A Method for Remaining Discharge Energy Prediction of Lithium-Ion Batteries Based on Terminal Voltage Prediction Model

In this paper, a new remaining discharge energy prediction method for dynamic Li-ion batteries based on second-order equivalent circuit model (ECM) is proposed. Assuming future current condition is known, the core of this method lies in building a terminal voltage prediction model, which involves estimation of future State-Of-Charge (corresponded with open circuit voltage) and future model parameters. Extended Kalman Filtering (EKF) method is used for estimation of present SOC and polarization voltage to determine the predicting initial point. Ampere-hour integral method is adopted for future SOC estimation, while future parameters are calibrated through correspondence to SOC. The method ensures estimation accuracy with lower than 5% error even in low SOC interval. Moreover, in order to simplify the algorithm, this method is combined with subtraction method to form a whole remaining discharge energy curve.

I. Introduction

In electric vehicle, limited by energy density of lithium-ion battery pack, driving range is relatively short. Besides, unable to estimate the remaining range accurately, especially in low range interval, causes drivers' “low range anxiety” problem, which further shorten the whole range. Therefore, accurate residual range estimation has long been an essential issue in electric vehicles.