Skip to content

Add analytical Alfvén perturbations to guiding-centre tracing - #177

Open
eduardolneto wants to merge 10 commits into
mainfrom
feat/analytical-alfven-perturbations
Open

eduardolneto wants to merge 10 commits into
mainfrom
feat/analytical-alfven-perturbations

Conversation

@eduardolneto

Copy link
Copy Markdown
Contributor

Adds prescribed electromagnetic perturbations to the existing Boozer and VMEC guiding-centre equations, including magnetic flutter, induction and the required parallel-force correction. Reuses existing integrators and preserves constant magnetic moment during wave energy exchange.
Includes a small CPU example, conservation benchmarks, and the 100-particle Paul et al. loss comparison with documented inputs, figures and reproduction commands. The Paul et al. comparison remains a small-sample comparison; quantitative paper reproduction is not established.

@rogeriojorge

Copy link
Copy Markdown
Member

Eduardo, status on the Alfvén perturbations.

Main has moved (loss API, tracing defaults, VMEC refactor). Your branch still merges main cleanly. I checked that the axis-stop results and the adjoint argument in Tracing survive the merge.

There are two PRs into this branch, meant to be merged in order:

#196, slimming. Same physics and API.

  • One Perturbation base for the Boozer and guiding-centre versions (+628 / -2877).
  • A Landau-resonance test (White): energy grows only at resonance and equals the parallel work, with mu constant.
  • The Paul benchmark example uses the parameters-at-top template.
  • The slow adaptive Dopri8 tests are shorter.
  • The 1.5 MB wout_LandremanPaul2021_QA_lowres.nc is replaced by the reactor-scale QA file already in the repo. It is the same equilibrium at lower resolution. The 100-marker loss table moves within its sampling error (Boozer with wave 21.6% -> 22.3%, VMEC with wave 21.6% -> 24.3%).

#235, FAX eigenmodes.

  • Builds on Simplify the analytical Alfven perturbation PR #196 and adds FaxEigenmode.from_fax(directory, amplitude).
  • Reads FAX v0.0.2 output (eigenvalue and per-(m, n) radial structure) into the existing Boozer perturbation, with no new dependency.
  • Includes one example and three tests: round trip against the FAX data, E.B0 = 0 with exact zero-amplitude limit, and energy change equal to wave work.
  • The committed mode is a demonstration. Its frequency depends strongly on FAX's radial-stiffness model, so it is not a validated AE yet.

If both look fine to you, merge #196 into this branch, then #235.

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.

2 participants