Adama Science and Technology University

Heat Transfer (MEng3208)

CHAPTER THREE: TWO DIMENSTIONAL HEAT CONDUCTION

Samuel G/mariam
March 2025
TOPICS
• Alternative Approaches
• The Method of Separation of Variables
• The Conduction Shape Factor and the Dimensionless
Conduction Heat Rate
• Finite-Difference Equations
• Solving the Finite-Difference Equations
Alternative Alternative Approaches to 2D Conduction
Approach
• Recall that, in any conduction analysis, there exist two major objectives. The 1st to
determine the temperature distribution T(x, y) and achieved by solving the appropriate
form of the heat equation. And the 2nd objective is to determine the heat flux components
and achieved by applying the rate equations.
• For two-dimensional, steady-state conditions with no heat generation and constant
thermal conductivity, this form is,

…..Equation 3.1

Two-dimensional conduction.
• Methods for solving Equation 3.1 the heat diffusion equation, include the use of
analytical, graphical, and numerical (finite-difference, finite-element or boundary-
element) approaches.

• The analytical method involves effecting an exact mathematical solution to Equation

3.1. The problem is more difficult than those considered in Chapter 2, since it now
involves a partial, rather than an ordinary, differential equation.

• To illustrate the nature and importance of analytical techniques, an exact solution is

obtained by the method of separation of variables

• In contrast to the analytical methods, which provide exact results at any point, graphical
and numerical methods can provide only approximate results at discrete points
 The Method of Separation of Variables

• Let we consider a system mentioned below.

• Three sides of a thin rectangular plate or a long rectangular rod are maintained at a
constant temperature T1, while the fourth side is maintained at a constant temperature T2
≠ T1.
• We are interested in the temperature distribution T(x, y), but to simplify the solution we
introduce the transformation.

………Equ. 3.2

Substituting Equ. 3.2 into Equ. 3.1, the transformed

differential equation is then

………Equ. 3.3
Since the equation is second order in both x and y, two boundary conditions are needed for
each of the coordinates. They are:

• Note that, through the transformation of θ, three of the four boundary conditions are now
homogeneous and the value of θ is restricted to the range from 0 to 1.
• We now apply the separation of variables technique by assuming that the desired
solution can be expressed as the product of two functions, one of which depends only on
x while the other depends only on y.
• That is, we assume the existence of a solution of the form
………Equ. 3.4
Substituting into Equation 3.3 and dividing by XY, we obtain
………Equ. 3.5
 and it is evident that the differential equation is, in fact, separable. That is, the left-hand side
of the equation depends only on x and the right-hand side depends only on y.
• Hence the equality can apply in general (for any x or y) only if both sides are equal to the
same constant. Identifying this, as yet unknown, separation constant as , we then have

………Equ. 3.6 and the partial differential equation has been reduced to
two ordinary differential equations. Note that the
designation of λ2 as a positive constant was not
………Equ. 3.7 arbitrary.

• The general solutions to the above equations are, respectively

in which case the general form of the two-dimensional solution is

………Equ. 3.8
• Applying the condition that θ(0, y) = 0, it is evident that C1 = 0. In addition from the requirement
that θ(x, 0) = 0, we obtain

• which may only be satisfied if C3 = - C4. Although the requirement could also be satisfied by
having C2 = 0, this would result in θ (x, y) = 0, which does not satisfy the boundary condition θ
(x, W) = 1. If we now invoke the requirement that θ (L, y) = 0, we obtain

• The only way in which this condition may be satisfied (and still have a nonzero solution) is by
requiring that λ assume discrete values for which sin λL = 0. These values must then be of the
form
………Equ. 3.9 the desired solution may now be expressed as

where the integer n=0 is precluded, since it implies ϴ (x, y)=0. The desired solution may now be
expressed as
………Equ. 3.10
• Combining constants and acknowledging that the new constant may depend on n, we obtain
 where we have also used the fact that (eny/L – e-ny/L) = 2 sinh(ny/L).

