SHELL ENERGY BALANCES & BOUNDARY CONDITIONS

The methodology adopted in the heat transfer section is similar to that adopted for momentum transfer.
Like in the basic momentum transfer lessons, here heat-conduction problems are solved by an analogous
procedure.:(a) an energy balance is made over a thin slab or shell perpendicular to the direction of heat flow
and this balance leads to a I order Differential Equation from which the heat flux distribution is obtained; (b)
then into this expression for the heat flux Fourier’s law of heat conduction is inserted, and this gives a first
order differential equation for the temperature as a function of position. The integration constants that
incidentally appear here are determined by use of boundary conditions that specify the temperature or heat
flux at the boundary surfaces.
For kinds of heat sources can be investigated; electrical, nuclear, viscous, and chemical. Also can be
considered are heat flux through composite walls and heat transfer from(extended surfaces) fins. Heat
transfer in moving fluids ccan also be analyzed in the form of forced convection and free (natural)
convection.
Shell Energy Balances:
We select a slab or shell, the surfaces of which are normal to the direction of heat conduction and then
write for this system a statement of the law of conservation of energy. For steady state (or time independent)
conditions we write:
– + = 0 [1]
Thermal energy may enter or leave the system by the mechanism of heat conduction according to
Fourier’s law of heat conduction. Thermal energy may also enter or leave the system by virtue of the overall
fluid motion (This type of transfer is sometimes referred to as convective transport) and the energy entering
and leaving this way is commonly called “sensible heat” in and out. Thermal energy can be produced by the
degradation of electrical energy, by the slowing down of neutrons and nuclear fragments liberated in the
fission process, by the degradation of mechanical energy (viscous dissipation) and by conversion of
chemical energy into heat.
When equation [1] has been written for a system consisting of a thin slab or shell, the thickness of the
slab or shell is allowed to approach zero. This procedure leads us ultimately to a differential equation for the
temperature distribution. When this differential equation is integrated, constants of integration appear. These
are evaluated by the use of boundary conditions.
The following are the most common types of boundary conditions.
a] The temperature at a surface may be specified e.g.: T = T0
b] The heat flux at a surface may be given. E.g.: q = q0 (this is same as specifying the temperature gradient
normal to the surface)
c] At a solid-fluid interface the hear flux may be related to the difference between the temperature at the
interface and that in the fluid, thus:
q = h (T – Tfluid) ---- [2]
This relation is called the “Newton’s law of Cooling”. Actually it is not a law but a definition for ‘h’, which
is called a “heat transfer co-efficient”
d] At solid-solid interfaces the continuity of temperature and the normal component of the heat flux may be
specified.

HEAT CONDUCTION WITH AN ELECTRICAL HEAT SOURCE

Let us consider an electrical wire of circular cross section with radius ‘R’ and electrical conductivity
Ke Ohm – 1cm – 1. Through this wire we suppose there is an electrical current with current density I
amps/cm2. The transmission of an electric current is an irreversible process, and some electrical energy is
converted into heat (thermal energy). The rate of heat production per unit volume is given by the expression:
Se = --- [3]

Thus Se is the heat source owing to electrical dissipation. It is assumed here that the temperature rise in
the wire is not so large as to reckon the dependence of thermal or electrical conductivity. The surface of the
wire is maintained at a temperature T0. Let us now find out the temperature distribution within the heated
wire.
We selected as our system for energy balance, a cylindrical shell of thickness r and length ‘L’. (See
Figure below) The various contributions to the energy balance may be listed as follows:
 (2 rL) (q|r) ---- [4]

[2(r+ r) L) (q|r+r) ---- [5]

(2 rr L) Se --- [6]

The notation ‘qr’ means “flux of energy” in the r-direction. And |r means “evaluated at r”. We take “in”
and “out” to be in the +ve r-direction. On substituting these terms into equation [1] and dividing by 2rL
and taking limit as r goes to zero we have:

Limit r0 = Se r --- [7]

The left hand side of the above equation is the mathematical definition of the first derivative of ‘rq r’ with
respect to r, so that equation [7] becomes
(r qr) = Se r --- [8]
This isa First order Differential equation for the energy flux that can be integrated o obtain:
qr = + --- [9]
B.C. 1: at r = 0; qr is not infinite. Hence C1 has to be zero. --- [10]
The energy flux will take the form: qr = --- [11]
Which indicates that the heat flux increases linearly with r.
We shall now substitute Fourier’s law in the form: qr = – k --- [12]

