Modeling And SOC Estimation Of Lithium Battery For Electric Vehicle

In recent years,energy crisis and environmental pollution have become more and more serious.Energy saving and environmental protection have become the theme of the development of all countries.Because of the safety,cleanness and efficiency of energy conversion,new energy electric vehicles have become the development hotspot of automobile industry.Battery management system（BMS）,as the core component of electric vehicle,whether it can run stably,efficiently and safely,has become one of the important indexes to evaluate the performance of electric vehicle.State of charge（SOC）estimation is an important functional module of battery management system,which is the premise of real-time monitoring,control strategy and safety guarantee of battery management system.So it is very important to study the accurate estimation of battery power in BMS.In the current engineering application,the ampere hour integration method and the open circuit voltage method are commonly used SOC estimation methods,which are simple in principle,fast in calculation,but prone to error accumulation and estimation delay.In order to improve the accuracy,model-based SOC estimation methods have gradually become the mainstream,such as particle filter,Kalman filter and its derivative algorithm.In addition,SOC estimation is also affected by battery charge discharge ratio,temperature,aging and other factors.In this paper,18650 lithium battery for electric vehicle is taken as the research object to explore the change of electric quantity in the process of charging and discharging.Firstly,a simplified electrochemical model based on P2 D model is established to describe the battery and its external characteristics.Using the idea of single particle and the polynomial approximation method,the P2 D model is simplified in stages,and the simplified process is supplemented and the approximate results are verified by numerical value,and the extended single particle model is obtained.Secondly,the characteristics of the selected lithium-ion battery are tested and analyzed.The lithium-ion battery test was carried out by Shenzhen Newell cell test system and the data were collected.By analyzing and processing the experimental data,the voltage characteristics,multiple discharge characteristics,capacity characteristics and temperature characteristics of lithium-ion battery are obtained,which provide the basis for parameter identification of battery model and power state estimation.Thirdly,based on the characteristics of the model parameters,genetic algorithm is used to identify the parameters.At the same time,considering the changes of model parameters in the whole process of battery charging and discharging,the model can be identified in different SOC stages to improve the accuracy of the model.Finally,based on the results of extended single particle model and parameter identification,extended Kalman filtering（EKF）and unscented Kalman filtering（UKF）are used to estimate SOC,and the results are compared to analyze the advantages and disadvantages of the algorithm.On the basis of Unscented Kalman filter,strong tracking filter and adaptive filter are added to improve the estimation accuracy.The accuracy and robustness of the proposed estimation method are verified by SOC Test under multiple operating conditions.