Faculty of Engineering

Alexandria University
Mechanical Engineering Department
Heat Transfer Branch

Numerical Analysis
MEC 155
Lecture (3)
Reference; Heat and mass transfer, Fundamentals & Applications, 5th
edition, by Yunus A. Çengel and Afshin J. Ghajar

Prof. Mohamed Khamis

 Numerical Methods In Heat Conduction

Many problems encountered in practice involve complicated geometries with

complex boundary conditions or variable properties, and cannot be solved
analytically. In such cases, sufficiently accurate approximate solutions can be
obtained by computers using a numerical method.

Analytical solution methods are based on solving the governing differential

equation together with the boundary conditions. They result in solution
functions for the temperature at every point in the medium.

Numerical methods are based on replacing the differential equation by a set

of n algebraic equations for the unknown temperatures at n selected points in
the medium (discrete points).
 Numerical Methods In Heat Conduction
There are several ways of obtaining the numerical formulation of a heat
conduction problem, such as;

- the finite difference method,

- the finite element method,
- the boundary element method, and
- the energy balance (or control volume) method.

In this chapter, we use primarily the finite difference approach.

 Numerical Methods In Heat Conduction
Finite Difference Formulation Of Differential Equations

Finite difference method is done by replacing the derivatives by differences.

For example; consider a function f that depends

on x, as shown in Figure.

the first derivative of f(x) at a point is equivalent

to the slope of a line tangent to the curve at that
point and is defined as;

which is the ratio of the increment f of the function to the increment x of

the independent variable as x 0.
 Numerical Methods In Heat Conduction
Finite Difference Formulation Of Differential Equations

If we don’t take the indicated limit, we will have the following approximate
relation for the derivative:

This approximate expression of the derivative in terms of differences is the

finite difference form of the first derivative. The equation above can also be
obtained by writing the Taylor series expansion of the function f about the
point x,

and neglecting all the terms in the expansion except the first two.
 Numerical Methods In Heat Conduction
Finite Difference Formulation Of Differential Equations

• The first term neglected is proportional to x2, and thus the error involved
in each step of this approximation is also proportional to x2.
• Therefore, the smaller the x, the smaller the error, and thus the more
accurate the approximation.

• Any body can be divided into number of elements centered by nodal points.
These elements are normally equal except the two elements adjacent to
boundaries.
• The finite difference technique is used to convert the differential equations
of the model to a group of linear or non-linear equations. The number of these
equations equals to the number of model nodal points having the same number
of unknowns. Solving this set of equation (simultaneous equations) can be
carried out using any mathematical techniques such as matrices or iteration
methods.
 Numerical Methods In Heat Conduction
Finite Difference Formulation Of Differential Equations

• Two dimensional model with 5 x 5 nodal points

9 internal nodal points.
16 boundary nodal points.

• One dimensional model with 5 nodal points

3 internal nodal points.
2 boundary nodal points.
 Numerical Methods In Heat Conduction
Finite Difference Formulation Of Differential Equations

• Now consider steady, one-dimensional heat conduction in a plane wall of

thickness L. The wall is subdivided into (M-1) sections of equal thickness x =
L/(M-1) in the x-direction, separated by planes passing through M points 1, 2, .
. . , m - 1, m, m + 1, . . . , M called nodes or nodal points, as shown in Figure.

• The x-coordinate of any point m is simply

𝑥𝑚 = 𝑚 − 1 ∆𝑥, and the temperature at that
point is simply T(𝑥𝑚 ) = 𝑇𝑚 .
 Numerical Methods In Heat Conduction
Finite Difference Formulation Of Differential Equations

The heat conduction equation involves the second derivatives of temperature

with respect to the space variables, such as d2T/dx2, and the finite difference
formulation is based on replacing the second derivatives by appropriate
differences.
The first derivative of temperature dT/dx at the
midpoints m – ½ and m + ½ of the sections
surrounding the node m can be expressed as;
 Numerical Methods In Heat Conduction
Finite Difference Formulation Of Differential Equations

Noting that, the second derivative is simply the derivative of the first
derivative, the second derivative of temperature at node m can be expressed as;

Which is the finite difference representation of the second derivative at a

general internal node m. While, the equations of external nodes (1 and M)
presented by the boundary conditions.
 Numerical Methods In Heat Conduction
Iteration Technique

The finite difference formulation of steady heat conduction problems

usually results in a system of N algebraic equations in N unknown nodal
temperatures that need to be solved simultaneously.

