Hello everybody,
whats is wrong with the code.it return the errors
"Error in makeGrid (line 83)
kgrid = makeGridStruct(varargin{:});
"Error in k_GRid (line 11)
kgrid=makeGrid(Nx, dx, Ny, dy);"
clear all;
clc;
x_size=3000;
y_size=3000;
f_max= 50000;
c0_min=1400;
dx=c0_min/(2*f_max);
dy=dx;
Nx=round(x_size/dx);
Ny=round(y_size/dy);
kgrid=makeGrid(Nx, dx, Ny, dy);
Hi Zia,
The problem is that your grid size is extremely large (each matrix is 343 GB), so you are likely seeing out of memory errors. Simulating 50 kHz waves in water over 3 km is a very large domain, even in 2D.
Brad.
Hi Bradely Treeby.
I need to simulate a sound source 3Km away from an array of 3-sensors 10m apart from each other. the medium is sea-water.The grid spacing used can be dx=dy. i have the latest core i7 machine for simulation. could you please suggest how i can do it.
Hi Zia,
It's easy enough to check why this isn't feasible using k-Wave if you check the numbers:
% simulation parameters
f = 50e3; % [Hz]
c = 1500; % [m/s]
d = 3e3; % [m]
% compute dx and Nx
ppw = 2;
dx = c/(ppw*f);
Nx = round(d/dx);
% compute memory used for each matrix
mem = Nx^2 * 8 / 1024^3;The memory required to store one 2D matrix of this size in double-precision is 300 GB!
Depending on your source and boundary conditions, you could look at using the boundary element method, or solving the Rayleigh integral.
Brad.
You must log in to post.