Scribd
Upload
148 views

Matlab Program For Second Order FD Solution To Poisson's Equation

The document describes a Matlab program that uses a second order finite difference method to solve Poisson's equation on a square domain with Dirichlet boundary conditions. The program takes in the functions that define the right hand side and boundary values, and outputs the numerical solution on a uniform grid. An example is also given that uses the program to solve the Poisson equation -∇^2u = -5π^2sin(πx)cos(2πy) on the domain (0,1)x(0,1) with boundary values u = sin(πx)cos(2πy). Plots of the numerical solution and error are generated to validate the method.

Uploaded by

ugo
Copyright
© © All Rights Reserved
Available Formats
Download as PDF, TXT or read online on Scribd
148 views

Matlab Program For Second Order FD Solution To Poisson's Equation

The document describes a Matlab program that uses a second order finite difference method to solve Poisson's equation on a square domain with Dirichlet boundary conditions. The program takes in the functions that define the right hand side and boundary values, and outputs the numerical solution on a uniform grid. An example is also given that uses the program to solve the Poisson equation -∇^2u = -5π^2sin(πx)cos(2πy) on the domain (0,1)x(0,1) with boundary values u = sin(πx)cos(2πy). Plots of the numerical solution and error are generated to validate the method.

Uploaded by

ugo
Copyright
© © All Rights Reserved
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 3

Matlab Program for Second Order FD

Solution to Poisson’s Equation


Code:
0001 % Numerical approximation to Poisson’s equation over the square [a,b]x[a,b] with
0002 % Dirichlet boundary conditions. Uses a uniform mesh with (n+2)x(n+2) total
0003 % points (i.e, n interior grid points).
0004 % Input:
0005 % ffun : the RHS of poisson equation (i.e. the Laplacian of u).
0006 % gfun : the boundary function representing the Dirichlet B.C.
0007 % a,b : the interval defining the square
0008 % m : m+2 is the number of points in either direction of the mesh.
0009 % Ouput:
0010 % u : the numerical solution of Poisson equation at the mesh points.
0011 % x,y : the uniform mesh.
0012 %
0013 function [u,x,y] = fd2poisson(ffun,gfun,a,b,m)
0014
0015 h = (b-a)/(m+1); % Mesh spacing
0016
0017 [x,y] = meshgrid(a:h:b); % Uniform mesh, including boundary points.
0018
0019 idx = 2:m+1;
0020 idy = 2:m+1;
0021
0022 % Compute boundary terms, south, north, east, west
0023 ubs = feval(gfun,x(1,1:m+2),y(1,1:m+2)); % Include corners
0024 ubn = feval(gfun,x(m+2,1:m+2),y(m+2,1:m+2)); % Include corners
0025 ube = feval(gfun,x(idy,m+2),y(idy,m+2)); % No corners
0026 ubw = feval(gfun,x(idy,1),y(idy,1)); % No corners
0027
0028 % Evaluate the RHS of Poisson’s equation at the interior points.
0029 f = feval(ffun,x(idy,idx),y(idy,idx));
0030
0031 % Adjust f for boundary terms
0032 f(:,1) = f(:,1) - ubw/h^2; % West
0033 f(:,m) = f(:,m) - ube/h^2; % East
0034 f(1,1:m) = f(1,1:m) - ubs(idx)/h^2; % South
0035 f(m,1:m) = f(m,1:m) - ubn(idx)/h^2; % North
0036
0037 f = reshape(f,m*m,1);
0038
0039 % Create the D2x and D2y matrices
0040
0041 % Full matrix version. Can be made faster with Matlab’s sparse library.
0042 z = [-2;1;zeros(m-2,1)];
0043 D2x = 1/h^2*kron(toeplitz(z,z),eye(m));
0044 D2y = 1/h^2*kron(eye(m),toeplitz(z,z));
0045
0046 % Solve the system
0047 u = (D2x + D2y)\f;
0048
0049 % Convert u from a column vector to a matrix to make it easier to work with
0050 % for plotting.
0051 u = reshape(u,m,m);
0052
0053 % Append on to u the boundary values from the Dirichlet condition.
0054 u = [ubs;[ubw,u,ube];ubn];
0055
0056 end
Example:
Use the above Matlab code to solve the Poisson problem

∇2 u(x, y) = −5π 2 sin(πx) cos(2πy) for (x, y) ∈ Ω = (0, 1) × (0, 1)


u(x, y) = sin(πx) cos(2πy) for (x, y) ∈ ∂Ω

0001 % Script for testing fd2poisson over the square [a,b]x[a,b]


0002 a = 0;
0003 b = 1;
0004 k = 6;
0005 m = 2^k-1; % Number of interior grid points in one direction
0006
0007 f = @(x,y) -5*pi^2*sin(pi*x).*cos(2*pi*y); % Laplacian(u) = f
0008 g = @(x,y) sin(pi*x).*cos(2*pi*y); % u = g on Boundary
0009 uexact = @(x,y) g(x,y); % Exact solution is g.
0010
0011 % Compute and time the solution
0012 tic
0013 [u,x,y] = fd2poisson(f,g,a,b,m);
0014 gedirect = toc;
0015
0016 fprintf(’Direct Gaussian elimination take %d s\n’,gedirect);
0017
0018 %% Plot solution
0019 figure, set(gcf,’DefaultAxesFontSize’,10,’PaperPosition’, [0 0 3.5 3.5]),
0020 surf(x,y,u), xlabel(’x’), ylabel(’y’), zlabel(’u(x,y)’),
0021 title(strcat(’Numerical Solution to Poisson Equation, h=’,num2str(h)));
0022
0023 % Plot error
0024 figure, set(gcf,’DefaultAxesFontSize’,10,’PaperPosition’, [0 0 3.5 3.5]),
0025 surf(x,y,u-uexact(x,y)),xlabel(’x’),ylabel(’y’), zlabel(’Error’),
0026 title(strcat(’Error, h=’,num2str(h)));

Numerical Solution to Poisson Equation, h=0.025

0.5

0
u(x,y)

−0.5

−1

−1.5
1
1
0.5
0.5

y 0 0 x

2
Error, h=0.2 Error, h=0.1

0 0.01

0
−0.05

−0.01
u(x,y)

u(x,y)
−0.1
−0.02

−0.15
−0.03

−0.2 −0.04
1 1
1 1
0.5 0.5
0.5 0.5

y 0 0 x y 0 0 x
Error, h=0.05 Error, h=0.025

−3 −3
x 10 x 10

2 0.5

0 0

−2 −0.5
u(x,y)

Error

−4 −1

−6 −1.5

−8 −2

−10 −2.5
1 1
1 1
0.5 0.5
0.5 0.5

y 0 0 x y 0 0 x

You might also like