Research On Estimation Of Battery SOC Based On Fractional Order Theory

With the decrease of non-renewable energy and the turbulence of global energy situation,as well as people's increasing attention to environmental protection,electric vehicles have become an important development direction of automobile industry.Lithium ion power battery is the main energy source of electric vehicle,which has a good development prospect.The accuracy of estimation for state of charge(SOC)directly determines the battery management system and vehicle control strategy,and then affects the performance of the vehicle.Therefore,accurate estimation of battery SOC has important theoretical value and engineering application significance for the development of electric vehicles.This topic comes from the key research and development project(industry prospect and key technology)in Jiangsu Province.This paper takes the LiFePO4 battery as the research object,conducts the characteristic test and identification test for parameters of equivalent circuit model under different influence factors,and uses the fractional extended Kalman filter to carry out the real-time and accurate estimation of battery SOC.The main work of this paper is as follows:(1)In this paper,the first-order fractional equivalent circuit model of power battery was established by taking the lithium iron phosphate battery as the research object and replacing CPE fractional element as the capacitor element.The discrete state expression of the equivalent circuit model was deduced by using G-L fractional calculus definition.(2)The battery charging-discharging test platform was built.At ambient temperatures for 0?,5?,15?,25?,30?,35?,the capacity of LiFePO4 battery with different charging and discharging ratio was tested respectively.In order to estimate the SOC value accurately,the concept of conversion coefficient of battery charging and discharging capacity was put forward in this paper.The curve of conversion coefficient of battery charging and discharging capacity at different temperatures was calculated and drawn based on the battery capacity in 30? ambient temperature and C/3 charging and discharging ratio.(3)At seven selected ambient temperatures,and the battery SOC was 1,0.95,0.9,…,0.1,0.05,0,the open circuit voltage test in charging and discharging state was carried out respectively.In order to adapt to the engineering application,the open circuit voltage value of the battery in charging and discharging state was simplified and unified on the premise of considering the battery hysteresis characteristics.The open circuit voltage model of the battery was established by fitting with the 8-order polynomial.(4)At seven selected ambient temperatures,and the battery SOC was 0.9,0.8,0.7,0.6,0.5,0.4,0.3,0.2,0.1,the double pulse method was used to identify the model parameters in the battery charging and discharging state respectively.Based on the battery terminals voltage data which was composed of 108 double pulses obtained from the identification test,the parameters of the first-order fractional equivalent circuit model were identified by using the particle swarm optimization algorithm with dynamic inertia weight.Under 26 consecutive DST conditions,compared with the first-order integral order equivalent circuit model,the results show that the first-order fractional equivalent circuit model were more accurate and can simulate the battery dynamic characteristics better.(5)Using Jacobian matrix linearization method,the recursive formula of fractional order extended Kalman filter was derived,and battery SOC was estimated in real time under the condition of continuous DST.By comparing with the SOC estimation results of integer order extended Kalman filter,it showed that the SOC estimated from fractional order extended Kalman filter was more accurate,the average error of the SOC estimation value was 0.0036,the relative error was 0.52%,and the accuracy was more improved than that in integer order extended Kalman filter.Finally,the battery SOC was estimated under different initial errors.The algorithm was robust because the results showed that the algorithm can reduce the SOC estimation error gradually and return to the accurate value under different initial errors.In this paper,a method of SOC estimation using extended Kalman filter based on the fractional order theory was proposed,which has a certain engineering application significance and a certain reference value for other researchers to study the battery SOC estimation method.