Research on prediction of the remaining useful life of lithium-ion batteries based on particle filtering

Abstract

The particle filtering is used to study the prediction of the remaining useful life (RUL) of lithium-ion batteries, and a simple and effective algorithm fusing the model method and the data-driven method for RUL predicting is proposed. The algorithm uses the fusion of the model method and the data-driven method to modify the double exponential empirical degradation model to reduce the model parameters and the parameter training difficulty, uses the particle filter algorithm to track the battery capacity degradation process, and uses the auto regression model to modify the observation value of the state space equation to improve the prediction accuracy. The experimental results show that the proposed algorithm can effectively predict the remaining useful life of lithium batteries. © 2017, Executive Office of the Journal. All right reserved.