Chapter 2 Conduction Analysis

Download as doc, pdf, or txt
Download as doc, pdf, or txt
You are on page 1of 16

MEC551 WANWM

CHAPTER 2

CONDUCTION ANALYSIS
2.1 Conductivity analysis : Plane walls
For a simple case of steady-state one-dimensional heat flow through a plane wall, the
temperature gradient and the heat flow do not vary with time and the cross-sectional area
along the heat flow path is uniform. The equation for rate of heat transfer in equation 1.1
(Fouriers Law) can be integrated and yields

Ak
T
(Thot Tcold )
L
L / Ak

2.1

In this equation, T is the difference between the higher temperature T hot and the lower
temperature Tcold, which is the driving potential that causes the flow of heat. The quantity L/Ak
is equivalent to a thermal resistance R that the wall offers to the flow of heat by conduction:

L
Ak

2.2

The resistance to the heat flow is directly proportional to the material thickness, and inversely
proportional to the material conductivity and area normal to the area of heat transfer.
A heat flow system can be analyzed as an electrical circuit. The analogy between conductive
heat transfer and electrical current flow is illustrated in Fig 6.

Figure 1. Analogy between conduction and electrical current flow


2.2 Conductivity analysis: Radial Systems with Fixed Surface Temperature
Fig 2 shows a single-layer cylindrical wall of a homogenous material with constant thermal
conductivity and uniform inner and outer surface temperatures. The area normal to radial heat
flow in a hollow cylinder by conduction is 2rL, where L is the cylinder length.

W.A.N.

Page

April 2005

MEC551 WANWM

Figure 2. Hollow cylinder with surface conditions


For a hollow cylinder, the heat transfer rate is

2kL (Ts ,1 Ts , 2 ) T

r2
R
ln
r1

2.3

where the thermal resistance for the hollow cylinder, R is

r2
r1
R
2kL
ln

2.4

For a sphere, the area at a given radius is 4r 2. Therefore, the radial conductive heat transfer
is given by the equation

4k (T1 T2 ) T

1 1
R

r1 r2

2.5

where the thermal resistance, R is

1 1
)
r1 r2
R
4k
(

2.6

Example 2.1
A steel pipe of 100 mm bore and 7 mm wall thickness, carrying steam at 260 oC, is insulated
with 40 mm of a molded high-temperature diatomaceous earth covering. This covering is in
turn insulated with 60 mm of asbestos felt. If the atmospheric temperature is 15 oC, calculate
the rate at which heat is lost by steam per m length of pipe.
The heat transfer coefficients for the inside and outside surfaces are 550 and 15 W/m 2 K,
respectively, and the thermal conductivity of steel, and diatomaceous earth, and asbestos felt

W.A.N.

Page

April 2005

MEC551 WANWM

are 50, 0.09, and 0.07 W/m K respectively. Calculate also the temperature of the outside
surface. Consider 1 m length of the pipe.

Solution

T Tst Tair

R
RT

where the total resistance of the heat flow,


RT = Rst + Rsp + Rrb + Rasb + Rair
For convective heat transfer of the steam film near the inner wall of the steel pipe,

R st

1
hst A

1
1

0.00579 K / W
hst ( 2 ri H ) 550 x 2x0.05 x1

For conduction heat transfer of the steel pipe section,

Rsp

r1

ri

ln

57

50

4.17 x10 4 K / W
2 (1)(50)
ln

2H k sp

For rubber covering (conduction),

r2
r1

97

57

Rrb

0.94 K / W
2H k r 2 (1)(0.09)
ln

ln

For asbestos insulation (conduction),

Rasb

ro
r2

ln

2H k asb

157

97 1.095 K / W

2 (1)(0.07)
ln

For air film on the outside surface (convection),


W.A.N.

Page

April 2005

MEC551 WANWM

Rair

hair A

1
1

0.0675 K / W
hair ( 2 ro H ) 15 x 2x 0.0157 x1

Therefore, the total resistance for the heat transfer,


RT = 2.1087 K/W
And the heat loss per meter length of piping,

Tst Tair 260 15

116 W
RT
2.1087

To calculate the temperature of the outside surface of the piping, the heat transfer analysis is
made for the layer between the outside surface and the air film adjacent to the surface. The
quantity of the heat transferred in this layer is equal to the heat transfer quantity from the
steam to the surrounding air.