There are numerous systematic

approaches available, and they are
classified as direct and iterative
methods.
In this section we develop the finite difference formulation of heat conduction
in a plane wall using the energy balance approach and discuss how to solve the
resulting equations.

I- One-dimensional Steady Heat Conduction Without Heat

Generation, Through Plane Wall.

Consider a plane wall of thickness L, thermal

conductivity k, and surface area A.

For Steady State, no heat generation;

𝑑2 𝑇
2
=0
𝑑𝑥
I- One-dimensional Steady Heat Conduction Without Heat
Generation Through Plane Wall.

Numerical Solution Procedure:

1. Select the nodal points (or nodes) M at which

the temperatures are to be determined and then
forming elements (or control volumes) over the
nodes by drawing lines through the midpoints
between the nodes.

2. Calculate x from;
𝐿
∆𝑥 =
𝑀−1
3. Convert the differential equation into (M-2) algebraic equations for the
internal nodes by using the finite difference method.

𝑑2 𝑇
2
=0
𝑑𝑥

𝑇𝑚−1 + 𝑇𝑚+1
Thus, 𝑇𝑚 = ,
2

4. Write the equations for the external nodes (1 and M) from the boundary
condition at each surface.

5. Apply the iteration technique until achieving the convergence condition

(error  specified error at each node).
 Thomas Algorithm
𝑎𝑖 𝑋𝑖−1 + 𝑏𝑖 𝑋𝑖 + 𝐶𝑖 𝑋𝑖+1 = 𝑑𝑖

𝑎𝑖 = 0 𝐶𝑛 = 0
Gauss-Seidel iterative method
 Applied Topics
5. Apply the iteration technique (Gauss-Seidal method) until achieving the
convergence condition (error  specified error at each node).
𝑇𝑚−1 + 𝑇𝑚+1
𝑇1 = 30 𝑇𝑚 = 𝑇9 = 𝑇8 + 3
Error = 0.01 2

Node 1 2 3 4 5 6 7 8 9
Iteration
Initial guess 30 50 50 50 50 50 50 50 50
1 30 40 50 50 50 50 50 50 53
2 30 40 45 50 50 50 50 51.5 53
3 30 37.5 45 47.5 50 50 50.75 51.5 54.5
30
30
30
236 30 33.01 36.02 39.04 42.04 45.05 48.05 51.06 54.06
237 30 33.01 36.02 39.03 42.05 45.05 48.06 51.05 54.06
Error = -0.00195 0.002947 -0.00556 0.005445 -0.00833 0.007114 -0.00982 0.0077
Example 1
 Example 2
Consider steady heat conduction in a plane wall having a thickness of L=4 cm
and thermal conductivity of k = 2 W/m·K, whose left surface is maintained at
30°C while the right surface is subjected to a heat flux of 1200 W/m2.
Determine numerically the temperature distribution within the wall for the
case of no heat generation. Also obtain the rate of heat transfer at the left
boundary.

Solution;

1. Select the nodal points M

M
take M = 9

2. Calculate x from;
𝐿 0.04
∆𝑥 = = = 0.005 𝑚
𝑀−1 9−1
 Applied Topics
3. Convert the differential equation into (M-2) algebraic equations for the
internal nodes by using the finite difference method.

