Discussion Forum

Hello Merlin Morlock

Your Discussion has gone 30 days without a reply. If you still need help with COMSOL and have an on-subscription license, please visit our Support Center for help.

If you do not hold an on-subscription license, you may find an answer in another Discussion or in the Knowledge Base.

Have you got any solution to this problem?

I tried myself to simulate ultrasonic waves in time domain

1) PML no effect, seems it could only be used in Frequency domain. I think COMSOL should give us some warning like "PML is not applicable in time-dependant analysis" at least, instead of saying nothing.
2) Infinite elements, no effect. Maybe it can only be used for static and diffusion type problems but not wave problems?
3) Low reflecting BC, some effect. Still there exists reflection, but better than PML and Infinite elements.

So, I suppose there is no perfect solution in time-dependant analysis?

I have run a few small tests in both 2D and 2D axisymmetric for simple beam and cylindrical geometries and noticed a few things. For the 2D beam case with an initial displacement or force on the incident boundary, if you place a periodic - continuity condition on the two outside boundaries you are left with a symmetric and flat stress wave-front and the low-reflecting boundary will not reflect any of the incident stress wave. For the axisymmetric case, if you place a symmetric boundary condition on the outside boundary you will again be left with a single flat stress wave-front and the low-reflecting boundary works well with zero elastic wave reflection. I think that the problem arises when the outside faces are left free and the wave-front is symmetric, but not flat which is how it would actually be in a real-world test. Maybe the low-reflecting boundary condition has problems with elastic waves coming in from different directions?

Hi,

Maybe the low-reflecting boundary condition has problems with elastic waves coming in from different directions?

This is so. The low-reflecting boundary is ideal only for normal incidence, but is does a decent job also for oblique waves if the angle of incidence is not too large. The theory can be found in the users guide for Structural Mechanics Module.

Regards,
Henrik

Thanks for the reply Henrik. In my experience the best method for reducing reflection from boundaries during transmission/absorption tests involving elastic waves is to make the incident and transmission domains very large compared to the absorption region (spherical cavity scattering). This is of course more expensive with respect to computation time. Is there a better method for cancelling boundary reflections? Maybe boundaries with the same geometry as the scattered waves like a hemispherical shape?