k1 , k 2 , n r ) n rS 0 (k1 , k 2 , n r )

sin ( pz)
t
2 1 n1 C n n
F e
q 2 t q 2 t
e dt
2
m 0 (33)
2
t
0
e q t F eq t dt
2
(28)
2
sin( pz) S 0 (k1 , k 2 , nb)
H ( z, t )
0

n1 m 0 n2
a 1 1 sin ( pz)
2G t
n2 2
t
2 1 n1 Cn m0
e q t F eq t dt
2
(34)
2S0 (k1, k2 , nr) nrS0 (k1, k2 , nr) 0
t VI. NUMERICAL RESULTS, DISCUSSION AND REMARKS
F e
q 2 t q 2 t
e dt
To interpret the numerical computation we consider material
0
properties of low carbon steel (AISI 1119), which can be used
a 1 1
sin ( pz) for medium duty shafts, studs, pins, distributor cams, cam
t shafts, and universal joints having mechanical and thermal
1 n 1 Cn m 0 n2 properties

r2 2 13.97[ m / s 2 ] 0.29, 51.9 [W /(m K )] and

S0 (k1 , k 2 , n r ) 2 n rS0 (k1 , k 2 , n r ) n S0(k1 , k2 , n r at 14.7 m / m 0 C .
2
Setting the physical parameter with a 2.5 , b 3 and
r 2 a 1 1 p 2 sin ( pz) S0 (k1 , k2 , n r ) h 500 . k1 0.25 , k2 0.25 , t 1sec .
t
2 1 n1 Cn m0 n2 VII. CONCLUSION
t
e q F e
q 2 t
dt
2
t
(29) In this paper, we modify the conceptual idea proposed by
0 Khobragade et al. [9] for hollow cylinder and the
temperature distributions, displacement and stress functions
1 p cos ( pz) on the curved surface z b occupying the region of the
rat 1
rz G
2
C n m 0
cylinder a r b, 0 z have been obtained with the

1 n1
n
known boundary conditions. We develop the analysis for the
temperature field by introducing the transformation defined
2S0 (k1, k2 , n r ) nrS0 (k1, k2 , nr ) by Zgrablich et al, finite Fourier cosine transform techniques
with boundary conditions of radiations type. The series
solutions converge provided we take sufficient number of
terms in the series. Since the thickness of cylinder is very

www.ijltemas.in Page 8
 International Journal of Latest Technology in Engineering, Management & Applied Science (IJLTEMAS)
Volume VI, Issue IV, April 2017 | ISSN 2278-2540

