Multifidelity Active Learning for Failure Estimation of TRISO Nuclear Fuel

TL;DR

This paper introduces a multifidelity active learning method to efficiently estimate small failure probabilities of TRISO nuclear fuel, reducing computational costs by leveraging cheaper models while maintaining accuracy.

Contribution

The paper develops a novel multifidelity active learning framework specifically for failure probability estimation in nuclear fuel, combining active learning with multifidelity modeling.

Findings

Efficiently predicts TRISO failure probability with fewer high-fidelity evaluations.

Outperforms traditional Monte Carlo methods in computational efficiency.

Accurately estimates small failure probabilities using combined low- and high-fidelity models.

Abstract

The Tristructural isotropic (TRISO)-coated particle fuel is a robust nuclear fuel proposed to be used for multiple modern nuclear technologies. Therefore, characterizing its safety is vital for the reliable operation of nuclear technologies. However, the TRISO fuel failure probabilities are small and the computational model is time consuming to evaluate them using traditional Monte Carlo-type approaches. In the paper, we present a multifidelity active learning approach to efficiently estimate small failure probabilities given an expensive computational model. Active learning suggests the next best training set for optimal subsequent predictive performance and multifidelity modeling uses cheaper low-fidelity models to approximate the high-fidelity model output. After presenting the multifidelity active learning approach, we apply it to efficiently predict TRISO failure probability and make comparisons to the reference results.