To Tair
Rair

116 W

To 15 o C
0.0675 K / W

To 22.8 o C
2.3 Critical Thickness of Insulation
The critical thickness of
insulation is used to
determine whether
insulation is required in a
system.

h, T

ri
Ti

ro

There are cases where added insulation will increase (not decrease!) the rate of heat transfer
from the inner part of a pipe to the surrounding, meaning that it increases the loss of energy
from a heat conveying fluid.
For an insulation material, calculate the critical radius value, rc

k
and then compare with
h

the outer radius of the pipe containing the heat conveying fluid.
If rc > ro, pipe
then the heat transfer from the fluid to the surrounding will increase if
insulation is added.
If rc < ro, pipe

then the heat transfer will decrease for any addition of insulation thickness.

2.4

Differential Equation for heat conduction in Cartesian coordinates

2.4.1

Plane wall

W.A.N.

dQ(z+dz)
Page

dQ(x+dx)
April 2005

MEC551 WANWM

dQ(y+dy)
dQy
y

dQx
x

Differential control volume,

1) An element of a
material.
2) Material assumed :
- homogeneous across
the grain of the solid
- conductivity constant

dQz

dV = dx.dy.dz

Heat flux in x-direction, dQx k .dy.dz

T
x

The rate of heat flux out of the element in the x-direction,

dQ( x dx ) k .dy.dz


T
T
2T

dx

k
.
dy
.
dz

k
.
dx
.
dy
.
dz

x
x
x
x 2

Therefore, the net rate of heat flow into the element in the x-direction
dQnet,x = dQx dQ(x+dx) k .dx.dy.dz
in the y-direction
dQnet,y = dQy dQ(y+dy) k .dx.dy.dz

2T
x 2
2T
y 2

in the z-direction
dQnet,z = dQz dQ(z+dz) k .dx.dy.dz

2T
z 2

So, the total rate of heat flow on the element


dQ = dQnet,x + dQnet,y + dQnet,z

2T 2 T 2T

2
y 2
z 2
x

k .dx.dy.dz

If there is heat generation and heat storage


Rate of heat
storage

Net rate of heat flow


in into the element

1 T 2T 2T 2T
q'
.

t x 2 y 2 z 2
k
where

k
.C p

Rate of heat
generation

2.7

thermal diffusivity

Special cases of heat conduction equation

1 T 2T 2T 2T
.

t x 2 y 2 z 2

(i) Fourier equation (no heat generation, q = 0)

W.A.N.

Page

April 2005

MEC551 WANWM

(ii) Poisson equation (steady state HT,

2T 2T 2T q '
2 2
2
k
y
z
x

T
0)
t

2T 2T 2T
2 2
2
y
z
x

(iii) Laplace equation (steady state with no heat generation)


Boundary conditions
i) constant surface temperature,

T(0,t) = Ts

ii) constant & finite heat flux,

iii) adiabatic or insulated surface,

T
x

iv) convection surface condition,

2.4.2

T
x

x0

x 0

T
x

qs

x 0

h T f T (0, t )

Cylindrical coordinate system

dz

dr

Rate of heat
storage

Net rate of heat flow


in into the element

1 T 1 T
1 2T 2 T q '
.

t
r r r
k
r 2 2 z 2

Rate of heat
generation

2.8

2.5 Heat generation Systems


Heat generation systems include electrical heating, nuclear reactors and chemically reacting
systems. In this section, heat transfer for one-dimensional cases with constant and uniform
heat generation is analyzed.
2.5.1 Plane walls

W.A.N.

Page

April 2005

MEC551 WANWM

Consider the plane wall of Figure 3 in which there is uniform energy generation per unit

volume ( q is constant) and the surfaces are maintained at T s,1 and Ts,2. By integrating the
Poisson equation for conductive heat transfer, the temperature distribution is

qL2
T ( x)
2k

x2
L2

Ts , 2 Ts ,1
2

Ts , 2 Ts ,1
x

L
2

2.9

The heat flux at any point in the wall can be determined by using equation 2.9

Figure 3. Conduction in a plane wall with uniform heat generation (a)


Asymmetrical boundary conditions. (b) Symmetrical boundary conditions. (c)
Adiabatic surface at midplane.
For a simpler case where Ts,1 = Ts,2 = Ts (figure 3-b), equation 2.9 reduces to

qL2
T ( x)
2k

x2
L2

Ts

2.10

The maximum temperature exists at the mid-plane of the wall,

qL2
T (0) T0
Ts
2k

2.11

If the temperature of the adjoining fluid, T , is known, then at the boundary surface, x = L, the
temperature can be determined as

qL
Ts T
h

2.12

where h is the heat transfer coefficient of the fluid.

