Solving coupled pdes with FFT
14 visualizaciones (últimos 30 días)
Mostrar comentarios más antiguos
I have been trying to generalise some working code from Professor Steve Brunton which solves the one way wave pde using the FFT method. This is from his book at http://databookuw.com The explanation of the method and the code is at https://www.youtube.com/watch?v=RilggJd1_LY I paste the code below and it works very well.
I would like to solve a pair of coupled 'two-way' wave equations by the same method. In fourier space the equivalent first order system is
uhat_t = phat
vhat_t = qhat
phat_t = mu^2 k^2 uhat + qhat
qhat_t - lambda^2 mu^2 k^2 vhat -phat
where lambda and mu are real constants, k is the spatial frequency and the fourier transformed vector of unknowns is [uhat,vhat,phat,qhat]
Clearly, there is a vector of unknowns instead of just one scalar unknown in the working example. It is not clear to me that ode45 can be used for this problem as the 2D array of output for the scalar problem would somehow need to be extended to a 3D array with the use of a 'page' dimension. I have attemepted this but cannot get anywhere.
Does anyone have an example of such a higher dimensional problem or is it possible to adapt the code below?
Many thanks in anticipation of your help.
clear all, close all, clc
% Define spatial domain
c = 2; % Wave speed
L = 20; % Length of domain
N = 1000; % Number of discretization points
dx = L/N;
x = -L/2:dx:L/2-dx; % Define x domain
% Define discrete wavenumbers
kappa = (2*pi/L)*[-N/2:N/2-1];
kappa = fftshift(kappa'); % Re-order fft wavenumbers
% Initial condition
u0 = sech(x);
uhat0 = fft(u0);
% Simulate in Fourier frequency domain
dt = 0.025;
t = 0:dt:100*dt;
[t,uhat] = ode45(@(t,uhat)rhsWave(t,uhat,kappa,c),t,uhat0);
% Alternatively, simulate in spatial domain
[t,u] = ode45(@(t,u)rhsWaveSpatial(t,u,kappa,c),t,u0);
% Inverse FFT to bring back to spatial domain
for k = 1:length(t)
u(k,:) = ifft(uhat(k,:));
end
% Plot solution in time
subplot(1,2,1)
h=waterfall(real(u(1:10:end,:)));
set(h,'LineWidth',2,'FaceAlpha',.5);
colormap(jet/1.5)
subplot(1,2,2)
imagesc(flipud(real(u)));
colormap jet
%% FIGURES (PRODUCTION)
figure
CC = colormap(jet(101));
dt = 0.025;
for k = 1:length(t)
u(k,:) = real(ifft(uhat(k,:)));
if(mod(k-1,10)==0)
plot(x,u(k,:),'Color',CC(k,:),'LineWidth',1.5)
hold on, grid on, drawnow
end
end
% xlabel('Spatial variable, x')
% ylabel('Temperature, u(x,t)')
axis([-L/2 L/2 -.1 1.1])
set(gca,'LineWidth',1.2,'FontSize',12);
set(gcf,'Position',[100 100 550 220]);
set(gcf,'PaperPositionMode','auto')
%
figure
subplot(1,2,1)
h=waterfall(u(1:10:end,:));
set(h,'LineWidth',2,'FaceAlpha',.5);
colormap(jet/1.5)
% view(22,29)
view(3,21)
set(gca,'LineWidth',1.5)
set(gca,'XTick',[0 500 1000],'XTickLabels',{})
set(gca,'ZTick',[0 .5 1],'ZTickLabels',{})
set(gca,'YTick',[0 5 10],'YTickLabels',{})
subplot(1,2,2)
imagesc(flipud(u));
set(gca,'LineWidth',1.5)
set(gca,'XTick',[0 500 1000],'XTickLabels',{})
set(gca,'YTick',[0 50 100],'YTickLabels',{})
colormap jet
set(gcf,'Position',[100 100 600 250])
set(gcf,'PaperPositionMode','auto')
__________________________________________________
function duhatdt = rhsWave(t,uhat,kappa,c)
duhatdt = -c*i*kappa.*uhat;
________________________________________________
function dudt = rhsWaveSpatial(t,u,kappa,c)
uhat = fft(u);
duhat = i*kappa.*uhat;
du = ifft(duhat);
dudt = -c*du;
0 comentarios
Respuestas (0)
Ver también
Categorías
Más información sobre Geometry and Mesh en Help Center y File Exchange.
Community Treasure Hunt
Find the treasures in MATLAB Central and discover how the community can help you!
Start Hunting!