I am unclear from the documentation whether attenuation (shear or longitudinal) is supported in elastic wave simulations. Empirically I see that if I set a fairly high attenuation I get unstable solutions.
Any insight about applying attenuation in elastic simulations would be very welcome.
Thanks,
Bill Walker
MATLAB Version: 9.0.0.341360 (R2016a)
Operating System: Mac OS X Version: 10.12.5 Build: 16F73
Java Version: Java 1.7.0_75-b13 with Oracle Corporation Java HotSpot(TM) 64-Bit Server VM mixed mode
----------------------------------------------------------------------------------------------------
MATLAB Version 9.0 (R2016a)
Signal Processing Toolbox Version 7.2 (R2016a)
k-Wave Toolbox Version 1.1
Hi Bill,
k-Wave implements the Kelvin-Voigt model of viscoelasticity, which gives absorption with a frequency squared dependence (unless you're operating at very high frequencies / absorptions). You can set these as explained in the Plane Wave Absorption Example under the Elastic Wave Propagation heading in the MATLAB help. As with the fluid model, very high values of absorption will require smaller time steps to remain stable.
Brad.
Hi Brad,
First of all, great effort creating such a useful toolbox. The simulation works really well, and I want to thank you for putting the resources developing it.
I am simulating acoustic wave field in an elastic medium that is biological in nature. The absorption power does not necessarily follow the frequency square, in fact, it is very close to 1. I know that you used fractional Laplacian to simulate absorption, but is there a way to customize the value of absorption power in the pstdElastic2D solver? I greatly appreciate your response.
Yishi
Hi Yishi,
The function pstdElastic2D solves a relatively simple visco-elastic wave model as described in this paper. This doesn't use the fractional Laplacian to implement the absorption, and gives frequency squared absorption. We have experimented with this for the elastic code (as described here), but the computational load is very heavy, so we haven't released this code.
If you're working with continuous wave inputs (or at least something with a centre frequency), one option would be to adjust the alpha_coeff values to give the correct absorption at the frequency of interest.
Hope that helps,
Brad.
You must log in to post.