The SPARC disruption modeling papers are the five publications from 2024 to 2026 that carry out the self-consistent disruption modelling 📝Ryan Sweeney's 2020 physics-basis paper described as still ongoing — a promise answered six years later.
"MHD stability and disruptions in the SPARC tokamak," one of the seven papers in 📝The SPARC Physics Basis (2020), fixed the loads 📝SPARC must survive and then said plainly that the hardest question was open. It reported GO+CODE simulations giving high 📝flattop-to-runaway current conversion in the absence of seed losses against NIMROD modelling predicting losses near 80 percent, and concluded: "self-consistent modelling is ongoing." Its structural numbers are the target the later work is written against — predicted unmitigated axisymmetric 📝halo current forces up to 50 meganewtons, with comparably large 📝eddy current loads forced to flow poloidally in the 📝vacuum vessel. Those are projected worst-case design loads for a machine engineered to withstand them, not measurements and not the expected outcome of a mitigated 📝disruption.
Three of the five papers model the two mitigation systems. Izzo and co-authors put both 📝massive gas injection and the 📝runaway electron mitigation coil through the 3D 📝magnetohydrodynamics code NIMROD (10.1063/5.0254080), predicting radiation fractions above 95 percent and toroidal peaking below 1.5 during the thermal quench for both the six-valve and four-valve layouts, and finding that the coil holds the field stochastic only while the on-axis 📝safety factor stays below 2. Kleiner and co-authors ran extended-MHD simulations in M3D-C1 with the conducting vessel structures meshed and narrow, realistic 📝deuterium-neon plumes (10.1088/1741-4326/ad9ec4), finding six injectors distribute radiation more evenly than two. Ekmark and co-authors used the DREAM framework with multi-objective Bayesian optimisation over injected densities (10.1017/s0022377825000455): during 📝D-D operation runaway generation can be avoided even counting Compton scattering, while under 📝D-T acceptable runaway current is reached only at the highest achievable deuterium densities, and neon outperforms helium or argon.
The remaining two ask what happens when mitigation fails. Datta and co-authors coupled a fluid 📝runaway electrons model self-consistently to the bulk MHD in M3D-C1 (10.1063/5.0272430), producing the first systematic benchmark of primary runaway sources — including activated 📝tritium beta decay and Compton scattering — and reproducing runaway plateau formation in unmitigated disruptions. Rizzi and co-authors then computed the damage: the first systematic thermal analysis of the 📝tungsten 📝plasma-facing components on SPARC's outboard off-midplane limiters under runaway beams formed during 📝vertical displacement events, reporting melt depth and vaporisation losses (10.1088/1361-6587/ae7d51). Every figure in all five papers is a code output. SPARC has not disrupted, because it has not run.
Our 2020 physics basis said the self-consistent disruption modelling was ongoing. These five papers are that modelling — and we published the unmitigated cases alongside the mitigated ones, runaway plateau and melt depth included, because a mitigation strategy is only worth reading next to the damage it exists to prevent.