Add magdrift_passing so one run reproduces MARS-K's orbit-class drift scope - #94
Closed
krystophny wants to merge 38 commits into
Closed
Add magdrift_passing so one run reproduces MARS-K's orbit-class drift scope#94krystophny wants to merge 38 commits into
krystophny wants to merge 38 commits into
Conversation
The direct GEQDSK path formed the canonical toroidal precession as drift_phi - q*drift_theta with the *global* safety factor. In geoflux coordinates theta is the geometric poloidal angle, so alpha = phi - q*theta is not constant along a field line and the precession derived from it is wrong. This is the mechanism behind NEO-RT #85, where the direct and straight-field backends disagree on Omega_d by 53%, 64% and 88% NRMS at rho_tor = 0.2, 0.6 and 0.8, with the correlation degrading to -0.87 at the outer surface. Along a field line dphi/dtheta = B^phi/B^theta, so the label gradient carries the local pitch, which reduces to q only when theta is already a straight-field-line angle. Use hctrvr(2)/hctrvr(3). Not yet validated against the straight-field backend: the direct path cannot currently be exercised on the TC24 equilibrium because of a second GEQDSK reader defect, so this is committed on a branch as a diagnosis with its fix, not as a closed issue.
…d line PR #91 substitutes the local pitch hctrvr(2)/hctrvr(3) for the global q in the field-line label, but nothing tested that the substituted quantity is the pitch. The oracle is deliberately not the tabulated q: it is a field line integrated in cylindrical coordinates from field_eq alone, dR/dphi = R B_R/B_phi , dZ/dphi = R B_Z/B_phi , advanced until the geometric angle about the magnetic axis completes one circuit, giving the rotational transform as Delta_phi/(2 pi). That shares no code with do_magfie_eqdsk's coordinate transformation, so it is an independent behavioral oracle rather than a restatement of the patch. The two agree to 7e-4 at s = 0.15, 0.35, 0.60. The test also asserts the local pitch genuinely varies with theta. Without that an implementation returning the global q would satisfy the averaged comparison, and the test would not discriminate the change it exists to protect. Scope: this validates the substituted quantity, not the full canonical precession. The straight-field-line angle map also depends on s, so its gradient carries a radial term that the inp_swi=11 branch still omits; see #85. Writing this test surfaced a separate defect, now fixed in libneo: geoflux built both q and the s label from the GEQDSK's stored qpsi record, which on circ.eqdsk disagrees with the rotational transform of the fixture's own field by a factor of about 1.55. That is why test_eqdsk_axis could only afford a factor-of-two bound on the same comparison. This test needs the libneo fix to pass.
…ift scope MARS-K has no passing magnetic-drift code path at all. Matching that scope in NEO-RT previously required two runs -- trapped with magdrift, passing without -- whose per-class outputs were added by hand. That works because NEO-RT serializes Tco, Tctr and Tt separately and the class contributions are additive, but it doubles the runs and puts the composition in the analysis layer where it is easy to get wrong. `magdrift_passing` switches the magnetic drift off for passing orbits only. It gates two places: the passing branch of Om_tB, which now returns exactly zero rather than an interpolated value, and init_canon_freq_passing_spline, which no longer builds a drift spline it will not use. The namelist control is an integer with a negative "follow magdrift" default, so every existing deck keeps its behaviour bit-identically and no deck has to name the new key. Only an explicit 0 or 1 overrides. VALIDATION IS PENDING and this should not merge without it. The oracle is the two-run equivalence above: trapped-with-drift plus passing-without, summed from the separate class columns, must reproduce a single run with magdrift=.true. and magdrift_passing=0. That compares against an existing, independent code path rather than restating the patch. A bespoke unit test was attempted and dropped: driving Om_tB directly needs more orbit-machinery setup than the check is worth, and the two-run comparison is the stronger oracle anyway. First of a planned set of code-mimicry flags, whose purpose is to let NEO-RT be run *as* MARS-K on the MARS field and *as* PENTRC on the GPEC field, so that residual differences are physics rather than configuration.
Member
Author
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
MARS-K has no passing magnetic-drift code path. Matching that scope needed two NEO-RT runs whose per-class columns were summed by hand; this does it in one.
magdrift_passinggates the passing branch ofOm_tB(now exactly zero) andinit_canon_freq_passing_spline. The namelist control is an integer with a negative follow magdrift default, so every existing deck stays bit-identical and no deck must name the new key.Draft: validation pending, do not merge yet. The oracle is the two-run equivalence — trapped-with-drift plus passing-without, summed from the separate
Tco/Tctr/Ttcolumns, must reproduce a single run withmagdrift=.true.andmagdrift_passing=0. That compares against an existing independent code path rather than restating the patch. A bespoke unit test was attempted and dropped: drivingOm_tBdirectly needs more orbit-machinery setup than the check is worth, and the two-run comparison is the stronger oracle.Full suite passes (14/14) with the default, as expected since the default changes nothing.
First of a planned set of code-mimicry flags (collisionality including passing, which #78 currently leaves at R=1, smoothing, spline variants), whose purpose is to let NEO-RT run as MARS-K on the MARS field and as PENTRC on the GPEC field, so residual differences are physics rather than configuration. Context: ITER TC24 NTV benchmark.