Data-Driven Soh Estimation Via Voltage Temporal and Capacity Peak Features for Batteries

The State of Health (SOH) is essential for assessing the condition of lithium-ion batteries (LIBs). However, accurately estimating SOH for commercial LIBs, particularly in complex environments and over extended periods, remains a challenge. This study introduces an SOH estimation method tailored for small-capacity batteries by integrating multi-feature extraction with artificial intelligence techniques. Incremental capacity (IC) characteristics and time-varying differential voltage features are extracted from the charging curves of LIBs. Based on these traits, a hybrid health indicator (HI) fusion technique underpinned by artificial neural networks is devised to augment the precision of SOH estimation. The proposed scheme is verified via extensive longterm degradation tests on LCO batteries, manifesting its exactitude, trustworthiness, and robustness against cell disparities.