Research in support of the SPARC Off-Normal Warning System

SPARC disruption prevention strategies will enable the accomplishment of Q>1 during its first campaign. The off-normal warning (ONW) system for asynchronous detection of disruptive instabilities under development at CFS is based on an extended physics-driven stability model [1]. Alcator C-Mod data, retrieved and validated via DisruptionPy [2], is used to test and validate radiative collapse and vertical displacement observers inside the ONW framework. Ongoing research informs potential improvements to the baseline SPARC ONW system. A novel control-oriented performance metric to monitor and minimize accumulated damage to the machine [3] has been explored. Additionally, a generic ML-based disruption predictor is found to achieve 100% of correct classification with 6% of false positives when analyzing the C-Mod ramp-up phase. Being a high-energy-density device, SPARC will require robust prediction of thermal collapses and consequent energy release [4] which risk melting the divertor. First results from TCV suggest that NTMs, loss of vertical control, and edge cooling are the most common precursors of melt risk events. The characterization of the disruptive precursors was conducted via DEFUSE [5], currently being ported to C-Mod for further explorations.

[1] Gerhardt et al 2013 Nucl.Fusion 53 063021

[2] Trevisan et al 2024 Zenodo 10.5281/zenodo.13935223

[3] Saperstein et 2025 al Nucl Fusion under review

[4] Sweeney et al J. Plasma Phys. (2020), vol. 86, 865860507

[5] Pau et al, 2023 IAEA FEC, IAEA-CN-316-2057