k-Wave User Forum » Topic: Steel plate in water immersion pulse-echo setup

k-Wave User Forum » Topic: Steel plate in water immersion pulse-echo setup http://www.k-wave.org/forum/topic/steel-plate-in-water-immersion-pulse-echo-setup Support for the k-Wave MATLAB toolbox en-US Wed, 13 May 2026 00:50:43 +0000 http://bbpress.org/?v=1.0.2 <![CDATA[Search]]> q http://www.k-wave.org/forum/search.php Alexll7 on "Steel plate in water immersion pulse-echo setup" http://www.k-wave.org/forum/topic/steel-plate-in-water-immersion-pulse-echo-setup#post-6651 Fri, 16 Nov 2018 08:57:07 +0000 Alexll7 6651@http://www.k-wave.org/forum/ Hi Marioj, I'm pleased that it's quite ok now. I think it is exactly that you said, a trade off. I continually have to do compromises between quality and speed (indeed, I think you can totally remove this oscillation with a very low CFL and/or a very low spatial step). Anyway, wait an answer of the developers to be sure. Best regards, Alex marioj on "Steel plate in water immersion pulse-echo setup" http://www.k-wave.org/forum/topic/steel-plate-in-water-immersion-pulse-echo-setup#post-6650 Thu, 15 Nov 2018 17:00:30 +0000 marioj 6650@http://www.k-wave.org/forum/ Hello Alex. Thanks you for the tips. In fact the sampling frequency were not in accord with the time step. To avoid typing errors I put the sampling frequency as 1/dt. With that change and with CFL=0.15 and dx=dy=50e-6m, for 4 MHz excitation, the signals from the surface and bottom of the steel plate are very similar to the excitation signal. Nevertheless, the are still present some oscillations between surface and bottom signals, but much attenuated than previous. For my code the surface amplitude signal is 145.4 (arbitrary units), first echo from bottom 21.4au and the oscillation arise after this interfaces signals and has the maximum value (0.275au) around the middle of the steel plate. Perhaps it may not be possible to completely extinguish this parasitic signal and should always exist a trade off between this signals and the simulation parameters, namely CFL and grid spacing. Best regards. Mário Alexll7 on "Steel plate in water immersion pulse-echo setup" http://www.k-wave.org/forum/topic/steel-plate-in-water-immersion-pulse-echo-setup#post-6648 Thu, 15 Nov 2018 10:15:46 +0000 Alexll7 6648@http://www.k-wave.org/forum/ Hi Marioj, I'm sorry but I haven't the time looking for the problem (and, probably, I will not find it). However, I have just seen that you use ToneBurst with a sampling frequency of 2e8 whereas your time step is not equal to 1/(2e8) so your Toneburst doesn't have any sense in this framework. The best is to create your time array thanks to maketime and to choose your sampling frequency equal to 1/kgrid.dt. Also, maybe, you can try to increase the CFL to remove your problem of oscillatory behavior. Once again, I'm not a developer. So, take my advices with prudence. Best regards, Alex marioj on "Steel plate in water immersion pulse-echo setup" http://www.k-wave.org/forum/topic/steel-plate-in-water-immersion-pulse-echo-setup#post-6647 Wed, 14 Nov 2018 10:16:10 +0000 marioj 6647@http://www.k-wave.org/forum/ Hello Alex. Thank you for your post! I think that is not the problem. Using a grid point spacing of 25e-6 m for a frequency of 4 MHz (that is 375/25=15 points per wavelength) the problem remains.The collected signal presents a oscillatory behavior, that was not visible for 1 MHz frequency with 100e-6 grid. Best regards. Mário Alexll7 on "Steel plate in water immersion pulse-echo setup" http://www.k-wave.org/forum/topic/steel-plate-in-water-immersion-pulse-echo-setup#post-6645 Tue, 13 Nov 2018 17:55:18 +0000 Alexll7 6645@http://www.k-wave.org/forum/ Hi Marioj, I didn't look at the code precisely and I'm not a developer but you're working with a very low number of points per wavelength (3.75 for 4MHz and under 2ppw from 7.5MHz...). I advice you trying to increase ppw and, particularly, to choose ppw in accordance with the frequency used for emission (using a constant spatial step not taking into account the frequency is not a good idea) Best regards, Alex marioj on "Steel plate in water immersion pulse-echo setup" http://www.k-wave.org/forum/topic/steel-plate-in-water-immersion-pulse-echo-setup#post-6644 Tue, 13 Nov 2018 13:23:41 +0000 marioj 6644@http://www.k-wave.org/forum/ Hello, I am simulating a simple immersion setup using pulse-echo configuration, where a plate of steel in immersed in water. Using 1 MHz frequency everything works well and the recovered signals from the surface and bottom of the steel plate are similar to the excitation signal, with normal smaller amplitudes. Increasing frequency, for instance 4 MHz or higher, the collected signal present a oscillatory behavior, different from the previous one. How can this problem be solved or is this a limitation of k-wave, when dealing with very different acoustic impedance materials among then, such as steel and water? Thanks in advance! Here is the code: % grid XX=21e-3; YY=20e-3; %grid point spacing dx=100e-6; dy=100e-6; % create the computational grid Nx=round(XX/dx); Ny=round(YY/dy); kgrid = makeGrid(Nx, dx, Ny, dy); % transducer d=10e-3; D=round(d/dx); Ti=round(Ny/2-D/2+1); Tf=round(Ny/2+D/2); line = makeLine(Nx, Ny, [1,Ti], [1,Tf]) source.p_mask = line; % define properties of the input signal source_strength = 1e6; % [Pa] tone_burst_freq = 1e6; % [Hz] tone_burst_cycles = 3; % create the input signal using toneBurst input_signal = toneBurst(1e8, tone_burst_freq, tone_burst_cycles); source.p = input_signal; %medium v_matrix=5830; ro_matrix=7920; medium.sound_speed=zeros(Nx,Ny); medium.density=zeros(Nx,Ny); %stell plate in water N1=round(8e-3/dx); %8 mm water N2=round(21e-3/dx)-1;% 13 mm stell N3=round(21e-3/dx) medium.sound_speed((1:N1),:)=1500; % water medium.density((1:N1),:)=1000; medium.sound_speed((N1:N2),:)=v_matrix; % steel medium.density((N1:N2),:)=ro_matrix; medium.sound_speed(N3,:)=1500; % water medium.density(N3,:)=1000; % receiver sensor.mask=line; % input arguments input_args = {'DisplayMask', source.p_mask, 'DataCast', 'single','PMLInside', false,'PMLSize', 20}; % run the simulation sensor_data = kspaceFirstOrder2D(kgrid, medium, source, sensor, input_args{:}); sensor.record ={'p'} pp=sum(sensor_data);