A State of Health Estimation Method of Lithium-Ion Batteries Based on Improved Gray Wolf and SVM Algorithm

Electrochemical energy storage serves as a foundational technology in contemporary electrical energy storage systems, with its operational safety and stability being crucial to socio-economic development. The estimation of the state of health (SOH) of energy storage batteries is an essential component for ensuring system safety warnings and lifecycle management. To address the challenges of redundant health feature dimensions, insufficient correlation of influencing factors, and limited prediction accuracy in existing SOH estimation methods, in this paper, a novel state of health estimation framework is introduced, leveraging an Improved Gray Wolf Optimization (IGWO) algorithm to optimize the parameters of a Support Vector Machine (SVM). This model achieves precise prediction of battery health states by extracting multidimensional health features, including the differential temperature, incremental capacity, time interval of equal charge voltage difference (DT-IC-TIECVD) and implementing the improved gray wolf optimization algorithm with support vector machine algorithm (IGWO-SVM). Validated on the Oxford battery aging dataset, the proposed model achieves mean absolute error (MAE), root mean squared error (RMSE), and coefficient of determination (R2) values of 0.43%, 0.55%, and 0.99, respectively. These results confirm the high accuracy and feasibility of the proposed method, while also providing a novel technical pathway for the health management of energy storage batteries.