W.A.N.

Page

April 2005

MEC551 WANWM

Example 2.2
A plane wall is a composite of two materials, A and B. The wall of material A has heat

generation, q = 1.5 x 106 W/m3, kA = 75 W/m.K and thickness L A = 50mm. The wall material
B has no generation, with kB = 150 W/m.K and thickness LB = 20mm. The noncontact surface
of material A is well insulated, while the noncontact surface of material B is cooled by a water
stream with T = 30oC and h = 1000 W/m2.K.
Sketch the temperature distribution which exists in the composite under steady-state
conditions, and determine the temperature To of the cooled surface.
1.
2.

Sketch the temperature distribution which exists in the composite under steady-state
conditions.
Determine the temperature T0 of the insulated surface and the temperature T2 of the
cooled surface.

ASSUMPTION:
1.
2.
3.
4.
5.

Steady-state conditions.
One dimensional conduction in x direction.
Negligible contact resistance between walls.
Inner surface of A is adiabatic.
Constant properties for materials A and B.

ANALYSIS:
From the prescribed physically conditions, the temperature distribution in the composites is
known to have the following features:
a)Parabolic in material A.
b)Zero slope at insulated boundary.
c)Linear in material B
d)Slope change = kB/kA = 2 at interface.
The temperature distribution in the water is characterized by
e)Large gradients near the surface.
These features are shown below.

W.A.N.

Page

April 2005

MEC551 WANWM

The outer surface temperature, T2, may be obtained by performing an energy balance on a
control volume about material B. Since there is no generation in this material, it follows that,
for steady-state conditions and a unit surface area, the heat flux into the material at x = LA
must equal the heat flux from the material due to convection at x = LA + LB. Hence
q = h (T2- T)

(1)

The heat flux q may be determined by performing a second energy balance on a control
volume about material A. In particular since the surface at x = 0 is adiabatic, there is no inflow
and the rate at which energy is generated must equal the outflow. Accordingly, for a unit
surface area,

q LA = q

(2)

Combining Equations 1 and 2, the outer surface temperature is

T2

= T + q L A

T2
T2

6
3
= 30 C + 1.5 10 W / m x 0.05m
1000W / m 2 k
o

= 105

The temperature at the insulated surface is

2
T0 = q L A + T1
2k A

(3)
where T1 may be obtained from the following thermal circuit.

That is T1 = T + (Rcond,B + Rconv) q


The resistances for a unit surface area are

W.A.N.

Page

April 2005

MEC551 WANWM

Rcond,B =

LB
kB

; Rconv =

1
h

Hence

C + 1500W.02/mmK 1000W1 / m K 1.5 10

T1 = 30 C + 85 C = 115 C
T1 = 30

W / m3 0.05 m

Substituting into Equation 3,


2

1.5 10 6W / m3 (0.05m)
115 C
2 75W / mK
T0 25 C 115 C
T0

T0 140 C
COMMENTS:
1. Recognize that material A, having heat generation, cannot be represented by a
thermal circuit element.
2. Note that [(T2 - T)/(T1 T2)] = (Rconv/Rcond,B) = 7.5.
2.5.2 Radial Systems
Heat generation may occur in a variety of radial geometries. Consider the long, solid cylinder
of figure 4, which could represent a current-carrying wire or a fuel element in a nuclear
reactor. For steady-state conditions the rate at which heat is convected from the surface
temperature to be maintained at a fixed value of Ts.

Figure 4. Conduction in a solid cylinder with uniform heat generation


The temperature distribution is

q ro2
T (r )
4k

1 r T
s
2

r0

2.13

Evaluating Equation 2.13 at the centerline and dividing the result into equation 2.13, we
obtain the temperature distribution in nondimensional form

r2
T (r ) Ts
1 2

T0 Ts
r0

W.A.N.

