Skip to content

Conversation

@amorimd
Copy link

@amorimd amorimd commented Nov 12, 2025

📝 Pull Request Summary

The PML is modified to allow customization of the conductivity function and to match the properties of the last layer for $\sigma$ and $\epsilon_r$

🔧 Changes Made

  • Change the way $\sigma$ and $\epsilon^{-1}$ tensors are filled for the PML layers
  • Add tests for a plane wave with resistive materials

To Do

  • Check that the tests are correct
  • Add the $\mu^{-1}$ tensor filling (it still assumes $\mu_r=1$)

✅ Checklist

  • Code follows the project's style and guidelines
  • Added tests for new functionality
  • Updated documentation (mentioned in docs/ or included in examples/ and notebooks/)
  • Verified that changes do not break existing functionality (automatic tests passing)

📌 Related Issues / PRs

Related to feature request #33

amorimd and others added 16 commits August 13, 2025 15:35
Besides the number of PML cells, the resistivity boundaries, permittivity and the function used to compute the resistivity profile in the PML are now arguments for the SolverFIT3D
To make the init clearer, the PML parameters are given default values within the initialization. To customize the values for PML resistivity ans eps, user can access the attributes and re-execute fill_pml_sigmas to update the PML cells.
This first attempt uses the properties of the central cell of the layer adjacent to the PML, assuming the beam is centered
Relaxation times were incorrect when sigma=0 in the simulation space, as the whole PML had sigma=0
… cases other than vacuum

The sigma for the PML layers is now the sigma of the previous non-PML layer + the function of resistivity. This function starts from the sigma value of the layer if it is non zero, or from a low sigma value
This is to improve field absorption, in particular if we have vacuum. One might need more layers of PML then
This still needs to be worked out, as there are some divergence in computations
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant