A hybrid pulse power characterization-elastic net framework for accurate state-of-health estimation in lithium-ion batteries under thermal aging conditions

This study presents a novel hybrid framework for accurate State-of-Health (SoH) estimation of lithium-ion batteries, critical for enhancing safety, reliability, and lifespan in electric vehicles and energy storage systems. The approach integrates an enhanced Hybrid Pulse Power Characterization (HPPC) protocol with a thermal degradation model utilizing Elastic Net regression. By extracting key electrochemical features—ohmic resistance (R 0 ) and polarization time constants (τ 1 , τ 2 )—across varying state-of-charge (SOC) levels and temperatures, the model minimizes cell stress while providing precise degradation indicators. Experimental validation on 18,650 LiFePO 4 cells over 400 cycles achieves an SoH prediction error of just 0.027 %. The framework demonstrates robustness in real-time applications, with minimal fitting errors and a strong correlation between the thermal indicator and SoH. This work introduces an effective, scalable approach for battery management systems, advancing predictive maintenance and condition monitoring. • Captures key parameters (R₀, τ₁, τ₂) across 20–90% SOC and 15–45 °C with minimal cell stress. • Achieves an RMSE of 0.027% and maintains SoH tracking within a ±0.02% error margin. • Ohmic resistance (R₀) increases by 22.8%, while τ₂ decreases by 17.4% over 400 cycles. • Normalized thermal indicator (Trel) shows strong correlation with SoH decline. • Suitable for predictive maintenance in electric vehicles and energy storage systems.