• In this form we have really obtained an infinite number of solutions that satisfy the differential
equation and boundary conditions.

• However, since the problem is linear, a more general solution may be obtained from a superposition of
the form
………Equ. 3.11

• To determine Cn we now apply the remaining boundary condition, which is of the form

………Equ. 3.12

• After some simplification, the value of Cn

………Equ. 3.13
 ………Equ. 3.13

we then obtain for the final solution

The Conduction Shape Factor and the Dimensionless Conduction Heat Rate
• In general, finding analytical solutions to the two- or three-dimensional heat equation is
time-consuming and, in many cases, not possible.

• Therefore, a different approach is often taken.

• For example, in many instances, two- or three-dimensional conduction problems may be

rapidly solved by utilizing existing solutions to the heat diffusion equation.
• These solutions are reported in terms of a shape factor S or a steady-state dimensionless
conduction heat rate, q*ss.
• The shape factor is defined such that

Where:
∆T1-2 is the temperature difference between boundaries.
• It also follows that a two-dimensional conduction resistance may be expressed as
• at locations removed from the object. For these infinite medium cases, useful results
may be obtained by defining a characteristic length

Where:
As:- is the surface area of the object. Conduction heat transfer rates from the
object to the infinite medium may then be reported in terms of a dimensionless
conduction heat rate

From Table 4.1, it is evident that the values of q*ss, which have been obtained analytically
and numerically, are similar for a wide range of geometrical configurations. As a
consequence of this similarity, values of q*ss may be estimated for configurations that are
similar to those for which q*ss is known.
 Example 1: Conduction Shape Factor
A long cylindrical pipe with a diameter of 10 cm is buried horizontally 1 m below the ground surface. The
pipe carries a hot fluid, maintaining its outer surface at 80°C, while the ground surface is at 10°C. The
thermal conductivity of the soil is k = 1.5 W/m*k. Calculate the heat loss per unit length of the pipe using
the conduction shape factor.
Step-by-Step Solution:
Identify the Geometry:
Buried horizontal pipe (cylinder) in a semi-infinite medium (soil).
Diameter D =10 cm =0.1 m
Depth z =1 m
Find the Shape Factor (S) for a Buried Horizontal Pipe:
The shape factor for this configuration is given by: S=
For unit length (L = 1 m), this simplifies to: S =

Substitute the given values we have the shape factor is:

S = 1.7m

The heat transfer is given by: -)

Q = 1.70×1.5×(80−10)=1.70×1.5×70
Substituting S=1.70 m, Tpipe=80°C and Tsurface we have =178.5 W/m
Example 2:
A cubical glass melting furnace has exterior dimensions of width W = m on a side and is constructed from
refractory brick of thickness L = 0.35 m and thermal conductivity k = 1.4 W/m*K. The sides and top of the
furnace are exposed to ambient air at 25°C, with free convection characterized by an average coefficient of h =
5W/m2*K. The bottom of the furnace rests on a framed platform for which much of the surface is exposed to the
ambient air, and a convection coefficient of h = W/m2*K may be assumed as a first approximation. Under
operating conditions for which combustion gases maintain the inner surfaces of the furnace at 1100°C, what is the
heat loss from the furnace?
Solution:
 Finite-Difference
Equations
The Nodal Network

• The first step in any numerical analysis must therefore be to select these points.
• Referring to the figure mentioned below, this may be done by subdividing the medium of interest
into a number of small regions and assigning to each a reference point that is at its center.
• The reference point is frequently termed a nodal point (or simply a node), and the aggregate of
points is termed a nodal network, grid, or mesh.
• The nodal points are designated by a numbering scheme that, for a two-dimensional system.
• The x and y locations are designated by the m and n indices, respectively.
• Each node represents a certain region, and its temperature is a measure of the average temperature
of the region.
• The numerical accuracy of the calculations depends strongly on the number of designated nodal
points.

