Chapter 1

Heat Conduction in 3D Solids

In order to obtain the temperature distribution T(x,y,z,t) in a 3D body as a function of the point
P(x,y,z) and of time t, it is necessary to study the problem of heat conduction, one of the three
kinds of heat transfer. The remaining two, that is convection and radiation have only a limited,
although very significant, role in the determination of temperature distribution in the case solids
and structures are concerned. Heat conduction is based on two basic principles of classical
physics:

V ∂V

O y

x
Figure 1.1: The considered volume V and its boundary ∂V

•   heat is transferred from high temperature points to low temperature points of the solid,
depending upon a material properties called thermal conductivity;

•   the accumulation of heat in a material point of the body increases the temperature of that
point, depending upon another property of the material called specific heat. In order to set
the mathematical model for the heat conduction problem we assume that a 3D material
body is occupying a volume V with boundary ∂V in a 3D cartesian framework (O,x,y,z).

For each point P(x,y,z) we define the temperature T(x,y,z,t) as a continuous function of space and
time coordinates and, considering an isothermal surface of normal n passing through P, we define
the heat flux qn as the heat Q per unit surface and per unit time passing through the surface. The
temperature and the heat flux are usually measured, respectively, in K and in W/m2. The heat flux
has the physical dimension of

[q] = [ F ⋅ L ⋅ L−2 ⋅ T −1 ].

P(x,y,z) dS

O y

Figure 1.2: The isothermal surface element dS of normal n passing through the point P.

By considering the derivative of T along n, and remembering that the heat flux qn is directed from
high temperatures to low temperatures, that is in the direction opposite to the normal n, the first
principle of heat conduction can be written as
∂T
qn = −λn (1.1)
∂n
where λn is the thermal conductivity of the solid along n, measured in W/m·K. In an anisotropic
body the thermal conductivity is different in all directions. It is possible to demonstrate that once
the heat flux is known along three orthogonal directions, namely the coordinates x,y,z, one can
obtain the value of qn provided that direction of n is given. By assuming that λx, λy, λz are the
thermal conductivities measured along x,y and z, the three components of heat flux q can be
written as

∂T ∂T ∂T
qx = −λx ; q y = −λ y ; qz = −λz . (1.2)
∂x ∂y ∂z
We now consider the second principle based on which the governing equations of heat transfer
are obtained. At point P we consider the balance of thermal energy of an infinitesimal volume of
the solid. For sake of simplicity we consider a volume element dxdydz.
 z

(dQx )in

dz
P
dx

dy
(dQx )out
O y

Figure 1.3: The heat balance in an infinitesimal volume towards the point P

We consider the amount of heat entering the element through all its surfaces. Through the surface
dydz passing through P the portion of heat enter the volume in the x direction in the time interval
dt is
∂T
( dQx )in = qx dydzdt = −λx dydzdt. (1.3)
∂x
At the surface distant dx from the previous one, the value of qx will be varied due to the fact that
we move from the coordinate x to the coordinate x+dx, and the heat is going out of the material
portion
𝜕𝑞# 𝜕𝑇 𝜕 𝜕𝑇
𝑑𝑄# $%& = 𝑞# + 𝑑𝑥 𝑑𝑦𝑑𝑧𝑑𝑡 = − 𝜆# + 𝜆# 𝑑𝑥 𝑑𝑦𝑑𝑧𝑑𝑡. (1.4)
𝜕𝑥 𝜕𝑥 𝜕𝑥 𝜕𝑥
The amount of heat remaining in the volume due to the heat flow along the x direction is

