Research On SOC Estimation Technology Of Li-Ion Battery For Electric Vehicle Based On Particle Filter

The car is not only an essential means of transportation for people to travel,but also an indispensable tool for all walks of life.The widespread use of automobiles will bring convenience and speed to people,but it will also bring environmental pollution problems and energy shortages.The use of electric vehicles can reduce environmental pollution and ease resource shortages to some extent.Therefore,the development of electric vehicles has received universal attention from all countries.Lithium-ion batteries are widely used in electric vehicles because they have advantages such as high energy density,long life,and no memory effect compared to other batteries.Lithium-ion batteries need to design a special battery management system(BMS)for protecting the battery and detecting the status of the battery.The state of charge(SOC)of the battery reflects the remaining battery state of the battery and is an important parameter that the BMS needs to estimate.Accurate SOC estimation is of great significance for the stable operation of the electric vehicle.Therefore,the research of high-performance SOC algorithms has become a hot issue in related research institutions.The research carried out in this paper is based on lithium-ion batteries.The appropriate battery model is selected and the parameters of the model are identified.Three SOC estimation algorithms based on particle filtering are designed and the performance of the algorithm is compared and analyzed.The main contents of this paper are as follows:1.Based on the Samsung ICR18650-22 P lithium-ion battery,the battery capacity test experiment,OCV-SOC correspondence experiment,and battery condition test experiment were designed and carried out.2.Analyze and discuss the chemical model and equivalent circuit model of the battery and select the second-order RC equivalent circuit model as the battery model.Use the polynomial fitting method to fit the relationship between the battery's OCV-SOC and the exponential fitting.Methods The R and C parameters in the model were identified offline.3.Aiming at the problem of the proposed density function for the particle filter,the Kalman filtering method was proposed to generate the recommended density function of particle filter.The SOC estimation method of EPF,CPF and UPF was designed.4.Taking the battery data of the City operating conditions measured by the equipment as a reference,the SOC estimation effects of the EKF,EPF,CPF,and UPF algorithms are systematically compared.In order to make the test closer to the real car,colored noise was added to the experimental data to compare the ability of the algorithm to resist noise.The test results show that the estimation effects of EKF,EPF,CPF,and UPF are satisfactory under the condition of no external interference.The EPF,CPF,and UPF algorithms are slightly better than the EKF algorithm.Under the conditions of adding colored noise,the EKF algorithm fails.The effect of EPF algorithm can describe the change of SOC more accurately but the accuracy is lower than that of CPF and UPF.The effect of CPF algorithm and UPF algorithm is similar and the accuracy is the highest.