search

jackhamel16 / Acoustics

Acoustic integral equation solver
1 stars 0 forks source link

Could solve at WS modes using other integral equations #85

Closed jackhamel16 closed 3 years ago

jackhamel16 commented 3 years ago

The sound soft IE is only required for construction of the scattering matrix and WS matrix. Once that is done, and the desired excitation for the WS mode is constructed, any of the IE can be solved with that excitation. I could go in and add this feature to my code, if it becomes necessary.

jackhamel16 commented 3 years ago

focus on adding CFIE with ACA for now

jackhamel16 commented 3 years ago

I need to find another way to compute the weighting coefficient computed here

Try printing out the scale factor for a few different meshes solving the CFIE. I have a hunch it might be close to some constant that I can hard code.

jackhamel16 commented 3 years ago

I need to find another way to compute the weighting coefficient computed here

Try printing out the scale factor for a few different meshes solving the CFIE. I have a hunch it might be close to some constant that I can hard code.

I tested a variety of meshes (rectangular plates, strips, circular plates, cubes, spheres) and these show the various computes weights image_of_weights Based on this, I probably dont need to weigh them at all since they are of the same magnitude, generally. This is probably because the propagation speed of my waves is 1.

jackhamel16 commented 3 years ago

Maybe test using the non-ACA CFIE solver without a scale factor on normal derivative of Z

jackhamel16 commented 3 years ago

Running the analytical sphere test with the CFIE with and without the weighting and it looks like using no weight yields lower errors. With weight: sphere_convergence_data_softCFIE.txt sphere_convergence_results_softCFIE without weight: sphere_convergence_data_softCFIE.txt sphere_convergence_results_softCFIE

jackhamel16 commented 3 years ago

I am going to remove the weight in this issue in both the ACA and non-ACA CFIE solvers

jackhamel16 commented 3 years ago

Everything is there, but I am now seeing problems with CFIE vs non-CIFE results. I think maybe these is an issue with using a WS excitation. Right now, in test.jl, results for the circular plate agree well, but in the test for the CFIE with ACA solve the CFIE results are double the non-CFIE results. The difference between the test and what is done in test.jl is the excitation (at least that is what I see initially)