0% found this document useful (0 votes)
40 views

Computational Assignment AE603

This document outlines two computational assignments for solving the 2D steady-state heat equation using finite difference methods. The first part involves writing a code to directly solve the linear system using Gaussian elimination. Sample inputs/outputs are provided. The second part involves implementing iterative Jacobi, Gauss-Seidel and SOR methods and comparing storage requirements and convergence rates between the direct and iterative methods.

Uploaded by

Abhijit Kushwaha
Copyright
© © All Rights Reserved
Available Formats
Download as PDF, TXT or read online on Scribd
0% found this document useful (0 votes)
40 views

Computational Assignment AE603

This document outlines two computational assignments for solving the 2D steady-state heat equation using finite difference methods. The first part involves writing a code to directly solve the linear system using Gaussian elimination. Sample inputs/outputs are provided. The second part involves implementing iterative Jacobi, Gauss-Seidel and SOR methods and comparing storage requirements and convergence rates between the direct and iterative methods.

Uploaded by

Abhijit Kushwaha
Copyright
© © All Rights Reserved
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 2

AE603 - Computational Assignment

Due November 14, 2014

1. Finite Difference (FD) Code for Solving the Heat Equation using direct solution
method (Gaussian elimination followed by back substitution):
Write a computer code that implements the Finite Difference (FD) method for solving the
two-dimensional steady state heat conduction equation with heat source on a square
plate. Essentially this results in a system of linear equations, Ax=b form. For simplicity
use the following assumptions:
a) The mesh is uniform.
b) A uniformly distributed source.
c) The material composition is homogeneous, i.e. constant thermal conductivity: k.
d) Temperature of the walls prescribed, on all four faces of the square plate.
Write your code in fortran/C/matlab only. Include a README text file that states:
i. How to compile the code: the line command necessary to compile.
ii. How to execute the code: the line command necessary to execute.
iii. Status of your code: operational/compiles-doesnt-run/doesnt compile.
iv. Briefly describe the problem solved by your code, expected input, resulting
output, any limitations or restrictions.
A sample input file (using the values listed below) and corresponding output file
produced by your code. Format of the input file may be in the following
sequence (all values in SI units):
Line
Number

Value

Meaning

Length of the plate

8/12/16/20

I = Number of computational cells per dimension

300.0

Tw = Wall temperature

100.0

s = Uniformly distributed source strength

50.0

k = Thermal conductivity

Note: your code should function for general (but permissible) input values not just the above;
the sample input serves only as a check of your successful completion of the assignment.

a. Compute temperature values at the following points (x, y) = (0.25,0.25), (0.5,0.5),


(0.75,0.75).
b. Plot each of these values with respect to the value of I, the number of computational cells
per dimension.
2. Repeat the problem now considering an iterative solver, Point-Jacobi, Gauss-Seidel,
SOR. Use the following table:

Line
Number

Value

Meaning

Length of the plate

8/12/16/20

I = Number of computational cells per dimension

300.0

Tw = Wall temperature

100.0

s = Uniformly distributed source strength

50.0

k = Thermal conductivity

6.

1.e-6

c = Relative convergence criterion

7.

0/1/2

Solver used, PJ=0, GS=1, SOR=2

The optimal weight for the SOR method, , is given by:

2
, where I is as
1 sin / I

defined above.
c. Compare the direct solution method with the GS method with respect to the storage
requirement (as a function of I).
d. Compare the three iterative methods with respect to the number of iterations needed for
the solution to converge to the pre-specified tolerance, i.e., Max( ) c , where is the
relative convergence error.

You might also like