• If this number is large (a fine-mesh), accurate solutions can be obtained.

Finite-Difference Form of the Heat Equation

• Determination of the temperature distribution

numerically dictates that an appropriate
conservation equation be written for each of the
nodal points of unknown temperature.
• The resulting set of equations may then be solved
simultaneously for the temperature at each node.
• If the system is characterized in terms of a nodal
network, it is necessary to work with an
approximate, or finite-difference form of this
equation.
• A finite-difference equation that is suitable for the
interior nodes of a two-dimensional system.
• Consider the second derivative, ­∂2T/­∂ x2. the value
of this derivative at the (m, n) nodal point may be
approximated as
• The temperature gradients may in turn be expressed as a function of the nodal temperatures. That is,

Substituting the above two equations in to the second derivative equations, we obtain

Proceeding in a similar fashion, it is readily shown that

Using a network for which ∆x = ∆y,

 The Energy Balance Method

• In many cases, it is desirable to develop the finite-difference equations by an alternative method

called the energy balance method.
• In the energy balance method, the finite-difference equation for a node is obtained by applying
conservation of energy to a control volume about the nodal region.
• Since the actual direction of heat flow (into or out of the node) is often unknown, it is convenient
to formulate the energy balance by assuming that all the heat flow is into the node.
• For steady-state conditions with generation,

• For two-dimensional conditions, energy exchange is influenced by conduction between (m, n) and
its four adjoining nodes, as well as by generation.

where i refers to the neighboring nodes, q(i) →(m, n) is the conduction rate between nodes, and unit depth
is assumed.
 Where:
- i refers to the neighboring nodes,
- q(i)→(m, n) is the conduction rate between nodes,

- and unit depth is assumed.

• Simplified forms of Fourier’s law may therefore be used.

• For example, the rate at which energy is transferred by conduction from node (m – 1, n) to
(m, n) may be expressed as

• The quantity (∆y * 1) is the heat transfer area, and the term (Tm–1,n - Tm,n)/∆x is the finite
difference approximation to the temperature gradient at the boundary between the two
nodes.
• The remaining conduction rates may be expressed as
• Note that, in evaluating each conduction rate, we have subtracted the temperature of the (m, n)
node from the temperature of its adjoining node.
• This convention is necessitated by the assumption of heat flow into (m, n).
• Substituting the above equations in to energy balance equation and we come up with
 Solving the Finite-Difference
Equations
• Once the nodal network has been established and an appropriate finite-difference
equation has been written for each node, the temperature distribution may be
determined.
• The problem reduces to one of solving a system of linear, algebraic equations.
Formulation as a Matrix Equation
• Consider a system of N finite-difference equations corresponding to N unknown
temperatures.
• Identifying the nodes by a single integer subscript, rather than by the double subscript (m,
n), the procedure for performing a matrix inversion begins by expressing the equations as
• where the quantities a11, a12, . . . , C1, . . . are known coefficients and constants involving
quantities such as ∆x, k, h, and T.
Using matrix notation, these equations may be expressed as

• Numerous mathematical methods are available for solving systems of linear, algebraic equations,
and many computational software programs have the built-in capability to solve the above matrix
for the solution vector [T].
Example 3
Steady-state temperatures (K) at three nodal points of a long rectangular rod are as
shown. The rod experiences a uniform volumetric generation rate of 5= 10e7 W/m3 and
has a thermal conductivity of 20 W/m*K. Two of its sides are maintained at a constant
temperature of 300 K, while the others are insulated.

(a) Determine the temperatures at nodes 1, 2, and 3.

(b) Calculate the heat transfer rate per unit length
(W/m) from the rod using the nodal temperatures.
Compare this result with the heat rate calculated
from knowledge of the volumetric generation rate
and the rod dimensions.
Chapter

End