Page

10

April 2005

MEC551 WANWM

where T0 is the centerline temperature. The heat rate at any radius in the cylinder may, of
course, be evaluate by using equation 2.13 with Fouriers law. To relate the surface
temperature, Ts, to the temperature T,of the moving fluid, either a surface energy balance or
an overall energy balance may be used.

qr
Ts T 0
2h

2.14

2.6 NUMERICAL ANALYSIS ON MULTIDIMENSIONAL STEADY-STATE CONDUCTION


2.6.1 Introduction
The steady-state temperature in a three-dimensional Cartesian coordinate obeys the Laplace
equation,

2 T 2T 2 T

0
x 2
y 2
z 2
this equation is true when the thermal conductivity, k, is constant.
The solution of this equation, T(x,y,z), can be differentiated and combined with Fouriers
equation to yield the components of the vector heat-transfer rate. These components are

q x kAx

T
x

q y kAy

T
y

q z kAz

T
z

where Ax is the area normal to qx, etc.


The Laplace equation can be solved using a number of different methods, one of it is the
numerical analysis.
2.6.2 Numerical Analysis
Figure 5 shows a general two-dimensional body with uniform thickness, L in the z-direction
and no temperature gradient in that direction.
y

4
x
1

2
Figure 5
An appropriate value of x and y is chosen, and the body is divided into a network of
rectangles. Each rectangle contains a single nodal point at its center.
Heat transfer is considered to occur between nodal points only, connected by fictitious rods
acting as conductors or resistors for the heat flow. Thermal energy is considered to be stored
at the nodal points only.

W.A.N.

Page

11

April 2005

MEC551 WANWM

The horizontal conductance,

kAh
kL(y )

x
x

The vertical conductance,

kAv
kL(x )

y
y

2.15

2.16

A steady-state energy balance on an interior nodal point, n, is

q1n q 2n q3n q 4n 0

2.17

where q is positive for heat flow into the nodal, n.


Using the product of conductance and a finite temperature difference for each conductive flux,
(2.17) becomes

T Tn
T Tn
kL(y ) 1
kL(x) 2
y
x

T Tn
T Tn
kL(y ) 3
kL(x) 4
x
y

When x y ,
T1 + T2 + T3 + T4 4Tn = 0

2.18

Equation 2.18 can be written for each interior nodal point of a body. The temperatures at the
nodal points can be solved by simultaneous solution of the sets of equations.
Example 2.3
Write the set of nodal temperature equations for a six-inch square grid for the square chimney
shown. Assume the material to have uniform thermal conductivity, uniform inside temperature,
Ti=300oC, and uniform outside temperature To = 100oC.
To=100oC
c

a
b
c

Ti

Ti=300oC

6
12
18

Solution
In the indicated quadrant of the chimney, the only unknown nodal temperatures are T a, Tb,
and Tc, since clearly Tb = Tb, and Tc = Tc.
The nodal equations from (2.18) are

W.A.N.

Node a :
Or

Tb + Tb + 100 + 100 4Ta = 0


Tb + 100 2Ta = 0

Node b :
Or

300 + Tc + 100 + Ta 4Tb = 0


400 + Tc + Ta 4Tb = 0

Page

12

April 2005

MEC551 WANWM

Node c :
Or

300 + Tb + 100 + Tb 4Tc = 0


200 + Tb 2Tc = 0

The set of equations can be rearranged in the form of


2Ta Tb + 0 = 100
-Ta + 4Tb Tc = 400
- Tb + 2Tc = 200
and the problem can be solved using linear algebraic equations.
2.6.3 Relaxation technique in solving linear algebraic equations
Linear algebraic equations can easily be solved by a computer using either the matrix
inversion technique or the Gaussian elimination technique. However, both computer-aided
techniques will not be discussed in this notes.
The relaxation technique is a method to solve a set of linear algebraic equations without using
computer. It is used for nonrepetitive-type problems, involving a relatively small number of
nodal points (not more than 10 nodals) in a conduction problem.
Example 2.4
Determine the steady-state temperatures at the 4 interior nodal points of the figure below.
T = 300oC

T = 400oC

T = 200oC
1

The nodal equations are,


Node 1 : 400 + 500 + T2 + T4 4T1 = 0
Node 2 : 500 + 200 + T1 + T3 4T2 = 0
Node 3 : 200 + 300 + T2 + T4 4T3 = 0
Node 4 : 300 + 400 + T1 + T3 4T4 = 0