𝑑2 𝑇
=0
𝑑𝑥 2

𝑇𝑚−1 + 𝑇𝑚+1
𝑇𝑚 = 𝑇2 = 0.5 (𝑇1 + 𝑇3 )
2
𝑇3 = 0.5 (𝑇2 + 𝑇4 )
𝑇4 = 0.5 (𝑇3 + 𝑇5 )
7 equations for
𝑇5 = 0.5 (𝑇4 + 𝑇6 ) 7 internal nodes
𝑇6 = 0.5 (𝑇5 + 𝑇7 )
𝑇7 = 0.5 (𝑇6 + 𝑇8 )
𝑇8 = 0.5 (𝑇7 + 𝑇9 )
 Applied Topics
4. Write the equations for the external nodes (1 and M) from the boundary
condition at each surface.

at (x=0) 𝑇1 = 30 °C (specified temperature)

𝑑𝑇
at (x=L =0.04 m) −𝑘 = −1200 (specified heat flux)
𝑑𝑥

𝑑𝑇 𝑇𝑀 − 𝑇𝑀−1 𝑇9 − 𝑇8
Where, = =
𝑑𝑥 ∆𝑥 0.005

Therefor,

𝑇9 − 𝑇8
−2 = −1200 𝑇9 = 𝑇8 + 3
0.005
 Applied Topics
5. Apply the iteration technique (Gauss-Seidal method) until achieving the
convergence condition (error  specified error at each node).
𝑇𝑚−1 + 𝑇𝑚+1
𝑇1 = 30 𝑇𝑚 = 𝑇9 = 𝑇8 + 3
Error = 0.01 2

Node 1 2 3 4 5 6 7 8 9
Iteration
Initial guess 30 50 50 50 50 50 50 50 50
1 30 40 50 50 50 50 50 50 53
2 30 40 45 50 50 50 50 51.5 53
3 30 37.5 45 47.5 50 50 50.75 51.5 54.5
30
30
30
236 30 33.01 36.02 39.04 42.04 45.05 48.05 51.06 54.06
237 30 33.01 36.02 39.03 42.05 45.05 48.06 51.05 54.06
Error = -0.00195 0.002947 -0.00556 0.005445 -0.00833 0.007114 -0.00982 0.0077
 Applied Topics

Maximum error = 0.00982  0.01 (specified error), which achieved after 237
iteration.

As a result, the temperatures distribution at the described points becomes;

node 1 2 3 4 5 6 7 8 9
x (cm) 0 0.5 1 1.5 2 2.5 3 3.5 4
T (°C ) 30 33.01 36.02 39.03 42.05 45.05 48.06 51.05 54.06

𝑑𝑇
The rate of heat transfer at the left boundary: 𝑞𝑙𝑒𝑓𝑡 = −𝑘
𝑑𝑥

𝑑𝑇 𝑇2 − 𝑇1 33.01 − 30
Where, = 𝑞𝑙𝑒𝑓𝑡 = −2 = −1204 𝑊/𝑚2
𝑑𝑥 ∆𝑥 0.005
Example 3
The outer surface of an engine is situated in a place where oil leakage can occur.
Some oils have auto-ignition temperatures of approximately above 250°C. When
oil comes in contact with a hot engine surface that has a higher temperature than
its auto-ignition temperature, the oil can ignite spontaneously. Treating the engine
housing as a plane wall, the inner surface (x = 0) is subjected to 6 kW/m2 of heat.
The engine housing (k = 13.5 W/m∙K) has a thickness of 1 cm, and the outer
surface (x = L) is exposed to an environment where the ambient air is 35°C with a
convection heat transfer coefficient of 20 W/m2∙K. To prevent fire hazard in the
event the leaked oil comes in contact with the hot engine surface, the temperature
of the engine surface should be kept below 200°C. Determine the variation of
temperature in the engine housing and the temperatures of the inner and outer
surfaces. Is the outer surface temperature of the engine below the safe
temperature?
Example 5
Example 6
Thank you