small, the series solution given here will be definitely ACKNOWLEDGEMENT

convergent. Assigning suitable values to the parameters and
The authors are thankful to University Grant Commission,
functions in the series expressions can derive any particular
New Delhi for providing partial financial assistance under
case. The temperature, displacement and thermal stresses that
Minor Research Project Scheme.
are obtained can be applied to the design of useful structures
or machines in engineering applications. REFERENCES
[1] Dange W K; Khobragade N W and Durge M H (2012): Thermal
APPENDIX Stresses of a finite length hollow cylinder due to heat generation,
Int. J. of Pure and Appl. Maths,
Finite Marchi-Zgrablich Integral Transform: [2] Gahane T T and Khobragade N W (2012): Transient
Thermoelastic Problem of A Semi-infinite Cylinder with Heat
The finite Marchi-Zgrablich integral transform of f (r ) is Sources, Journal of Statistics and Mathematics, Vol. 3, Issue 2, pp.
defined as 87-93, BIO INFO Publication.
[3] Gao-Lian Liu and Dao-Fang Zhang.: Numerical methods for
b
f p (m) r f (r ) S p ( , , m r ) dr
inverse problem of heat conduction with unknown boundary based
(A) on variational principles with variable domain, Journal of Thermal
a
Science, Volume 1, Number 4, 241248, 1992.
where 1 , 2 , 1 and 2 are the constants involved in the
[4] Gao-Lian Liu: A novel variational formulation of inverse problem
of heat conduction with free boundary on an image plane, Journal
boundary conditions of Thermal Science, Volume 5, Number 2, 8892, 1996.
[5] Khobragade N W (2013): Thermal stresses of a hollow cylinder
1 f (r ) 2 f (r ) r a 0 and 1 f (r ) 2 f (r ) r b 0 with radiation type conditions, Int. J. of Engg. And Innovative
Technology, vol. 3, Issue 5, pp.25-32.
[6] Khobragade N W (2013): Thermoelastic analysis of a solid
for the differential equation circular cylinder, Int. J. of Engg. And Innovative Technology, vol.
f (r ) (1 r ) f (r ) ( p / r ) f (r ) 0 , f p (n) is the transform
2 2 3, Issue 5, pp. 155-162.
[7] Khobragade N W (2013): Thermoelastic analysis of a thick hollow
of f (r ) with respect to kernel S p ( , , m r ) and weight cylinder with radiation conditions, Int. J. of Engg. And Innovative
Technology, vol. 3, Issue 4, pp.380-387.
function r [8] Khobragade N W and Patil V: Some Aspects of Thermoelastic
The inversion of equation (A) is given by Problems on Different Solids LAP LAMBERT Academic
Publishing, Germany, ISBN: 978-3-330-061412-2, 2017.
f p (m) S p ( , , m r ) [9] Khobragade N W: Some Thermoelastic Problems on a Cylinder,
f (r ) b 2 LAP LAMBERT Academic Publishing, Germany, ISBN: 978-3-
m 1
a [r S p ( , , m r )] dr 659-97446-5, 2016
[10] Khobragade N W: Thermoelastic Problems on Circular Plate and
where kernel function S p ( , , m r ) can be defined as Annular Disc, LAP LAMBERT Academic Publishing, Germany,
ISBN: 978-3-330-01018-5, 2016
S p ( , , m r ) J p (mr ) [Yp ( , ma) Yp ( , mb)] [11] Lamba N K and Khobragade N W (2012): Integral transform
methods for inverse problem of heat conduction with known
boundary of a thin rectangular object and its stresses, processes,
Yp (mr ) [ J p ( , ma) J p ( , mb)] Journal of Thermal Science, Volume 21, Number 5, 459465,
2012.
and J p ( r ) and Y p ( r ) are Bessel function of first and [12] Michael J Cialkowski and Andrzej Frckowiak.: Solution of the
stationary 2D inverse heat conduction problem by Treffetz method,
second kind respectively. Journal of Thermal Science, Volume 11, Number 2, 148162,
2002.
OPERATIONAL PROPERTY: [13] Michael J Cialkowski.: New type of basic functions of FEM in

r f (r ) (1 / r ) f (r ) ( p
b application to solution of inverse heat conduction problem, Journal
2 2
/ r 2 ) f (r ) S p ( , , m r ) dr of Thermal Science, Volume 11, Number 2, 163171, 2002.
a [14] Noda N; Hetnarski R B and Tanigawa Y: Thermal Stresses,
(b 2 ) S p ( , , mb) 1 f (r ) 2 f (r ) r b second edition Taylor & Francis, New York (2003), 260.
[15] Xijing Li, Hongtan Wu, Jingwei Zhou and Qun He.: Calculation
error of numerical solution for a boundary-value inverse heat
(a / 2 ) S p ( , , m a)1 f (r ) 2 f (r ) r a m
2
f p ( m) conduction problem, Journal of Thermal Science, Volume 5, No. 1,
2427, 1996.
[16] Yoon Hwan Choi; Yeon Won Lee; Kwang Choi; Deog Hee Doh
and Kyoung Joon Kim: Temperature distribution and thermal
stresses in various conditions of moving heating source during line
heating process, Journal of Thermal Science, Volume 21, Number
1, 8287, 2012.

www.ijltemas.in Page 9