Research on SOH Prediction of Lithium Battery Based on Improved SSA-BP Neural Network

The rapid growth of electric vehicles has led to widespread adoption of lithium batteries in energy systems, valued for their high energy density and extended cycle life. Nonetheless, frequent charge-discharge cycles and environmental conditions contribute to capacity loss, which ultimately reduces vehicle longevity. As a vital metric, the state of health (SOH) of lithium batteries is challenging to measure directly, and traditional methods often lack precision in SOH prediction. This study gathers experimental data on battery aging under various charge-discharge conditions and incorporates NASA's B05 battery data to identify health metrics. Correlation analysis is used to select optimized input variables. A BP neural network model, enhanced by the sparrow search algorithm (SSA), is developed. Furthermore, combining Levy flight and the sine cosine algorithm (SCA) improves prediction accuracy for SOH.