Computational and Analytical

Analysis of Heat Conduction

Through a Slab
Nishad Vijay, MTech student (22MEM1R22)
Thermal Engineering, Department of Mechanical Engineering, National Institute of Technology,
Warangal, Telangana, India-506004.

Abstract
In this work, two-dimensional analysis is used to study heat conduction through a slab under
certain assumptions such as steady state condition, no heat generation with a slab and only one non-
homogenous boundary condition. The temperature distribution comparison was done between the
computational result obtained from ANSYS Fluent 2022 R2 and Analytical results using the heat
transport equation as governing equation.

Keywords: Two-dimension, ANSYS, Variable separable methods, Heat transport equation, Temperature
distributions, MATLAB.
 1. Introduction

The conduction heat transfer is the heat transport from highly energetic particles to less energetic
particles or from high temperature to lower temperature within a substance. There are two states of
conduction such as steady state and non-steady state heat conduction. Steady state conduction is
observed when temperature at a given point does not change with respect to time provided the system
is under equilibrium conditions. Non-steady state conduction is observed when the temperature at a
given point does change with respect to time. Non-steady is most realistic heat conduction but it is
difficult to solve governing equation to extract solution from it. The term two dimensional is used in the
heat conduction problems when only two space coordinates are required to describe temperature
distribution at any point within the heat conducting body.

This study aims to analysis the theory behind two-dimensional steady state heat conduction with
no heat generation in a rectangular-shaped conducting body. This is first done by using the heat
transport equation as governing equation, a problem and analytical solution is taken from J.P Holman,
Heat Transfer [1] and calculated using MATLAB R2021b [2]. Also, computational analysis is done using
ANSYS Fluent 2022 R2[3]. A brief comparison between the results of analytical methods and
computational methods is discussed in the present work.

Nomenclature:

W Width of plate, m
L Length of plate, m
T Temperature, K
𝑇ℎ Hot surface temperature, K
𝑇𝑐 Cold surface temperature, K
A, B, C, D, E Integration constant
λ Constant of variable separable method
 2. Solution Methodology

2.1. Geometry Model

A Geometric model of two-dimensional flat plate in system is considered in present study is

show in figure 1. The length of plate in x-direction is, L=1m and length of plate in y-direction is W=1m.

The heat transfer through two-dimensional

plate is complex problem. To make simple,
temperatures of right, left, bottom side of plate is
taken as constant 𝑇𝑐 = 323𝐾 and top side can have
temperature distribution impressed on it as 𝑇 =
𝑓(𝑥). Which can be non-homogeneous function. For
our study, we taking it as constant 𝑇 = 𝑇ℎ = 373𝐾.

The computational method for present study used

ANSYS Workbench in ANYSY Fluent 2022 R2[3] and
model into ANSYS Design Modeler. The efficiency of
thermal CDF simulation depends on many factors
such as creation of geometry, grid generation, choice Figure 1: Geometry parameters of problem
of suitable numerical methods.

2.2. Governing equations

The governing equations for heat transfer in two-dimensional plate were solved in cartesian
coordinates system because CFD model for present study is solved in coordinates system in ANSYS
Fluent 2022 R2.

General steady state two-dimensional heat conduction equation with no heat generation is
given as follow and furthermore, it has 4 integration constants as there are two second order partial
differential equation.
𝜕2 𝑇 𝜕2 𝑇
∴ ( 𝜕𝑥 2 + 𝜕𝑦2 ) = 0 (1)

It is necessary to note that above equation is independent of thermal conductivity and density of
material. Therefore, present study is independent of material and depend only on x and y spatial
coordinates.

2.3. Boundary conditions

Computations was carried out under assumptions such as steady state, two-dimensional, no
heat generation and no homogenous temperature distribution. As per geometric model, following
boundary condition can be obtained.

1. Left side of plate: At x=0, for all values for y, 𝑇 = 𝑇𝑐 = 323𝐾

2. Bottom side of plate: At y=0, for all values for x, 𝑇 = 𝑇𝑐 = 323𝐾
3. Right side of plate: At x=L, for all values for y, 𝑇 = 𝑇𝑐 = 323𝐾
4. Top side of plate: At y=W, for all values for x, 𝑇 = 𝑇ℎ = 373
On simplifying the boundary condition with 𝜃 = 𝑇 − 𝑇𝑐
New boundary condition we get,
1. Left side of plate: At x=0, for all values for y, 𝜃 = 0
2. Bottom side of plate: At y=0, for all values for x, 𝜃 = 0
3. Right side of plate: At x=L, for all values for y, 𝜃 = 0
4. Top side of plate: At y=W, for all values for x, 𝜃 = 50

2.4. Analytical method

To solve governing equation 1, the separation of variables method is used in present study
provided there is only one non-homogeneity in boundary condition.

