"Irregular frequencies" in BEM coefficiencts --> Interpolate?

[dtgaebe]

I noticed some numerical instabilities in the BEM coefficients for the WaveBot and frequencies >0.9 Hz. Refining the mesh didn't solve this. Those are issues well-known to BEM solvers. We might want to address this in the post-porcessing?

The authors in the following suggest some post-processing techniques and have implemented them in an open-source tool BEMRosetta , written in C++. https://www.researchgate.net/publication/355198128_A_post-processing_technique_for_addressing_%27irregular_frequencies%27_and_other_issues_in_the_results_from_BEM_solvers

Me might have some function for our users that detect and interpolate the irregular frequencies as well.

[ryancoe]

I think that capytaine may soon tackle internal frequencies directly (i.e., not a post-processing step; https://github.com/capytaine/capytaine/issues/44), but this paper is useful to know about.

[ryancoe]

We've heard more recently that capytaine is indeed trying to address this problem internally. Thus, for wecopttool, we'll continue to consider this problem out of scope. However, if there's anyone that needs a more immediate solution, you may find it useful to reference:

• Kelly et al. 2021
• Example code provided here: https://github.com/capytaine/capytaine/issues/44#issuecomment-1714854805