Merge pull request #2 from ucl-bug/large-domain

test: 512 homogeneous domain

examples/test_against_kWave/example_accuracy_test_homog_large.m

@@ -0,0 +1,69 @@
+% Comparison between k-Wave and the GMRES Helmholtz solver in the case of
+% heterogeneous absorption
+
+
+%% Setup
+N = 512;                                     % number of grid points in each direction
+size_x = 40/1000;                            % Domain size [m]
+dx = size_x / N;                             % step size [m]
+
+omega = 2*(10^6)*2*pi;                       % angular frequency [rads/s], 2*pi*f_0
+c0 = 1300;                                   % Homogeneous speed of sound [m/s]
+abs_coeff = zeros(N,N); % absorption coefficient [dB/cm]
+
+src_location = [256,256];                      % Source location in gridpoints
+pml_size = 40;                               % PML size in gridpoints
+sigma_star = 4*omega;                        % chosen for reasons of stability?
+
+% This is run by example_accuracy_test_hetero_abs.m
+
+% run the GMRES Helmholtz solver
+p = heterog_helmholtz_solver(omega, c0, abs_coeff, dx, N, src_location, pml_size, sigma_star);
+
+% run the k-Wave solver for a single frequency 
+p_kwave = kwave_solution(N, c0, abs_coeff, src_location, omega, c0, 0.0, 0.1, 2, dx);
+
+% Fix amplitude in kwave solution
+p_kwave = conj(p_kwave);
+p_kwave = p_kwave*p(src_location(1),src_location(2))/ ...
+    p_kwave(src_location(1),src_location(2)); 
+
+
+% PLOTS
+figure;
+subplot(2,2,1);
+imagesc(real(p))
+axis off
+axis equal
+caxis([-0.2, 0.2])
+title("GMRES result")
+colorbar
+axis image
+
+subplot(2,2,2);
+imagesc(real(p_kwave))
+axis off
+axis equal
+title("kWave solution")
+caxis([-0.2, 0.2])
+colorbar
+axis image
+
+subplot(2,2,3);
+imagesc(100*abs(p - p_kwave)/max(abs(p_kwave(:))))
+axis off
+axis equal
+title("Percent Error")
+colorbar
+axis image
+caxis([0, 2])
+drawnow
+
+subplot(2,2,4);
+imagesc(abs_coeff)
+axis off
+axis equal
+title("absorption coefficient [dB/cm]")
+colorbar
+axis image
+drawnow