𝑇 = 𝑃(𝑥)𝑄(𝑦) or 𝜃 = 𝑃(𝑥)𝑄(𝑦) (2)

Differentiating above equation with respect to x and y

we can get,
𝜕2 𝑇 𝜕2 𝑃 𝜕2 𝑇 𝜕2 𝑄
∴ 𝜕𝑥 2 = 𝜕𝑥 2 (𝑄) and 𝜕𝑦 2
= 𝜕𝑦2 (𝑃)

And substituting back into equation 1

𝜕2 𝑃 𝜕2 𝑄
∴ 𝜕𝑥 2 (𝑄) + 𝜕𝑦2 (𝑃) = 0

If we observe carefully, as per our boundary condition, temperature distribution along y-direction need
to be exponential and temperature distribution along x-direction need to be periodic. Therefore, P(x)
should give periodic solution and Q(y) should give exponential solution.

As per separation of variables method, we can write,

𝜕2 𝑃 1 𝜕2 𝑄 1
∴ (−1) 𝜕𝑥 2 (𝑃) = 𝜕𝑦2 (𝑄) = 𝜆2

Above equation gives, two ordinary differential equations

𝜕2 𝑃 𝜕2 𝑄
∴ − 𝜆2 𝑃 = 0 and + 𝜆2 𝑄 = 0
𝜕𝑥 2 𝜕𝑦 2

Solutions of ordinary differential equation is given by

𝑃 = 𝐴 𝑠𝑖𝑛𝜆𝑥 + 𝐵 cos 𝜆𝑥 and 𝑄 = 𝐶 sinh 𝜆𝑦 + 𝐷 cosh 𝜆𝑦

Therefore, we can write general solution as follow

𝜃 = (𝐴 𝑠𝑖𝑛𝜆𝑥 + 𝐵 cos 𝜆𝑥)(𝐶 sinh 𝜆𝑦 + 𝐷 cosh 𝜆𝑦) (3)

On applying boundary conditions

1. Left side of plate: At x=0, for all values for y, 𝜽 = 𝟎

Applying 1st boundary condition to equation 3, we get B=0

2. Bottom side of plate: At y=0, for all values for x, 𝜽 = 𝟎

Applying 2nd boundary condition to equation 3, we get D=0

General solution becomes,

∴ 𝜃 = (𝐴 𝑠𝑖𝑛𝜆𝑥 + 𝐵 cos 𝜆𝑥)(𝐶 sinh 𝜆𝑦 + 𝐷 cosh 𝜆𝑦)

∴ 𝜃 = (𝐴 𝑠𝑖𝑛𝜆𝑥)(𝐶 sinh 𝜆𝑦)
∴ 𝜃 = 𝐸(𝑠𝑖𝑛𝜆𝑥)(sinh 𝜆𝑦) (Take E=AC) (4)

3. Right side of plate: At x=L, for all values for y, 𝜽 = 𝟎

Applying 3rd boundary condition to equation 4, we get,

∴ 𝜃 = 𝐸(𝑠𝑖𝑛𝜆𝑥)(sinh 𝜆𝑦)
∴ 0 = 𝐸(𝑠𝑖𝑛𝜆𝐿)(sinh 𝜆𝑦)
We get,

∴ 𝑠𝑖𝑛𝜆𝐿 = 0
𝑛𝜋
For 𝜆 values, 𝜆 = where n=1,2,3,4……
𝐿

Using property theorem, General solution becomes

𝑛𝜋 𝑛𝜋
𝜃 = ∑∞
𝑛=1 𝐸𝑛 (𝑠𝑖𝑛 𝐿
𝑥)(sinh 𝐿 𝑦) (5)

4. Top side of plate: At y=W, for all values for x, 𝜽 = 𝟓𝟎

Using simplified equation 5 and applying 4th boundary condition We get,

𝑛𝜋 𝑛𝜋
∴ 50 = ∑∞
𝑛=1 𝐸𝑛 (𝑠𝑖𝑛 𝐿
𝑥)(sinh 𝐿 𝑊)
𝑛𝜋
To solve above equation, Property of orthogonality is used as 𝑠𝑖𝑛 𝐿
𝑥 is orthogonal function.

We can write,
𝐿 𝑛𝜋 𝑛𝜋
∫0 (𝑇ℎ −𝑇𝑐 )(𝑠𝑖𝑛 𝐿 𝑥)(sinh 𝐿 𝑊)𝑑𝑥
∴ 𝐸𝑛 = 2
𝐿 𝑛𝜋 𝑛𝜋
∫0 ((𝑠𝑖𝑛 𝐿 𝑥)(sinh 𝐿 𝑊)) 𝑑𝑥