∂ ⎛ ∂T ⎞
( dQx )in − ( dQx )out = ⎜ λx ⎟ dxdydzdt. (1.5)
∂x ⎝ ∂x ⎠

Similar expressions can be obtained for the flow in y and z directions

∂ ⎛ ∂T ⎞
( dQ ) − ( dQ )
y in y out = ⎜ λy ⎟ dxdydzdt ,
∂y ⎝ ∂y ⎠ (1.6)

and

∂ ⎛ ∂T ⎞
( dQz )in − ( dQz )out = ⎜ λz ⎟ dxdydzdt. (1.7)
∂z ⎝ ∂z ⎠
The amount of heat that remains in the volume dxdydz due to heat conduction through the surface
of the elemental volume in the time interval dt is
𝜕𝑞# 𝜕𝑞4 𝜕𝑞5
𝑑𝑄3 = − + + 𝑑𝑥𝑑𝑦𝑑𝑧𝑑𝑡 = −𝑑𝑖𝑣 𝐪 𝑑𝑥𝑑𝑦𝑑𝑧𝑑𝑡 =
𝜕𝑥 𝜕𝑦 𝜕𝑧
𝜕 𝜕𝑇 𝜕 𝜕𝑇 𝜕 𝜕𝑇 (1.8)
= 𝜆# + 𝜆4 + 𝜆5 𝑑𝑥𝑑𝑦𝑑𝑧𝑑𝑡.
𝜕𝑥 𝜕𝑥 𝜕𝑦 𝜕𝑦 𝜕𝑧 𝜕𝑧
In presence of an interval heat generation per unit volume per unit time Q, the amount of heat
generation in the volume dxdydz in the time interval dt is
𝑑𝑄9 = 𝑄𝑑𝑥𝑑𝑦𝑑𝑧𝑑𝑡. (1.9)
On the other side the heat dQ3=dQ1+dQ2 will produce an increase of the temperature dT that is
related to dQ3 as follows

dQ3 = cρ dTdxdydz, (1.10)

where c is the specific heat for unit mass, measured in J/kg·K , and ρ is the density, measured in
kg/m3. Since dQ3=dQ1+dQ2, by dividing both members by dxdydzdt we obtain

∂T ∂ ⎛ ∂T ⎞ ∂ ⎛ ∂T ⎞ ∂ ⎛ ∂T ⎞
cρ = ⎜ λx ⎟ + ⎜ λy ⎟ + ⎜ λz ⎟ + Q. (1.11)
∂t ∂x ⎝ ∂x ⎠ ∂y ⎝ ∂y ⎠ ∂z ⎝ ∂z ⎠

This equation, called Fourier equation or heat conduction equation, holds for a nonhomogeneous
anisotropic solid. It is partial differential equation of the second order in space variables and of
the first order in time. For an homogeneous body λx, λy and λz do not depend upon the point P(x,y,z)
and the equation reduces to

∂T ∂ 2T ∂ 2T ∂ 2T
cρ = λx 2 + λy 2 + λz 2 + Q. (1.12)
∂t ∂x ∂y ∂z
If we add the simplifying of isotropic body, for which

λx = λ y = λz = λ (1.13)
we obtain

∂T ⎛ ∂ 2T ∂ 2T ∂ 2T ⎞
cρ = λ⎜ 2 + 2 + 2 ⎟+Q (1.14)
∂t ⎝ ∂x ∂y ∂z ⎠

that can also be written as

∂T
cρ = λ∇ 2T + Q, (1.15)
∂t

∂ 2 (⋅) ∂ 2 (⋅) ∂ 2 (⋅)

where ∇2 (⋅) = + 2 + 2 is the laplacian operator. Introducing the so called thermal
∂x 2 ∂y ∂z
diffusivity
 λ
k= (1.16)
cρ
measured in m2/s, equation 1.15 can be recast in the form
1 𝜕𝑇 𝑄
= ∇9 𝑇 + (1.17)
𝑘 𝜕𝑡 𝜆
If no heat is generated inside the volume the equation reduces to
1 𝜕𝑇
= ∇9 𝑇 (1.18)
𝑘 𝜕𝑡
If no variation in time is observed (stationary case) we have
𝑄
∇9 𝑇 + = 0, (1.19)
𝜆
and if Q=0 the equation reduces to the Laplace equation:
∇9 𝑇 = 0. (1.20)

        Density   Specific  heat     Thermal  conductivity   Thermal  expansion  coeff.  

