Mechanical information enhanced battery state-of-health estimation

Abstract

Accurate estimation of the state of health (SOH) is crucial for the safe operation of batteries. Mechanical features, in particular, offer significant potential for improving SOH estimation by directly reflecting key internal processes within batteries. However, research on the contribution of mechanical features to SOH estimation remains limited. This study demonstrates the effectiveness of mechanical features for SOH estimation in pouch cells under various operating conditions and scenarios. The results show that mechanical features provide reliable SOH estimates across different temperatures, C-rates, and charging profiles, and they are especially robust under real-world driving conditions. The mechanical features typically achieve at least a 28.26% reduction in prediction error. Notably, in the driving scenario, the mean absolute percentage error reaches an impressive low of 0.65%. Furthermore, this work introduces an evaluation framework to systematically benchmark features derived from electrical, thermal, and mechanical signals based on their overall predictive capabilities. Finally, detailed physical interpretations are provided to explain the effectiveness of mechanical features.