Assignment-5: 1-d Transient Heat Conduction Through Cylinder
Assignment-5: 1-d Transient Heat Conduction Through Cylinder
Assignment-5 ID : - 51120
Thermal Engineering(2016-18)
1-d transient heat conduction through cylinder
Problem: A 20 cm diameter cylindrical shaft made of stainless steel comes out of an oven at a
uniform temperature of 600oC. The shaft is then allowed to cool slowly in an air chamber at
200oC with an average heat transfer coefficient h=80 W/m2-oC. Determine the temperature
distribution along the radius by taking 10 points along the radius after 45min from the cooling
process starts. Take k=14.9 W/m-oC, cp= 477 J/kg-0C, = 3.9510-6 m2/s and =709 W-kg/m3,
using finite difference method (FDM). Compare the results with analytical solution Heisler
Chart.
Solution:
The governing equation for one dimensional heat conduction in cylindrical coordinates is:-
1 1
( ) =
Where, T is temperature, r is radius of cylinder, t is time and is thermal diffusivity which is,
= k/C . (where k= thermal conductivity, C= Thermal capacity). These all terms have some
units.
changing above equation in non-dimensional form :-
2 1
= +
R2 R R
t r TTi
Where, =
r2o
; R= ; =
r0 T Ti
Discretization -
Governing equation is discredited using forward and central difference as follows.
For numerical solution the governing equation is discretized using central difference method
taking grid-size equals to on both the axis.
+1 = + (1
2 + +1
)+
(1
+ +1 ) (for interior grid points)
2(1)
At R=1
+1 2 2 2 2
+1 = +1 + (2 2+1 +1 + ) + (2 +1 + )
2
At R=0
1+1 = 1 + 2(22 21 ).
Where i, = 1, 2, 3m are the grid points along x and m is the number of grids.
The following program is developed to solve these equations in MATLAB
MATLAB Program:
clc; clear;
q = 2000 ; k = 385 ; T0 = 25 ; L = 0.5;
h=1/4 ; m=1/h; n=m*m;
a1(1:n)=4;
a2(1:n-1)= -1; a2(1:m:n-(m-1))=-2; a2(m:m:n-m)=0;
a3(1:m)= -2; a3(m+1:n-m)= -1;
a4(1:n-1)= -1; a4(m:m:n-m)= 0;
a5(1:n-m)= -1;
A=diag(a1)+diag(a2,1)+ diag(a3,m)+ diag(a4,-1)+ diag(a5,-m);
b(1:n)= h^2;
P = A;
for k = 2:n
fori = k:n
b(i)= b(i)-(P(i,k-1)/P(k-1,k-1))*b(k-1);
P(i,:)= P(i,:)-(P(i,k-1)/P(k-1,k-1))*P(k-1,:);
end
end
X(n,1) = b(n)/P(n,n);
fori = n-1 : -1 :1
X(i,1)=(b(i)-P(i,i+1:n)*X(i+1:n,1))/P(i,i);
end
Y=zeros(m,m);
for j = 1:m;
Y(1:m,j)=X(((j-1)*m+1):j*m);
end
Z=zeros(2*m+1,2*m+1);
fori = 1 : m
Z(m+i,m+1: 2*m)=Y(i,1:m);
Z(m+i,m-(0:m-2))= Z(m+i,m+(2:m));
end
Z(m-(0:m-2),:)=Z(m+(2:m),:);
T(1:2*m+1,1:2*m+1)=Z(1:2*m+1,1:2*m+1)*q*L^2/k+T0;
x=linspace(0,1,2*m+1);
y=linspace(0,1,2*m+1);
contour(x,y,T)
Results:
Following figure shows the steady state temperature distribution obtained for the Steel plate.
The length and width of the plate is shown along x and y axis and the different curves shown
are the isotherms at different temperature.
Fig.3 Curves of different isothermal lines obtained for non dimensional parameters
The above stated problem is solved for square Cu plate thermal conductivity (k=385W/m-
k) of side 1 m. The heat is generating inside the plate at the rate of 2000 W/m2. The temperature
at the edges of the plate is taken to be 25oC. solve same for.
Fig.4 Curves of different isotherms (in oC) obtained for Cu plate
Comments:
The computed temperature satisfy the finite difference equations, equation are
approximate whose accuracy may be improved by reducing the grid size or increasing the
number of nodal points.
Isothermal lines are square if we take small grid size.