Data-driven approach for the prediction of remaining useful life

Abstract

Based on the data of the life cycle degradation of lithium ion batteries in Maryland University, this paper analyzes the life degradation process of lithium ion batteries and selects the empirical degradation model. A method for the prediction of lithium ion battery based on EKF / KF is proposed. The EKF algorithm is used to estimate the historical data. Then, the remaining useful life of the lithium ion battery is estimated based on the KF algorithm. The validity of the algorithm is verified by the lithium ion battery data of the University of Maryland, and the algorithm is evaluated by the MAE index.