SOH Estimation of Energy Storage Batteries Based on ICA and Data-driven Fusion Model

The assessment of the State of Health (SOH) plays a crucial role in diagnosing the condition of lithium-ion batteries (LIBs). However, SOH estimation for large-capacity batteries commonly used in energy storage systems remains insufficiently explored, particularly under conditions of high charge-discharge rates and deep discharge depths. This study proposes a novel approach for estimating the SOH of large-capacity batteries by integrating multi-feature extraction with artificial intelligence techniques. Specifically, various health indicator (HI) sets reflecting reconstructed Incremental Capacity (IC) curve characteristics are extracted from the LIB charging curves. Subsequently, an artificial neural network-based method is introduced to fuse these HIs, enabling precise SOH estimation. The proposed methodology was validated through extensive long-term aging experiments on lithium iron phosphate (LFP) batteries. The results demonstrate the significant advantages of the approach, including high estimation accuracy, reliability, and robustness against cell inconsistencies.