Sponge Source term at the boundaries

[cjlauer16]

I want to add a source term near the inlet and outlet to dampen out any waves and keep the boundaries more stable and am having trouble finding an example for how to do that. Can you help point me in the right direction for how to implement this?

[baperry2]

For now, to add arbitrary source terms, the procedure is to modify the fill_ext_source function as needed:

PeleC/Source/External.cpp

Line 40 in 33ad884

PeleC::fill_ext_source(

When #743 gets merged, custom source terms will be able to be added by defining a function in the prob files, and there may even be similar capability designed for sponge zones!

[cjlauer16]

Thank you! That's great for now, and I'll keep an eye out for the new capabilities.

[cjlauer16]

What is the best way to get coordinate information into the fill_ext_source function? I have a sponge source set up based on the (i,j,k) indices used to loop over the mesh, but I'd like to use (x,y,z) coordinate values instead to better deal with changing meshes when running with amr. Thank you!

[baperry2]

We typically manually compute this as needed based on the i,j,k index and the the AMReX geometry data. Use const auto geomdata = geom.data(); to get the GPU-safe AMReX GeometryData from the Geometry object that is part of the PeleC class. Then compute the coordinate values as is done here:

PeleC/Source/Derive.cpp

Line 757 in 3cef3ca

AMREX_D_TERM(coord_arr(i, j, k, 0) = prob_lo[0] + (i + 0.5) * dx[0];

[baperry2]

#743 has now been merged. You can now create custom source terms with the problem_modify_ext_sources function or add artificial viscosity using the problem_modify_transport_coeffs function. The EB-ConvergingNozzle case now has an example of using the latter approach to dampen waves at the boundaries - probably not the ideal approach for that but it did substantially reduce spurious wave reflections happening in that case.