Machine learning control for disruption and tearing mode avoidance

Real-time feedback control based on machine learning algorithms (MLA) was successfully developed and tested on DIII-D plasmas to avoid tearing modes and disruptions while maximizing the plasma performance, which is measured by normalized plasma beta. The control uses MLAs that were trained with ensemble learning methods using only the data available to the real-time Plasma Control System (PCS) from several thousand DIII-D discharges. A “tearability” metric that quantifies the likelihood of the onset of 2/1 tearing modes in a given time window, and a “disruptivity” metric that quantifies the likelihood of the onset of plasma disruptions were first tested off-line and then implemented on the PCS. A real-time control system based on these MLAs was successfully tested on DIII-D discharges, using feedback algorithms to maximize βN while avoiding tearing modes and to dynamically adjust ramp down to avoid high-current disruptions in ramp down.