Adapting neural operators for mechanics decisions under changing operating conditions

2026-09-27

AI Summary

This study addresses the accuracy degradation and insufficient decision reliability of neural operators when extrapolating beyond their training conditions. We propose an adaptive updating framework leveraging high-fidelity finite element data, which periodically fine-tunes the neural operator using online-acquired loading paths and integrates an empirical error estimator to optimize mechanical command selection. Experiments demonstrate that minimal data suffices to substantially restore predictive accuracy and improve command decisions. Furthermore, this work reveals a discrepancy between enhanced forward prediction accuracy and decision confidence, confirming that prediction errors must be independently evaluated to ensure overall system reliability.

Abstract

Neural operators can accelerate repeated nonlinear mechanics calculations, but their accuracy can deteriorate as operating conditions move beyond the training range. This work studies whether high-fidelity solutions acquired during use can be reused to adapt a neural operator and improve subsequent mechanics-based command selection. Two hard-magnetic soft-material systems are simulated using high-fidelity finite-element (FE) models, providing reference solutions for evaluating surrogate predictions and selected commands. A neural operator predicts deformation from known material, loading, and magnetic-field inputs, while an empirical error estimator determines which predictions may be used for command selection. Selected FE evaluations supplement these predictions, and their complete loading paths are retained for periodic updates of the neural operator and estimator. In both examples, the fixed operator loses substantial accuracy when stiffness and loading move outside the training range. Updates using 16 acquired paths recover much of the lost accuracy while preserving accuracy in the nominal regime. Under the same FE evaluation budget, the updated operators also improve command selection, although the benefit varies with the operating condition. Error estimation is less consistent, with inaccurate predictions sometimes accepted and accurate predictions rejected. These results demonstrate that reusing high-fidelity loading paths can extend the useful operating range of a neural operator. However, improved forward accuracy alone does not guarantee reliable prediction acceptance, highlighting prediction-specific error assessment as a separate requirement for trustworthy decision making.

Problem

• neural operator
• domain adaptation
• mechanics-based decision making
• error estimation
• soft materials

Innovation

• Neural operators
• Adaptive updating
• Error estimation
• Finite-element simulation
• Mechanics-based decision making