Poynting flux Adjoint#2191

Draft

Alex-Kaylor wants to merge 10 commits intoNanoComp:masterfrom

Alex-Kaylor:poynting_flux

Commits on Aug 5, 2022

Implemenation mostly typed up, but not yet working due to a bug.
Alex-Kaylor
committed
Added test file test_poynting_flux.py in main directory to reproduce bugs.
Alex-Kaylor
committed

Commits on Aug 6, 2022

Fixed forward evalution to match the return of meep's poynting flux function.
Alex-Kaylor
committed

Commits on Aug 9, 2022

Forward run works using dft_fields objects. Adjoint run throws an error in step_db.cpp - Assertion 'changed_materials' failed. Creating a pull request will let me difftool code changes to debug the…
Alex-Kaylor
committed
=Fixed JK_TRANSVERSE, removed some last edits in FourierFields, and added some more TODOs
Alex-Kaylor
committed

Commits on Aug 10, 2022

Fixed JK_TRANSVERSE so the adjiont simulation finishes without error. Off by a factor of ten in initial tests, debugging today.
Alex-Kaylor
committed

Commits on Aug 16, 2022

Fixed sign and conjugate bugs. Changed the implementation from four line sources (one for each tangential H or E component) to a series of point sources to accomodate end-to-end interpolation - we …
Alex-Kaylor
committed

Commits on Aug 25, 2022

Changed implementation back to FourierFields, passing the gradient as dJ. Works well in 2D, fails in 3D. Multiple frequencies untested.
Alex-Kaylor
committed

Commits on Sep 1, 2022

Still only works in 2D. Fixed cases when some field components are zero (due to symmetry, etc.). Tested with eigenmode and point sources radiating toward line monitors.
Alex-Kaylor
committed

Commits on Sep 15, 2022

Fixed a bug where we take the 0th element the return of FourierFields's place_adjoint_source function. This works in 2D, but it returns 9 separate sources in 3D. The extra sources are handled and c…
Alex-Kaylor
committed