Plasma State Monitoring and Disruption Characterization using Multimodal VAEs

2025-04-24

AI Summary

Tokamak plasma disruption prediction suffers from limited interpretability and insufficient understanding of underlying physical mechanisms. To address this, we propose a plasma-physics-informed triple-extended variational autoencoder (VAE), the first to jointly achieve: (1) structure-preserving projection of continuous discharge trajectories into latent space; (2) disentangled representation of operational modalities (e.g., density, current, magnetic field); and (3) explicit separation of disruption states in latent space. Leveraging multimodal diagnostic data, we construct statistically grounded, interpretable disruption-rate and “disruptiveness” continuous metrics, and introduce counterfactual attribution to identify critical disruption-triggering parameters. Validated on 1,600 TCV discharges, our model achieves strong correlation (r > 0.85) between predicted disruption risk and actual disruption timing, accurately distinguishes disruption types, and substantially enhances both physical interpretability and engineering utility.

Technology Category

• Computer Vision: Multi-modal Vision
• Reasoning under Uncertainty: Relational Probabilistic Models
• Machine Learning: Multimodal Learning

Application Category

• User Modeling, Personalization and Recommendation: Explainable and interpretable methods for personalization
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking

Abstract

When a plasma disrupts in a tokamak, significant heat and electromagnetic loads are deposited onto the surrounding device components. These forces scale with plasma current and magnetic field strength, making disruptions one of the key challenges for future devices. Unfortunately, disruptions are not fully understood, with many different underlying causes that are difficult to anticipate. Data-driven models have shown success in predicting them, but they only provide limited interpretability. On the other hand, large-scale statistical analyses have been a great asset to understanding disruptive patterns. In this paper, we leverage data-driven methods to find an interpretable representation of the plasma state for disruption characterization. Specifically, we use a latent variable model to represent diagnostic measurements as a low-dimensional, latent representation. We build upon the Variational Autoencoder (VAE) framework, and extend it for (1) continuous projections of plasma trajectories; (2) a multimodal structure to separate operating regimes; and (3) separation with respect to disruptive regimes. Subsequently, we can identify continuous indicators for the disruption rate and the disruptivity based on statistical properties of measurement data. The proposed method is demonstrated using a dataset of approximately 1600 TCV discharges, selecting for flat-top disruptions or regular terminations. We evaluate the method with respect to (1) the identified disruption risk and its correlation with other plasma properties; (2) the ability to distinguish different types of disruptions; and (3) downstream analyses. For the latter, we conduct a demonstrative study on identifying parameters connected to disruptions using counterfactual-like analysis. Overall, the method can adequately identify distinct operating regimes characterized by varying proximity to disruptions in an interpretable manner.

Problem

• Monitoring plasma state and characterizing disruptions in tokamaks.
• Developing interpretable models for plasma disruption prediction.
• Separating operating regimes and identifying disruption risk indicators.

Innovation

• Uses multimodal VAEs for plasma state representation
• Extends VAEs for continuous plasma trajectory projections
• Separates disruptive regimes via latent variable modeling