Remaining useful life prediction of lithium-ion batteries based on performance degradation mechanism analysis and improved Deep Extreme Learning Machine model

The remaining useful life (RUL) of lithium-ion batteries is a decisive factor in the stability of electric vehicle systems. Aiming at the problem of limited robustness of Deep Extreme Learning Machine (DELM) in lithium-ion battery RUL prediction, an improved whale optimization algorithm (IWOA) is proposed to improve the prediction ability of DELM. Four health features are extracted from the battery aging data, the outliers in the feature data are detected and removed using Hampel filtering, and the health features are dimensionality reduced using principal component analysis to avoid data overfitting. Then, chaotic tent mapping, positive cosine algorithm, and chaotic adaptive inertia weights are used to improve the whale optimization algorithm and increase the search diversity. The introduction of IWOA to optimize the parameter selection of the DELM model effectively solves the problems of low efficiency and poor stability of parameter selection. The method was fully validated using the cycle battery dataset and the prediction results were compared with the conventional method. The results show that the IWOA-DELM method has small prediction errors, strong state tracking fitting ability, good generalization ability, and robustness.