An Optimized Random Forest Regression Model for Li-Ion Battery Prognostics and Health Management

This study proposes an optimized random forest regression model to achieve online battery prognostics and health management. To estimate the state of (SOH), two aging features (AFs) are extracted based on incremental capacity curve (ICC) quantify degradation, further analyzed through Pearson’s correlation coefficient. predict remaining useful life (RUL), AFs extrapolated degradation trends closed-loop least square method. capture underlying relationship between capacity, a is developed; meanwhile, hyperparameters determined using Bayesian optimization (BO) enhance learning generalization ability. The method co-simulation MATLAB LabVIEW introduced develop management system (BMS) for verification proposed Based open-access datasets, results mean error estimated SOH 1.8152% predicted RUL 32 cycles, which better than some common methods.