(1)
(2)
(3)
(4)

T = 500oC

There are 4 unknown nodal temperatures. The relaxation method of solution is as follows:
1. Assume values for the 4 unknown temperatures. Good initial guesses helps to
minimize the calculation work.
2. Since the initial guesses will usually be in error, the right side of each nodal equation
will differ from zero. A residual (R) will exist due to inaccuracies in the assumed
values. Replace the zeros in equations (1) to (4) with R1, R2, R3 and R4 :

900 + T2 + T4 4T1 = R1
700 + T1 + T3 4T2 = R2
500 + T2 + T4 4T3 = R3
700 + T1 + T3 4T4 = R4

(5)
(6)
(7)
(8)

3. Set up a unit change table as Table 2.1.The table shows the effect of a one-degree
change of temperature at one node upon the residuals.

W.A.N.

Page

13

April 2005

MEC551 WANWM

Table 2.1
R1

R 2

R3

R4

T1 1

-4

+1

+1

T2 1

+1

-4

+1

0
+1

+1
0

-4
+1

+1

-2

-2

-2

T3 1
T4 1
Block Change =
+1

-4
-2

4. Calculate the initial residuals for the initially assumed temperatures using the
residual equations (5) through (8).
5. Set up a relaxation table such as Table 2.2. Begin with the initially assumed
temperatures and the resulting initial residues.
The left-hand column records the changes from the initially assumed temperature
values. Notice that the procedure begins by relaxing the largest initial residual.
Table 2.2
R2

T1

R1

T2

T3

400
400

-25
-5

325
345

+75
-5

275
275

400

-5

345

+20

400

350

R3

T4

R4

+75
+95

350
350

-25
-25

300

-5

350

300

350

Initial Values

T2 20
T3 25
T2 5
Check by
equations
Solution

0
400

0
350

0
300

0
350

In this problem, begin by reducing R 2 or R3. Choose R2 first because it is larger and proceed
by relaxing slightly ( T2 20 ). Table 2.2 is convenient because it allows rapid calculation
of the changes in the residuals.
Notice that the +20 change in T2 reduced the residuals at nodes 1 and 2 but increases R3.
The new residuals and temperatures are then noted in the table, where the only temperature
change is bolded. Proceed to relax the largest resulting residual, which is now R 3.
Following a temperature change of +25 o at node 3, we see that R 4 = 0. But this does not
necessarily mean that we have obtained the correct temperature at node 4.
Proceed, the largest residual now is R2 again, which is reduced to 0 by a +5o change in T2.
this also reduces all remaining residuals to zero.
A check is made by substituting the temperatures into equations (1) to (4).
2.6.4 Exterior Nodal Points
Temperatures at the exterior nodal points are usually not specified but can be determined as
part of the numerical analysis. Consider the boundary nodal point subjected to convective
heat transfer in figure 6.

W.A.N.

Page

14

April 2005

MEC551 WANWM

x
2

y
1

Figure 6

n
h

The steady-state energy balance on nodal point n is

kL

T Tn
y T1 Tn

kL(x) 2
2 x
y

kL

y T3 Tn

hL(x)(T Tn ) 0
2 x

2.17

where L is the thickness in the z-direction.


For a square grid, 2.17 simplifies to

1
hx
hx

(T1 2T2 T3 )
(T )
2 Tn 0
2
k
k

2.18

Example 2.5
Derive the nodal temperature equation for the case of an exterior corner node with one
adjacent side insulated and one adjacent side subjected to a convective heat transfer.
Solution

h, T

insulated

x
The rates of energy conducted between nodes 1 and n and between nodes 2 and n are,
respectively,

kL

y T1 Tn

2 x

and

kL

x T2 Tn

2
y

where L is the depth perpendicular to the xy-plane.


The rate of energy convected from T to Tn is

hL

y
(T Tn )
2

For steady state, the summation of energy transfer rates onto node n must be zero, and so,

W.A.N.

Page

15

April 2005

MEC551 WANWM

kL

y T1 Tn
x T2 Tn

+ kL
2
y
2 x

+ hL

y
(T Tn ) = 0
2

For a square grid, x y and the equation simplifies to

T1 T2

W.A.N.

hx
hx

(T )
2 Tn 0
k
k

Page

16

April 2005

You might also like