3
        𝜌  (kg/m )   cp  (J/kg  K)   𝜆  (W/m  K)   𝛼  (1E-­06/K)  
Aluminium   2700   900   247   23.6  
Metals  

Steel   7800   486   52   12  

Titanium   4500   460   7   8  
Gold   19320   138   315   14.2  
Magnesia  (MgO)   3600   940   38   13.5  
Ceramics  

Alumina  (Al2O3)   3690   775   39   7.6  

Glass   2500   795   0.78   9  
Silica  (SiO2)   2650    700   1.4   0.4  
Polypropylene   946   1925   0.12   145    180  
Polyethylene   920   1850   0.46  -­  0.50   106  -­  198  
Polymers  

Polystyrene   1000   1170   0.13   90  -­  150  

Teflon   2200   1050   0.25   126  -­  216  
Polystyrene  foam   46   1130   0.026    

Table 1.1: Thermal characteristics of different materials.

 
Boundary and Initial Conditions
The equation governing the temperature field in a homogeneous isotropic solid is given by eq
𝜕𝑇 𝜕9𝑇 𝜕9𝑇 𝜕9𝑇
𝑐𝜌 =𝜆 + + +𝑄 (1.21)
𝜕𝑡 𝜕𝑥 9 𝜕𝑦 9 𝜕𝑧 9

It is necessary to consider the boundary and initial conditions in order to obtain the temperature
field in a solid.

Boundary Conditions

We list some surface conditions to be prescribed for the heat transfer and for the temperature on
the boundary surface.
1)   Prescribed surface temperature (Dirichlet)
The surface temperature is prescribed as a function of position and time
𝑇 𝑥, 𝑦, 𝑧, 𝑡 = 𝑇 𝑥, 𝑦, 𝑧, 𝑡
(1.22)
for  𝐏 𝑥, 𝑦, 𝑧       ∈    𝜕𝑉  
where 𝑇 𝑥, 𝑦, 𝑧, 𝑡 is a prescribed temperature function of the position and the time.

2)   Adiabatic surface
When there is no heat flux through the boundary surface, the boundary conditions is
FG
expressed via eq. 𝑞 = −𝜆 :
FH

𝜕𝑇 𝑥, 𝑦, 𝑧, 𝑡
=0 (1.23)
𝜕𝑛
3)   Prescribed heat flux through the boundary surface (Neumann)
When the heat flux is prescribed on the boundary surface, the boundary condition follows
FG
eq. 𝑞 = −𝜆 :
FH

𝜕𝑇 𝑥, 𝑦, 𝑧, 𝑡
−𝜆 = 𝑞H 𝑥, 𝑦, 𝑧, 𝑡 (1.24)
𝜕𝑛
where 𝑞H 𝑥, 𝑦, 𝑧, 𝑡 denotes the heat flux prescribed on the boundary surface.

4)   Heat transfer by convection

In many case, the surface of the solid is in contact with fluid or gas. When heat transfer
between the boundary surface of the solid and the surrounding medium occurs by convection,
as in figure 1.4,
 Figure 1.4: Convection: temperature distribution

the heat 𝑄J transferred is:

𝑄J = ℎJ 𝑇 − 𝑇L 𝑑𝑆𝑑𝑡 (1.25)
R9 R3
where ℎJ denotes the heat transfer coefficient with dimensions [𝑊 ∙ 𝑚 ∙ 𝐾 ], ds is a
surface area element, and 𝑇∞ is the temperature of the surrounding medium which is a given
function of position and time. The amount of heat which flows through the surface of the
solid due to conduction is given by the Fourier equation
𝜕𝑇
−𝜆 𝑑𝑠𝑑𝑡 (1.26)
𝜕𝑛
If there is heat generation 𝑞W per unit area per unit time as the boundary surface, it may be
expressed by:
𝑞W 𝑑𝑠𝑑𝑡 (1.27)
The sum of the amount of heat due to heat conduction and the amount of heat due to surface
heat generation should be equal to the amount of the heat 𝑄J due to convection on the
boundary surface:
𝜕𝑇
−𝜆 𝑑𝑠𝑑𝑡 +   𝑞W 𝑑𝑠𝑑𝑡 =   ℎJ 𝑇 − 𝑇∞ 𝑑𝑠𝑑𝑡 (1.28)
𝜕𝑛
or after dividing by dsdt
𝜕𝑇
−𝜆 +   𝑞W =   ℎJ (𝑇 − 𝑇∞ ) (1.29)
𝜕𝑛
 Table 1.2: Some illustrative values of convective heat transfer coefficients

