State of health estimation for lithium-ion batteries based on incremental capacity analysis and Mamba model optimized by improved whale optimization algorithm

Accurate estimation of the state of health (SOH) for lithium-ion batteries (LIBs) is paramount for battery management systems (BMS) to ensure safe operation and extend the lifespan of applications such as electric vehicles (EVs). However, existing data-driven methods face inherent limitations regarding feature engineering complexity, long-term estimation accuracy, and hyperparameter optimization, restricting their effective application in practical BMS. To address these limitations, the novel SOH estimation method based on the improved whale optimization algorithm (IWOA)-Mamba with six strongly correlated aging features is proposed. The method employs incremental capacity analysis (ICA) to extract six features strongly correlated with and effectively characterizing battery degradation. Subsequently, the Mamba model is constructed using these features to predict SOH. Furthermore, the IWOA, enhanced with four improvement strategies compared to the standard whale optimization algorithm (WOA), is employed to optimize Mamba’s hyperparameters. This approach overcomes the drawbacks of manual hyperparameter tuning and improves the estimation accuracy of the resulting IWOA-Mamba model. The proposed method is validated using two public battery datasets: NASA and CALCE. Evaluation results demonstrate that the IWOA-Mamba model effectively enhances SOH estimation accuracy, with mean absolute error (MAE) and root mean squared error (RMSE) values predominantly remaining below 0.7%.