Enhancing Disruption Prediction through Bayesian Neural Network in KSTAR

TL;DR

This paper introduces a Bayesian neural network approach for predicting tokamak plasma disruptions, improving uncertainty quantification and reducing false alarms in disruption forecasts for fusion reactors.

Contribution

It applies Bayesian deep probabilistic learning with a Temporal Convolutional Network to enhance disruption prediction accuracy and uncertainty estimation in KSTAR plasma data.

Findings

Effective disruption prediction on KSTAR data

Improved uncertainty quantification over traditional methods

Reduced false alarms in disruption forecasts

Abstract

Disruption in tokamak plasmas, stemming from various instabilities, poses a critical challenge, resulting in detrimental effects on the associated devices. Consequently, the proactive prediction of disruptions to maintain stability emerges as a paramount concern for future fusion reactors. While data-driven methodologies have exhibited notable success in disruption prediction, conventional neural networks within a frequentist approach cannot adequately quantify the uncertainty associated with their predictions, leading to overconfidence. To address this limit, we utilize Bayesian deep probabilistic learning to encompass uncertainty and mitigate false alarms, thereby enhancing the precision of disruption prediction. Leveraging 0D plasma parameters from EFIT and diagnostic data, a Temporal Convolutional Network adept at handling multi-time scale data was utilized. The proposed framework demonstrates proficiency in predicting disruptions, substantiating its effectiveness through successful applications to KSTAR experimental data.