Research On Battery SOC Estimation Method Based On Stochastic Disturbance

The rapid development of automobile industry has brought great convenience for people travelling, however it also leads to increasing pollution of the environment and makes the energy depletion more serious than ever before, which has become concerning of human beings. Therefore, countries around the world take electric vehicle(EV) as the future direction of auto industry, and accurately estimating the battery charged State(State of charge, SOC) is the key prerequisite for battery management system to ensure safety of EV and improvement of the vehicle performance. But owing to the complexity of power battery electrochemical property and working condition, many existential factors impact the SOC estimation which makes accuracy SOC very difficult to implement. Based on the study of battery model and parameters identification, researches are carried out in this paper on the SOC estimation of lithium-ion battery.This paper firstly introduces the background and significance of the battery SOC estimation and explains the working principle and characteristics of lithium-ion battery in detail. Secondly, lithium-ion battery charging and discharging process and influencing factors of the residual capacity have been studied. Then some common prediction algorithms for SOC estimation have been discussed, especially the Kalman filtering(KF) algorithm. The superiority-inferiority of KF has been seriously analyzed considering the accuracy requirement of EV SOC estimation and project implementation, which paves the way for proposing the new method of SOC estimation which is based on stochastic disturbance.Then, several kinds of commonly used battery equivalent circuit model are presented, and the Thevenin equivalent circuit model is selected finally. Through the parameters identification experiment with matlab simulation toolbox, open voltage curve and relevant parameters of Thevenin equivalent circuit have been gained respectively, as well as the Thevenin model discretization equations for battery status.Finally, extended Kalman filtering(EKF) algorithm is adopted to estimate SOC,and after analysis of the results, the effectiveness of EKF algorithm is verified. This paper proposes a new method of SOC estimation which is based on stochastic disturbance, and carries on simulation experiments under three conditions. Finallyeffects of EKF algorithm and the designed algorithm which is based on stochastic disturbance have been compared. Results show that the designed SOC estimation method has better estimation accuracy than EKF algorithm, when the battery model is difficult to build accurately, it is useful and effective to restrain filtering divergence of the results gained from EKF.