Failure-Mechanism Transferability of Cumulative-Damage Features for Health State Estimation of SiC Power Modules

2026-08-08

AI Summary

Existing data-driven approaches struggle to generalize across different failure mechanisms in state-of-health estimation for silicon carbide power modules. To address this limitation, this work proposes a physics-informed neural ordinary differential equation (NODE) model that integrates cumulative electrothermal features, specifically targeting two representative failure modes: solder layer fatigue and bond wire lift-off. The model’s transferability is rigorously evaluated through multi-mechanism accelerated aging experiments and k-fold cross-validation. Results demonstrate that the choice of input feature representation influences transfer performance as significantly as model architecture. The proposed approach, leveraging cumulative damage-based features, maintains consistent accuracy across both failure mechanisms, with error variations confined within inter-fold variance bounds, substantially outperforming conventional methods and NODE variants using baseline features.

Abstract

Data-driven health-state estimators for SiC (Silica-Carbide) power modules typically report their performance on a single accelerated-aging campaign, and how that performance transfers to a different failure mechanism is rarely tested. We benchmark five reference methods from the prognostics and condition-monitoring literature against a physics-informed NODE (Neural Ordinary Differential Equation) on two SiC power-cycling campaigns driven by structurally different failure mechanisms, solder-layer fatigue and wire-bond lift-off, under a per-module $k$-fold protocol. The NODE is evaluated under two input regimes that share the rest of the pipeline: the baseline electrical precursors and a set of cumulative thermoelectric features. Every reference method degrades on the wire-bond campaign, with average errors growing and precision decreasing with respect to their performance on the soldered campaign. The NODE fed with the cumulative features keeps its soldered-campaign metrics on both mechanisms, with differences inside the fold-to-fold variance, while the same architecture fed with the baseline precursors falls back to the reference-method cluster. The input representation contributes at least as much as the architecture to failure-mechanism transferability of a health-state estimator.

Problem

• health state estimation
• failure mechanism transferability
• SiC power modules
• cumulative-damage features
• prognostics

Innovation

• failure-mechanism transferability
• cumulative-damage features
• Neural ODE
• health state estimation
• SiC power modules