Deep sequence to sequence learning-based prediction of major disruptions in ADITYA tokamak

Major disruptions in tokamak plasmas need to be identified well before their occurrence and appropriately mitigated. Otherwise, it may dump the heat electromagnetic load vessel its surrounding plasma-facing components. A predictor system based on precursor diagnostics help forecasting disruptive events plasma raise alert beforehand take necessary actions prevent major damages inside vacuum vessel. This paper describes a built with few selected diagnostic signals from ADITYA trained time-sequence long short-term memory network predict of disruption 7–20 ms advance an accuracy 89% testing set 36 6 non-disruptive shots. real-time can infer one time-step results under 170 µs Intel Xeon processor running python, suggesting minimal computation cost best suited for control applications.