Comparison of Prognostic Algorithms Based on PF and EKF for Estimating Remaining Useful Life of Li-Ion Batteries

Abstract

To estimate remaining useful life (RUL) of Li-ion batteries is a key factor for correct and safe battery management, particularly for the development of Battery Management System (BMS). A lumped parameter model which integrates the non-linear open-circuit voltage, current, temperature, cycle number, and remaining capacity and other dynamic characteristics is created based on the battery electrical characteristics is presented. A particle filter (PF) algorithm which syncretizes Li-ion battery electrochemical working process is proposed according to the sequence importance of re-sampling to predict its discharge end time in single cycle time and cycle life. Besides, for comparison, a extended Kalman filter (EKF) algorithm is also proposed to estimate the RUL based on the same statistics. Simulation results show that the PF algorithm according to lumped parameter model has a better precision in estimating RUL compared with the EKF algorithm.