Estimation of battery SOC based on improved EKF algorithm

摘要

This paper studies the estimation of the state of lithium battery (SOC), and develops an improved extended Kalman filter algorithm for this problem. To compensate deficiencies of the simple polynomial fitting, the neural network algorithm firstly is adopted to simulate the relation curve between the SOC and the parameters of circuit model, which is constructed based on Thevenin circuit. And then the state space equation among the battery's SOC and the voltage of the ends of the RC loop is established, also does the measurement equation which is based on the battery output voltage. In addition, extended Kalman is applied to estimate battery SOC. In the last, the effectiveness of the proposed method is verified using an experimental testing, and the results show that our method can estimate the SOC more accurately comparing with the standard extended Kalman algorithm.