Into equation [11] to obtain –k = --- [13]

If k is assumed to be constant, this first order differential equation may be integrated to give :
T= – + C2 --- [14]
To determine this integration constant C2 we use the B. C. 2 at r = r; T = T0 --- [15]
T0 = – + C2 i.e. C2 = T0 + --- [16]

T - T0 = --- [17]

Equation [17] gives the temperature rise as a parabolic function of the distance r from the axis of the wire.
From the temperature distribution expression [17] various other useful information can be obtained.
1] Maximum temperature rise (at r = 0)
Tmax – T0 = --- [18]
2] Average temperature rise:

<T> – T0 = = ---[19] ($ refer to box below)

The temperature rise averaged over the cross section is just one half the maximum rise; clearly <T> – T0 is
the same as <T – T0 > .
3] Heat flow at the surface (for length ‘L’ of the wire)
Q|r=R = 2  R L qr |r =R = 2  R L =  R2 L Se --- [21]
This vindicates one’s perception that at steady state all the heat produced all the heat produced by electrical
dissipation must leave through the wall.

$ T - T0 = ; <T> – T0 =

<T> – T0 = = = --- [20]

HEAT CONDUCTION WITH A VISCOUS HEAT SOURCE

Let us consider the flow of an incompressible Newtonian fluid between two co-axial cylinders. As the
outer cylinder rotates each cylindrical shell of fluid rubs against an adjacent shell of fluid. This rubbing
together of fluid produces heat; i.e. mechanical energy is steadily degraded into thermal energy. The volume
heat source resulting from this “viscous dissipation” may be designated by S v. Its magnitude depends on the
local velocity gradient; the more rapidly the two adjacent layers move with respect to one another, the
greater will be the viscous dissipation heating. The surfaces of the inner and outer cylinders at maintained at
T = T0 and T = Tb respectively. Clearly T will be a function of r alone.
 If the slit width b is small in comparison with the radius R of the outer cylinder, we can solve the
problem approximately by using the simplified system shown in Fig. 9.4-2 i.e. we ignore curvature effects
and solve the problem in Cartesian co-ordinates. For this modified problem, the viscous heat source is given
by:

Sv = – xz = --- [42]

(This equation may b accepted as such at this stage.) for the steady laminar flow of a fluid with constant
viscosity in a slit, we know the velocity profile will be linear represented by:
vz = --- [43]
so that the rate of viscous heat production per unit volume is:

Sv =  --- [44]
We can now introduce this quantity into an energy balance equation. Thermal energy balance over a thin
shell of thickness x, width W and length L gives for the steady state.

W L qx |x – W L qx |x+x + W L x  = 0 --- [45]

Division by W L x and letting x  0 gives:

= --- [46]
For constant viscosity  this equation may be integrated

qx =  x + C1 --- [47]
C1 cannot be determined now since nothing is known about heat flux at any value of x. Insertion of Fourier’s
law of heat conduction into equation [47] gives

–k = x + C1 --- [48]
In which k is the thermal conductivity of the fluid. Equation [48] may be integrated with respect to x to
obtain ( for constant k)
T= – + C1 + C2 --- [49]
C1 and C2 , the integration constants can be evaluated using the two boundary conditions:
B.C. 1: at x = 0, T = T0 --- [50]
B.C. 2: at x = b, T = Tb --- [51]
On solving we get:
= + ½ Br --- [52]

Here Br = is the “Brinkman number” which is a measure of the extent to which

viscous heating is important relative to the heat flow resulting from the impressed temperature difference (Tb
– T0). If Br >2, there is a maximum temperature at a position intermediate between the two walls.
In all flow problems viscous heating is not important. It is important in engineering work where large
velocity changes occur over short distances.
a] flow of a lubricant between fast moving parts
b] flow of plastics through dies in high speed extrusion
c] flow of air in the boundary layer near an earth satellite or a rocket.

