Research on Battery Health Status Estimation and Remaining Life Prediction Based on a Novel Inception-BiLSTM Model

To address the issues such as the single extraction of health features and the low estimation accuracy in the process of estimating the health status and predicting the remaining service life of lithium batteries, this paper proposes a novel battery health status estimation and remaining life prediction method based on InceptionBiLSTM model. By analyzing the state during the charging process of lithium batteries, nine health factors are extracted. The strongly correlated health factors are selected through the Pearson correlation coefficient and are used as the input of the model. The initial position of the dung beetle is generated using the chaotic initialization tent mapping, and the position of the thief dung beetle is optimized using the sine and cosine strategy, which solves the local convergence problem caused by the initialization of the DBO (Dung Beetle Optimizer) algorithm and optimizes the balance of the DBO algorithm, thereby improving the stability of the prediction. Practical trials manifest that, against the backdrop of current sophisticated prognostic models, the fresh design yields more accurate residual-life outlooks for lithium battery packs.