State-of-Charge Estimation of Lithium-Rich Manganese-Based Batteries Based on WOA LSTM and Extended Kalman Filter

Abstract

Recent years, electric vehicles gradually become popular, but their cruising range has become one of the main problems that plague car companies and users. The lithium-rich manganese-based cathode material batteries with higher energy density stand out. The state of charge is an important parameter. This paper selects a 19Ah lithium-rich manganese-based cathode material battery for research, using extended Kalman filter based on second-order Equivalent circuit model estimate its state of charge. However, the impedance spectrum of lithium-rich manganese battery is different from that of 18650 lithium-ion battery, and the second-order equivalent circuit model will have errors, resulting in the low accuracy of SOC estimation. In order to solve this problem, this paper proposes two schemes: EKF-LSTM and LSTM-EKF. The whale optimization algorithm (WOA) is used to select the preset parameters. The results show that the LSTM-EKF method has the highest estimation accuracy, with a maximum error of 1.46%.