Application of the interpretable disruption prediction model to the J-TEXT high-density disruption experiment

Nuclear fusion energy generated by magnetic confinement reactors plays a central role in preventing global energy insecurity and atmospheric pollution. Disruption as a plasma instability can terminate the fusion reaction and damage key components. Predicting and mitigating disruption represents one of the pressing challenges for tokamak devices. Moreover, the disruption predictor can be used for more than just predicting disruption owing to the learned disruption physics. In our previous research, an interpretable disruption predictor(IDP-PGFE) was developed to predict disruption and brought a new sight for disruption experiments under RMPs. Electron Cyclotron Resonance Heating (ECRH) can affect the density limit in experiments conducted on J-TEXT. Distinguishing the causes of disruption in high-density experiments under ECRH helps the understanding of physics mechanisms. In this paper, the application of IDP-PGFE in high-density disruption experiments under ECRH has been introduced. Benefiting from strong interpretability, IDP-PGFE can help researchers accurately identify various types of disruptions, including typical density limit disruptions, radiation-enhanced disruptions, and those driven by increased Magnetohydrodynamics (MHD) instabilities. This approach provides valuable insights into the mechanisms underlying disruption events and contributes to advancing disruption prediction capabilities in fusion research.