Disruption Event Characterization and Forecasting in Tokamaks*

The Disruption Event Characterization and Forecasting (DECAF) code, being developed to meet the challenging goal of high reliability disruption prediction in tokamaks, automates data analysis to determine chains of events that lead to disruptions and to forecast their evolution. The relative timing of magnetohydrodynamic modes and other events including plasma vertical displacement, loss of boundary control, proximity to density limits, reduction of safety factor, and mismatch of the measured and desired plasma current are considered. NSTX/-U databases are examined with analysis expanding to DIII-D, KSTAR, and TCV. Characterization of tearing modes has determined mode bifurcation frequency and locking points. In an NSTX database exhibiting unstable resistive wall modes (RWM), the RWM event and loss of boundary control event were found in 100{\%}, and the vertical displacement event in over 90{\%} of cases. A reduced kinetic RWM stability physics model [1] is evaluated to determine the proximity of discharges to marginal stability. The model shows high success as a disruption predictor (greater than 85{\%}) with relatively low false positive rate. [1] J.W. Berkery, et al., Phys. Plasmas \textbf{24} (2017) 506103. $^{\mathrm{\ast }}$Supported by US DOE Contracts DE-FG02-99ER54524, DE-AC02-09CH11466, and DE-SC0016614.