Discussion Forum

Hello all!

I was wondering if anyone has had any success with using a perfectly matched layer in an eigenfrequency analysis in the Solid Mechanics module. I'm modeling essentially a cylinder mounted on an infinite substrate made out of approximately the same material and am interested in capturing resonance frequencies just in the cylinder.

I'm interested in a relatively large range of frequencies, so I have parameterized many of the settings. To remain computationally efficient, I break the problem down into a series of small eigenfrequency analyses set to solve for frequencies within a much smaller band and then optimize settings for that frequency band. Utilizing the region setting for the eigenfrequency solver, and setting the imaginary component bounds both to 0, I am able to extract just resonance modes in the cylinder and can completely exclude modes appearing in the substrate domain. I configure the PML to be optimized for the longest wavelength for a particular frequency band and mesh to ensure the smallest wavelength in the band can be captured sufficiently. Doing this, I am able to almost completely eliminate the imaginary part of these frequencies.

So, I have previously been told that what I am doing shouldn't be done -- that is, that PMLs and eigenfrequency analyses in Solid Mechanics should not be attempted. But, my experience so far has been that this seems to be working. So, my question again to everyone is if anyone has had any success with this previously and if there are any problems or issues that anyone is aware of that might crop up down the road.

Thanks!