5)   Heat transfer by radiation

A body at absolute temperature T exchange heat by radiation. The amount of heat 𝑄Z
exchanged by radiation is:
𝑄Z =  𝜎𝜀 𝑇 ] − 𝑇∞ ] 𝑑𝑠𝑑𝑡 (1.30)
where 𝜎 is the Stefan-Boltzmann constant, 𝜀 is the emissivity of the surface, and 𝑇∞ is the
absolute temperature of the surrounding medium
𝜎 = 5,67 ∙ 10Ra   𝑊 (𝑚9 ∙ 𝐾 ] )
(1.31)

Thus the boundary condition due to heat transfer by radiation is given by:
𝜕𝑇
−𝜆 +   𝑞W =  𝜎𝜀 𝑇 ] − 𝑇∞ ] (1.32)
𝜕𝑛
This equation is not a linear function of temperature, so this condition offers some difficulty
in obtaining the temperature field in the solid from the Fourier heat conduction equation.
If the temperature difference between T and 𝑇∞ is not large it may be approximated to a
linear form:
𝑄Z =  𝜎𝜀 𝑇 ] − 𝑇∞ ] 𝑑𝑠𝑑𝑡 =
(1.33)
9 9
 = 𝜎𝜀 𝑇 − 𝑇∞ 𝑇 + 𝑇∞ 𝑇 + 𝑇∞ 𝑑𝑠𝑑𝑡 =
 ≅ 𝜎𝜀 𝑇 − 𝑇∞ 2𝑇∞ 2𝑇∞ 9 𝑑𝑠𝑑𝑡 =

= 4𝜎𝜀 𝑇∞ e 𝑇 − 𝑇∞ 𝑑𝑠𝑑𝑡

𝑄Z =   ℎZ 𝑇 − 𝑇∞ 𝑑𝑠𝑑𝑡
where:
ℎZ = 4𝜎𝜀 𝑇∞ e (1.34)
The boundary condition due to heat transfer by radiation is now:
𝜕𝑇
−𝜆 +   𝑞W =   ℎZ 𝑇 − 𝑇∞ (1.35)
𝜕𝑡
This equation has the same form as the equation of the convection.

6)   Heat transfer by convection and radiation

In many cases, hat transfer on the boundary occurs both by convection and radiation. The
boundary condition in this cases is given by:
𝜕𝑇
−𝜆 +   𝑞W = ℎJ 𝑇 − 𝑇∞ +  𝜎𝜀 𝑇 ] −   𝑇∞ ] (1.36)
𝜕𝑡
If the temperature difference between T and 𝑇∞ is not large, such equation can be
approximated by use of a compound coefficient ℎ, as follows:
𝜕𝑇
−𝜆 +   𝑞W = ℎ 𝑇 − 𝑇∞ (1.37)
𝜕𝑡
where:
ℎ =   ℎJ +   ℎZ (1.38)

7)   Contact between two solids

It very often happens that surface of a solid is in contact with the surface of another solid. If
the surface of two solids are in perfect thermal contact, the temperature on the contact surface
and the heat flow through the contact surface are the same for both solids:
𝑇3 = 𝑇9 (1.39)
and:
𝜕𝑇3 𝜕𝑇9
𝜆3 =   𝜆9   (1.40)
𝜕𝑛 𝜕𝑛
where subscripts 1 and 2, respectively, refer to solid 1 and solid 2 and n is the common
normal direction on the contact surface.

The seven boundary conditions discussed may be present simultaneously on different portions of
the boundary.
  

Initial condition
For solution of the transient heat conduction equation, it is necessary to specify an initial
condition which expresses the temperature distribution in the solid at initial time,
𝑇 𝑥, 𝑦, 𝑧, 𝑡f =   𝑇 𝑥, 𝑦, 𝑧 (1.41)
where 𝑇 𝑥, 𝑦, 𝑧 is the initial temperature distribution in the solid.
On the other hand, if the temperature has reached a steady state, it is not necessary to specify an
initial condition.