关键词: 车辆工程, 锂电池SOC, 扩展卡尔曼滤波算法, 自适应BP神经网络算法, 联合估计

Abstract: Accurate estimation of battery SOC is a prerequisite and important basis for the rational implementation of battery management in electric vehicles. Aiming at the problem that the estimation accuracy of power battery SOC for electric vehicles is improving continuously, the joint estimation method was used to study the SOC of lithium battery. Based on the Thevenin battery model and the modified ampere-hour integration algorithm, the output equation and state space model of the lithium battery were derived. The parameters of the battery model were identified by collecting relevant data during the experiment and applying the recursive least square method. The principles of the extended Kalman filter (EKF) algorithm and the adaptive BP neural network algorithm were analyzed, and two algorithms were combined. On this basis, the self-adaptive BP-EKF algorithm (ABP-EKF) was proposed. The proposed algorithm was used to jointly estimate the lithium-ion battery SOC. Finally, the data of the lithium battery SOC was estimated by comparing the two algorithms of ABP-EKF and EKF. The research results show that compared with the EKF algorithm, the proposed ABP-EKF algorithm reduces the mean error term and root mean square error term by 3.9% and 3.79%, respectively.

Key words: vehicle engineering, lithium battery SOC, extended Kalman filter algorithm, adaptive BP neural network algorithm, joint estimation