Cycle-Consistent and Uncertainty-Aware Neural Surrogates for Tokamak Edge Plasmas

2026-07-23

AI Summary

This work addresses the high computational cost of traditional tokamak edge plasma simulations, such as SOLPS-ITER, which hinders parameter scans and real-time control, while existing machine learning surrogates lack inverse inference capability and reliability assessment. The authors propose a cycle-consistent, uncertainty-aware neural surrogate that integrates a conditional U-Net for forward modeling with an optimization-based inverse solver. For the first time in edge plasma modeling, they introduce cycle consistency regularization and uncertainty quantification, enabling self-assessment of inference quality without ground-truth labels and achieving high-fidelity parameter inversion. Trained end-to-end on SOLPS-ITER grids, the model attains forward prediction normalized root-mean-square errors below 2.6% (correlation coefficient >0.95), improves cycle R² from 0.59 to 0.99 via cycle consistency, and achieves inversion correlation coefficients ≥0.97 across five control parameters, with inference speeds five to six orders of magnitude faster than SOLPS-ITER.

Abstract

The boundary and divertor plasma govern how a tokamak exhausts power and particles, setting heat fluxes, target conditions, and the onset of detachment. Predicting these quantities is essential for operating current and future devices, but edge simulations that resolve them are too slow for parameter scans, optimization, or real-time control. Machine-learning surrogates offer a fast alternative, yet most are forward-only: they cannot recover input parameters from observations or assess the reliability of their predictions. We introduce a cycle-consistent neural surrogate for edge plasmas, combining a conditional U-Net forward model with an optimization-based inverse method built on the frozen forward network. The forward model maps five control parameters to two-dimensional plasma-state fields on the SOLPS-ITER mesh; the inverse method enforces consistency between forward and inverse predictions, a self-supervised quality check needing no ground-truth labels at inference. An ensemble of multilayer perceptrons also predicts electron temperature and density profiles at the outboard midplane and divertor targets, with uncertainty estimates that flag where more simulations are needed. The forward model achieves normalized root-mean-square errors below 2.6% and Pearson correlations above 0.95 for all fields. Cycle-consistency regularization raises the average cyclical $R^2$ from 0.59 to 0.99 without degrading forward accuracy and enables recovery of the core fueling rate; all five control parameters are recovered with Pearson $r\ge0.97$. A $k$-d tree warm start yields a database completion rate above 95%, versus roughly 30% outright failures when cold-started. With about $4\times10^6$ parameters, the model produces full 2D predictions in milliseconds, five to six orders of magnitude faster than SOLPS-ITER, enabling real-time control, parameter scans, uncertainty analysis, and digital twins.

Problem

• tokamak edge plasmas
• neural surrogates
• inverse problem
• uncertainty quantification
• real-time control

Innovation

• cycle-consistent surrogate
• uncertainty-aware neural network
• inverse plasma modeling
• real-time tokamak control
• SOLPS-ITER acceleration