Linear JOREK n-scan of the TORAX ITER hybrid state: first job
This is the first of up to five jobs for case (b), the TORAX ITER hybrid state, on the equilibrium accepted in entry 71. The protocol is case (a)’s (entry 68), with only the case directory and the health gate’s reference flux changed: n = 1, 2, 3, 5, 8, 12, 16, 20, one run per n from round-off amplitude, Δt = 1, the 80 × 128 grid, 32 cores per run. The axisymmetric start passed its health gate.
What happened
Where the modes grew at all, they grew far more slowly than in case (a). No run reached JOREK’s solver abort: all eight ran to the job’s time limit (no GMRES abort: all 8 runs reached the job time limit at t ≈ 4790 (about 4770 steps); final energies 1e-24 to 1e-20, below the 1e-15..1e-7 fit band, so no γ is fittedC-236). Their energies stayed far below the band where case (a)’s growth rates were fitted, so no growth rate γ is fitted or quoted yet. The table gives local rates late in the job:
| n | late local rate of change of ½ ln E, in units of 1/√(μ₀ρ₀) |
|---|---|
| 1–16 | at t ≈ 4400: n = 1–8 −0.0003 to −0.0001 (decaying), n = 12 +0.00029, n = 16 +0.00173 (rising)C-238 |
| 20 | γ√(μ₀ρ₀) ≈ 0.0029, steady over t 4000–4771; flow perturbation peaks at psi_N 0.95–1C-237 |
Only n = 20 grows cleanly so far, at a rate about 15 times below case (a)’s pedestal modes. Case (b) is the lower-pressure state (66.6 keV in case (a) (EQDSK pressure split by TORAX density) against 14.4 keV in case (b) (TORAX at 80 s)C-239), which is consistent with slower growth, but we have not separated the causes.
The location rule needed a change. In case (a) a mode’s location was read from the peak of its temperature perturbation. In case (b) every n keeps a broad temperature perturbation from the start of the run, peaked at the magnetic axis, so that rule reads “axis” for every n. The flow perturbation does not carry this start-up component. For n = 20 it sits at ψ_N 0.95–1, in the pedestal. The analysis script now also reports the location from the flow perturbation. Case (a)’s classes are unchanged under it.
These rates come from the same linear model as case (a), on a fixed boundary at the separatrix (entry 67), which stabilises peeling modes artificially. They are not yet checked for time step or grid.
Next
The second job continues every n from its latest saved time slice with unchanged inputs. It is running now. JOREK’s time slices are full restart files: a restart of n = 20 on a login node resumed at the right time and step. The grid check is harder than planned. These modes emerge from round-off only late in the run, and repeating that on the doubled grid from scratch would take many more jobs than remain. The options go to the owner once the second job shows which pedestal modes grow cleanly.
Technical details → JOREK. Decision → OQ-41.
JOREK acknowledgement. JOREK is developed by the JOREK team and is open source under the GNU LGPL v3. Results here use JOREK at commit 01a52be3. Citations: M. Hoelzl et al., Nuclear Fusion 61, 065001 (2021); M. Hoelzl et al., Nuclear Fusion 66, 116006 (2026). For a list of JOREK team members, refer to the author list of M. Hoelzl, G.T.A. Huijsmans, F.J. Artola, E. Nardon, M. Becoulet, N. Schwarz, A. Cathey, S.J.P. Pamela et al., Nuclear Fusion 64, 112016 (2024).
Technical details → JOREK › case b n scan
Decisions → OQ-41