Repository navigation
ForceFreeStates/DCON Eliminates the large solution even when D_I < -1/4 #199
Description
According to section 2.3.3 of Jong-Kyu's thesis discusses internal boundary conditions at rational surfaces. He brings up how DCON systematically eliminates the large solution near rationals, which is necessary for mercier unstable perturbations to integrate to integrate to a finite delta W. However, for the purposes of perturbed equilibria, the large solution is a valid solution, and we are needlessly throwing it out. Enabling this solution should be investigated.
Activity
matt-pharr commentedon Apr 3, 2026
I implemented this in a screw pinch and found that including the large solution does slightly modify the final perturbed equilibrium. Furthermore, this is confirmed in the fortran version of the code -- RDCON includes the large solution and this affects the final perturbed equilibrium.
logan-nc commentedon Oct 27, 2026
Not a PR tonight, for two reasons. The elimination is now explicit and centralised: #489 removes the resonant solution at ideal crossings through
apply_gaussian_reduction!with the resonant rows, and #520 is rewriting the trigger around it, so akeep_large_solutionswitch would be a few lines there but would land on code two open PRs are changing. And it is a convention change, not a fix: keeping the large solution changes Δ′ and every perturbed-equilibrium quantity downstream, so it needs the maintainers' decision on the default and a!harness run of its own. Suggest scheduling it right after #520, as an opt-in flag first (Matt's screw-pinch and RDCON evidence above says the effect is real but modest).logan-nc commentedon Oct 29, 2026
@matt-pharr I still think this comes after #520, which is now outside the v2.0 scope and targeting v2.1. This seems like it also belongs in v2.1.
matt-pharr commentedon Oct 29, 2026
Agree