Model Identification And SOC Estimation Of LiFePO4Batteries In Hybrid Electric Vehicles

As the energy and environmental issues become serious increasingly, hybrid electric vehicle is highly valued by the international community for the advantage of zero emissions and low noise. One of the key technology of the electric vehicle industrialization is the design of BMS, which needs accurate SOC estimation of battery. All of these plays an important role on the improving of the life of the battery and the performance of the vehicle. Nowadays, LiFePO4battery has became more and more popular because of the outstanding performance in high energy rate.According to the nonlinear characteristic of battery and the special difference of LiFePO4battery, this paper improved PNGV equivalent circuit model of battery, identified the model parameter based on the Least Square method, introduced two algorithm of SOC estimation based on EKF and AUKF, which are derivative algorithm of KF, and compared the error of the two algorithm at last Because of the disadvantage of EKF and AUKF is existed,such as heavily computation, long computational time and uncertain stability,LPV subspace identification method is proposed to implement the model identification of battery and the SOC estimation, which provided new thought on the issue of control-oriented battery model and SOC estimation.