Applying BC1 to Equation [49] we find that C2 = T0; and applying BC 2 we have
Tb = – + C1 + T0 ; thus C1 = (Tb – T0)+

C1 = (Tb – T0)+ we can now introduce the values of the integration

constants into the Equation [49] and rearrange to get:
= + ½ Br --- [52]

HEAT CONDUCTION WITH A NUCLEAR HEAT SOURCE

 Let us consider a nuclear fuel element of spherical form as shown in the figure above. It consists of a
sphere of fissionable material with Radius RF, surrounded by a spherical shell of alluminium cladding with
outer radius RC.
Inside the fuel element fission fragments are produced which have very high kinetic energies. Collissions
between these fragments and atoms of the fissionable material provide the major source of thermal energy in
the reactor. Such a volume source of thermal energy resulting from nuclear fission we call ‘Sn’(cal cm –3 sec
–1
) This source will not be uniform throughout the sphere of fissionable material; it will be smallest at the
centre of the sphere. We assume, for the sake of this problem, that the source can be approximated by a
simple parabolic function.
Sn = Sn0 --- [22]

Here Sn0 is the volume rate of heat production at the centre of the sphere, and b is a dimensionless constant
between 0 and 1.
Let us begin by making a thermal energy balance over a spherical shell of thickness r within the sphere
of fissionable material.
thermal energy in at r: qrF |r 4  r2 --- [23]
thermal energy out at r+r : qr |r+r 4  r2 --- [24]
F

thermal energy produced : Sn 4  r2 r --- [25]

substitution of these terms into equation [1] for the thermal energy balance, divide by 4  r and taking
limits as r  0.

Limit r  0 = Sn r2 --- [26] whence

= Sn0 r2 --- [27]

The differential equation for the heat flux qrC in the cladding is of the same form as equation [27] except to
that there is no source term.
= 0 --- [28]
On integrating these two equations [27] and [28] we have:

qrF = Sn0 + --- [29]

qrC = ---[30]

in which and are the integration constants to be evaluated by means of the boundary conditions:
B.C. 1: at r = 0; qrF is not infinite --- [31]
B.C. 2: at r = RF; qrF = qrC ---[32]
On applying these boundary conditions we get expressions for qrF and qrC as:
qrF = Sn0 --- [33]

qrC = Sn0 ---[34]

These are the heat flux distributions in the fissionable sphere and in the spherical shell cladding. We can
now substitute into these equations the Fourier’s law of heat conduction:
– kF = Sn0 --- [35]

– kC = Sn0 --- [36]

These equations may be integrated for constant kF and kC to give:
TF = – + --- [37]

TC = + --- [38]

The integration constants and are determined using the following boundary conditions:
B.C. 3: at r = RF TF = TC --- [39]
C C
B.C. 4 : at r = R T = T0 --- [40]
where T0 is the known temperature at the outside of the cladding. The temperature profiles can be obtained
as:

TF – T0 =

+ --- [41]

TC – T0 = --- [42]
The maximum temperature in the sphere of fissionable material can be obtained by setting r = 0 in equation
[41]

– T0 = + --- [43]

HEAT CONDUCTION WITH CHEMICAL HEAT SOURCE

A chemical reaction is being carried out in a fixed bed flow reactor as shown in the figure below.

The reactor extends from z = –  to z = +  and is divided into three zones. Reaction zone (0<z<L) is
packed with catalyst pellets and the entrance and exit zones are packed with pellets physically similar to the
catalyst pellets but non-catyalytic. The radial velocity gradients re neglected. (Plug flow is assumed)
The reactor wall is well insulated so that the temperature can be considered essentially independent of ‘r ‘.
We wish to find out the steady state axial temperature distribution T(z) when the fluid enters at z = –  with
a uniform temperature T1 and a superficial velocity v1 =
Such problems can be solved by assuming that the axial heat conduction follows Fourier’s law, with an
‘effective thermal conductivity’ for the packed bed. The axial and radial effective conductivities k zeff and
kreff generally have quite different values.
The volume rate of thermal energy production by chemical reactions Sc, is in general a complicated
function of pressure, temperature, composition and catalyst activity. For simplicity we shall consider Sc as a
function of temperature alone and assume a linear temperature dependence.
Sc = Sc1 --- [53]
 Here T is the local temperature in the catalytic bed (assumed equal for catalyst and fluid) and Sc1 and T0
are empirical constants for the given reactor inlet conditions. For the shell balance let us select a disk of
radius R and thickness z in the catalyst zone and the z is so chosen as to be much larger than the catalyst
particle dimensions. At steady state the different terms of the thermal energy balance can be written as
follows
R2 qz |z --- [54]

R2 qz |z+z --- [55]

R2 1 v1 Cp (T – T0 )|z --- [56]

R2 1 v1 Cp (T – T0 )|z+z --- [57]

( R2 z ) Sc --- [58]