𝐿 𝑛𝜋
𝑇ℎ −𝑇𝑐 ∫0 𝑠𝑖𝑛 𝐿 𝑥 𝑑𝑥
∴ 𝐸𝑛 = 𝑛𝜋 𝐿 𝑛𝜋 2
sinh 𝑊 ∫ (𝑠𝑖𝑛 𝑥) 𝑑𝑥
𝐿 0 𝐿

𝐿 𝑛𝜋
𝑇ℎ −𝑇𝑐 ∫0 𝑠𝑖𝑛 𝐿 𝑥 𝑑𝑥
∴ 𝐸𝑛 = 𝑛𝜋
sinh 𝑊 ∫𝐿(𝑠𝑖𝑛𝑛𝜋𝑥)2 𝑑𝑥
𝐿 0 𝐿

𝑇ℎ −𝑇𝑐 2(1−(−1)𝑛 )
∴ 𝐸𝑛 = 𝑛𝜋
𝑛𝜋
sinh 𝑊
𝐿

Final general solution becomes, we will call it as “Analytical solution”

∞ 𝒏𝝅 𝒏𝝅
𝟐(𝑻𝒉 − 𝑻𝒄 ) 𝟐(𝟏 − (−𝟏)𝒏 ) (𝒔𝒊𝒏 𝑳 𝒙) (𝐬𝐢𝐧𝐡 𝑳 𝒚)
∴𝜽= ∑ 𝒏𝝅
𝝅 𝒏 𝐬𝐢𝐧𝐡 𝑳 𝑾
𝒏=𝟏
2.4. Grid independence study

A two-dimensional grid was generated in ANSYS Fluent to solve for temperature field from heat
transport equation as governing equation. In present study, structural grid was generated using sizing
on each side of plate and face facing scheme with quadrilateral elements. In order to establish grid
independency and arrive to least number of elements can yield accurate computational results, four
different grid size were selected for discretizing the domain of the flat plate.

Table 1 summarize the Grid independence study and temperature values for four different grids
at center of plate/slab respectively. On comparing computational results with analytical results. The
error percentage was calculated and the value of error percentage for all four selected grids are nearly
same and equal to zero. The least error percentage was found with 20x20. Therefore, after grid
independency, 20x20 was accepted for present study.

Table 1: Grid independence study

Grid Computational Analytical Percentage Error

temperature in K Temperature in K In %
5x5 335.4946 0.00298
10x10 335.4898 335.5000 0.00304
20x20 335.5182 0.00244
100x100 335.5268 0.00799

2.5. Model validation

To validate computational results, the analytical solution is given in J.P Holman, Heat Transfer
[1] book. In this book, the same heat conduction study was done on a two-dimensional plate in the
cartesian coordinate system by keeping only one non-homogeneous boundary condition at top of
the plate and keeping the other three sides of the plate at the same temperature. With help of the
final analytical equation, temperature values were evaluated and later, it was compared with
computational results obtained from ANSYS Fluent 2021 R2[3]
 3. Results and Discussions

Fig 2.1. show. the temperature contours for entire two-dimensional plate in cartesian
coordinates system was generated using ANSYS Fluent 2022 R2[3].

Figure 2.1: Temperature contour

Using an analytical solution, temperature

distribution along the x-direction and y-direction
can be plotted on a graph. To avoid truncation
error of higher order terms in the analytical
solution, MATLAB R2021b [2] is used for the
calculation of temperatures along x-direction and
y-directions upto n=20 terms of the analytical
solution. The result obtained from ANSYS Fluent is
compared against the result obtained from
analytical solutions.

Fig 2.2 show, temperature distribution along x-

Figure 2.2 Temperature distribution along x-direction
direction such that it passes through center
(0.5,0.5). Both ANSYS Fluent and Analytical method
have periodic solutions and both curves are
coinciding with each other and also, both curves
are symmetric about x=0.5 line.

Fig 2.3 show, temperature distribution along y-

direction such that it passes through center
(0.5,0.5). Both ANSYS Fluent and Analytical method
have exponential solutions and both curves are
coinciding with each other.

Figure 2.3: Temperature distribution along y-direction

 4. Conclusion
The computational method using ANSYS Fluent 2022 R2[3] and Analytical method of two-dimensional
plate (under certain assumptions like steady state condition, no heat generation and only one non-
homogeneous boundary Condition) was calculated in present study.

The computational results were compared with analytical solution and temperature distribution along
the x-direction and the y-direction such that it passes through center of plate.

It was observed that:

1. if non-homogenous boundary condition along the y-direction then, temperature distribution

along the y-direction is exponential and temperature distribution along the x-direction is
periodic
2. if non-homogenous boundary condition along the x-direction and keeping other conditions
same. temperature distribution along x-direction is exponential and temperature distribution
along the y-direction is periodic.

5. References

1. J.P Holman, Heat Transfer, sixth edition.

2. MATLAB R2021b, https://in.mathworks.com/products/matlab.html.
3. ANYSY Fluent 2022 R2 (ANSYS Inc., Canonsburg, PA, USA). https://www.ansys.com/en-
in/academic/students