High-Accuracy Disruption Event Characterization and Forecasting for database analysis and real-time application on KSTAR*

Disruption prediction and avoidance is critical for ITER and reactor-scale tokamaks to maintain steady plasma operation and to avoid damage to device components. Physics-based disruption event characterization and forecasting (DECAF) research determines the physical and technical events leading to disruption and can provide event onset forecasts with high accuracy and early warning for disruption avoidance [1]. Real-time application of DECAF on the KSTAR tokamak to over 50 experimental plasmas subject to disruption by locking MHD instabilities, and which produced a nearly equal number of disrupted / non-disrupted cases, were forecast with 100% accuracy. These real-time forecasts triggered controlled plasma shutdown, disruption mitigation, and disruption avoidance actuators. The warnings were issued well before (0.5s – 1.5s) the expected plasma disruption time and early warning guidance given for ITER disruption mitigation. High accuracy exceeding 99% was also found in DECAF analysis of tokamak databases. This fully automated analysis now expands to examine the plasma state as a general dynamical system to best validate DECAF physical events and event chains, thereby allowing reliable extrapolation of models across devices and to future machines.

[1] S.A. Sabbagh, et al., Phys. Plasmas 30 (2023) 032506; https://doi.org/10.1063/5.0133825