Cp is assumed constant for the reaction mixture and the mass flow rate R2 1 v1 has been expressed in
terms of the entrance conditions. Energy (thermal) is expressed relative to a datum temperature T0.(not
possible to be expressed absolutely) The thermal energy balance can now be written as follows:
+ 1 v1 Cp = Sc --- [59]

This we have obtained by dividing the entire equation by R2 z. We now take limits as z  0. (not
continuum- still considering= knowing that resulting equation describes, not point values – but average
values of qz , T and Sc) This limiting operation yields:
+ 1 v1 Cp = Sc --- [60]
On inserting Fourier’s law, we have (same argument as per the above Differential Equation)
– kzeff + 1 v1 Cp = Sc --- [61]
Equation [61] is valid for zones I and III if Sc equal to zero. The resulting D.E.s for the temperature
distribution in the three zones are:
Zone – I (z<0) – kzeff + 1 v1 Cp = 0 --- [62]

Zone – II (0<z<L) – kzeff + 1 v1 Cp = Sc1 --- [63]

Zone – III (z>L) – kzeff + 1 v1 Cp = 0 --- [64]

We have used same value of kzeff for all three zones. These three second order differential equations are to be
solved with the following six boundary conditions:
at z = – ; TI = T1 --- [65]
at z = 0 ; TI = TII --- [66]
at z = 0 ; kzeff = kzeff --- [67]
at z = L ; TII = TIII --- [68]
at z = L ; kzeff = kzeff --- [69]
at z =  ; TIII is finite --- [70]
Equations [66] to [69] express the continuity of temperature and heat flux between zones and [70] is from
physical grounds. The following dimensionless quantities may be introduced.
Z = z / L -- [71];  = -- [72]; B = --[73] & N = --[74]

The above differential equations [62], [63] and [64] will thus become:

– + = 0 --- [75]
 – + = N  --- [76]

– + = 0 --- [77]
The solutions of these three linear homogenous Differential Equations are:
I = C1 + C2 --- [78]
II = C3 + C4 --- [79]
III = C5 + C6 --- [80]
in which m3 = B --- [81]
m4 = B --- [82]
boundary condition given by equation [70] renders C6 =0; hence the temperature in zone III is constant.
Application of Boundary condition given by [65] with [78] renders C1 to be unity. The rest of the boundary
conditions as given by equations [66] to [69] give the following temperature profiles in the three zones.
{when [1 –(4N/B)] >0}

I = 1 + --- [83]

II = (m3 + m4) ----[84]

III = --- [85]

HEAT CONDUCTION THROUGH COMPOSITE WALL

In industries conduction through walls made up of layers of various materials, each with its own
characteristic thermal conductivity is very common. We can see here how the various resistances to heat
transfer combine into a total resistance.

The above figure depicts a plane composite wall comprising of three different materials with thicknesses and
thermal conductivities as follows:
Thicknesses: x1 – x 0 ; x2 – x1 and x3 – x2
Thermal conductivities: k 01 k12 k23
At x = xo substance ‘01’ is in contact with a fluid with ambient temperature T0 , and at x = x3 substance ‘23’
is in contact with a fluid at temperature Tb. The heat transfer at the boundaries x = x o and x = x3 is given by
“Newton’s law of cooling” with heat transfer coefficients h0 and h3 respectively. The temperature profile is
sketched as above.
Differential equation for heat conduction in region ‘01’. A thermal balance on the slab of volume WH x
gives:
qx01|x WH – qx01|x+x WH = 0 --- [86]
After division by WH x and taking the limit as x  0, we get: = 0 ---[87]
Integration of this equation gives: qx01 = q0 ( a constant) --- [88]
The constant q0 is the heat flux at the place x = x0. On physical grounds, we know that at steady state the
heat flux in all three regions will be the same. Hence:
qx01 = qx12 = qx23 = q0 --- [89]
We also know that: qx01 = – k01 --- [90]
with similar relations for qx12 and qx23. Combination of these relations with equation [89] then gives:
– k01 = q0 --- [91]; – k12 = q0 --- [92] and – k23 = q0 --- [93]
Integration of these equations then gives for constant k01 ; k12 ; k23
T0 – T1 = – q0 --- [94]

T1 – T2 = – q0 --- [95]
 T2 – T3 = – q0 --- [96]
In addition we have the two statements regarding the heat transfer at the surfaces:
Ta – T1 = --- [97] and T3 – Tb = --- [98]

Addition of all the five equations [94] to [98] gives:

