Answers are grounded in this paper and cite their sources.
In this paper, we propose an artificial intelligence (AI)–driven framework for predicting fission gas release (FGR) in liquid metal–cooled reactors. Three AI models [transformer, mamba, and random forest (RF)] are trained on a comprehensive FGR database comprised of 10 065 data points generated from multiphysics simulations. The results revealed that RF achieved superior prediction accuracy, with a mean absolute percentage error of 0.63% for the optimal parameter combination (linear power, fuel grain size, fuel porosity, fuel burnup, and time). While RF excels in database predictions, the transformer model demonstrated exceptional capability in capturing parameter-effect trends, such as the positive correlation between FGR and fuel linear power or time. This study underscores the potential of AI models, particularly RF and transformer, in advancing FGR forecasting for reactor safety analysis.
No takes yet. Share an insight, caveat, or question.
Li et al. (2026) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: