I am trying to simulate the photoacoustic signals detected by an ultrasound transducer. The 2D axisymmetric geometry consists of the carrier medium, water, and a transducer. I fed pressure as an initial value to k-wave.
While running the solver, the acoustic waves look similar to the literature in the case of a homogenous medium. However, after adding the second medium, i.e., the transducer's physical properties, the propagating acoustic waves get noisy.
I appreciate any help in this regard.
Hi Farzin,
Can you post some more details (or preferably a simple example) of what you mean?
Brad.
Dear Brad,
Thank you for your response. The attached codes are a simplified example in MATLAB.
Farzin
P.S. The output figure: https://i.ibb.co/MR2KDCZ/transducer-ct-pt.png
Nz = 560;
Nr = 125;
kgrid = kWaveGrid(Nz, 1e-04, Nr, 1e-4);
source.p0 = zeros (Nz, Nr);
source.p0(30:100,1:50) = 1;
sensor.mask = [520;1;530;95];
medium.sound_speed = 1500 * ones(Nz, Nr);
medium.density = 1000 * ones(Nz, Nr);
sensor_data_hom = kspaceFirstOrderAS(kgrid, medium, source, sensor);
medium.sound_speed(520:end,:) = 2200;
medium.density(520:end,:) = 1780;
sensor_data_het = kspaceFirstOrderAS(kgrid, medium, source, sensor);
figure('Name','Trace from center of transducer');
plot(kgrid.t_array*1e6,squeeze(sensor_data_hom(1).p(1, 1, :)), 'r-');
xlabel('Time (us)');
ylabel('Pressure');
hold on
plot(kgrid.t_array*1e6,squeeze(sensor_data_het(1).p(1, 1, :)), 'b-');
legend('Homogenous', 'Heterogeneous');
axis tight;
hold off
Hi Farzin,
By default, the reference sound speed used in the k-space operator is set to the maximum to make the simulation unconditionally stable. However, because most of the medium is 1500 m/s in your simulation, this means dispersion errors accumulate. Try setting medium.sound_speed_ref = 1500 (or alternatively reduce the size of the time step) and it should solve the problem. See here (Sec V.A) if you want to read more.
Brad
Thank you for the informative hint and the reference.
You must log in to post.