Ta – Tb = q0 --- [99]

or q0 = = --- [100]

Sometimes this result is rewritten in a form reminiscent of Newton’s law of cooling:

q0 = U (Ta – Tb) or Q0 = U (WH) (Ta – Tb) --- [101]
and the quantity U (called overall heat transfer coefficient) can be expressed as:

–1
U= --- [102]

HEAT CONDUCTION THROUGH COMPOSITE CYLINDRICAL WALLS

Let us perform a development similar to that we have adopted for a composite cylindrical wall such
as that shown in the figure below:
Solution: A thermal energy balance on a shell of volume 2rLr for region ‘01’ is:
qr01 |r 2  r L – qr01 |r+r 2  (r+r) L =0 --- [103]
Division by 2  L r and taking limit as r – 0 gives:
( r qr01 ) = 0 --- [104]
Integration of this equation for constant k01
r qr01 = r0 q ( a constant) --- [105]
in which r0 is the inner radius of region’o1’ and q0 is the heat flux there. For steady state we write, analogous
to Equations [91] to [93];
– k01 r = r0 q0 --- [106]

– k12 r = r0 q0 --- [107]

– k23 r = r0 q0 --- [108]

Integration of these equations gives for constant k01 . k12 and k23
T0 – T1 = r0 q0 --- [109]

T1 – T2 = r0 q0 --- [110]

T2 – T3 = r0 q0 --- [111]
At the two fluid-solid interfaces we have the requirements that
Ta – T0 = --- [112]

T3 – Tb = = --- [113]

Addition of all the five equations [109] to [113] gives an expression for Ta – Tb , which may be solved for q0
to give
Q0 = 2 L r0 q0 = --- [114]

We now define an “over-all heat transfer coefficient based on the inner surface” U0 :
Q0 = U0 (2  r0 L)(Ta – Tb) --- [115]
Combining of Equations [114] and [115] gives:
U0 = ( r0 ) – 1 –1 [115]

The subscript ‘0’ on U0 indicates that the over-all heat transfer coefficient is referred to he radius r0. This
result may be generalized to include a cylindrical tube made of ‘n’ laminae by replacing ‘3’ by ‘n’ in three
places in Equation [115]
HEAT CONDUCTION IN A COOLING FIN
We shall now consider another simple, but practical, application of heat conduction is in the
calculation of the efficiency of a cooling fin. Such fins are used to increase the area available for heat
transfer between metal walls and poorly conducting fluids such as gases. A simple rectangular fin is
sketched in the figure below.

A reasonably good description of the system can be obtained by approximating the true physical
situation by a simplified model:
True Situation Model
1] is a function of both x and z, but the 1] t is a function of z alone
dependence on z is more important
2] A small quantity of heat is lost from the 2] No heat is lost from the end or from the
fin at the end(area 2BW) and the edges edges.
(area 2BL + 2BL)
3] The heat transfer coefficient is a function 3] The heat flux at the surfaces is given by q =
of position h(T – Ta), in which h is constant and T = T(z)

A thermal energy balance on a segment z of the bar gives:

qz |z .2B W – qz|z+z .2B W – h(2 w z ) (T – Ta) = 0 --- [116]
Division by 2BWz and taking the limit as z  0 gives:
– = (T – Ta) --- [117]
Insertion of Fourier’s law (qz = – k dT/dz) in which k is the thermal conductivity of the metal, gives for
constant k:
= (T – Ta) --- [118]
This equation is to be solved with the boundary conditions:
B.C. 1: at z = 0, T = Tw --- [119] & B.C. 2: at z = L, = 0 --- [120]
We shall introduce the following dimensionless quantites:
= = dimensionless temperature --- [121]

= = dimensionless distance --- [122]

N= = dimensionless heat transfer coefficient --- [123]

This problem will thus get restated as:

= N2  with | = 0 =1 and | = 1 = 0 --- [124,125 & 126]

Equation [124] may be integrated to give hyperbolic functions. When the two integration constants have
been determined we get”
 = Cosh N – (tanh N) Sinh N ----- [ 127]
This may be rearranged to give:
= --- [128]

It should be emphasized that this expression is reasonable only if the heat lost at the edges is negligible.
The Effectiveness of a Fin is defined by:
= --- [129]

The theoretical value of ‘’ for the problem considered here is:

 = --- [130]

Or = = = ---[131]

In which N is the dimensionless quantity defined in equation [123]