Research On The Methods Of SOC Estimation And RUL Prediction Of Lithium-ion Battery

Nowadays,more and more serious pollution of the environment have push forward the development of electric vehicles(EVS).China is a country rich in coal and lack of oil and gas.With the increase of domestic car ownership year by year,and the research of electric vehicles has become one of the hot spots in China.In order for EVS to operate safely and efficiently,the battery is the core,and battery management system(BMS)that effectively manages and regulates the battery is essential.At the same time,BMS also plays a key role in promoting the development of EVS.The accurate estimation of Sate of charge(SOC)in BMS can not only prevent overcharge and overdischarge,but also can be used as a low battery current limit threshold and as a vehicle control strategy threshold.Remaining Useful life(RUL)prediction in BMS can provide users with an estimate of battery cycle life,so that they can make decisions such as battery replacement.The above two aspects are essential to the safe operation and reliability of EVS.This paper mainly studies the methods and implementation of SOC estimation and RUL prediction of LIBs.(1)The model is introduced and established.In the process of SOC estimation,the second-order equivalent circuit model is selected by considering the accuracy and complexity of the model.The relationship curve of open circuit voltage and SOC required by the model is obtained by experiments,and the resistance capacitance in the model parameters is obtained by off-line identification;In the process of RUL prediction,the data of the battery capacity attenuation are fitted by different fitting methods,and three aging models for the prediction of RUL are established.The fitting effect of data is displayed directly by adj R~2and RMSE.The results show that the fitting effect of the ensemble model is the best,and the applicability and the global simulation of the ensemble model to different lithium-ion batteries are better than the other two models.(2)The accuracy of SOC estimation can be improved by selecting the appropriate algorithm.In this paper,SOC estimation is estimated by using AUKF.The tracking accuracy,convergence time and robustness of the estimation results are evaluated by using convergence speed and MAE under two conditions.The results are compared with EKF,UKF and AEKF.The experimental results show that compared with the other three algorithms,AUKF has better precision and faster convergence speed.Finally,the performance of AUKF is further verified under different operating conditions.(3)Based on the aging model of different lithium-ion batteries,the parameters of the model are updated in real time by particle filter algorithm for RUL prediction.By comparing the predicted results with the actual measurement values,the estimation error is obtained.Compared with the other two models,the ensemble model shown in this paper has better regression characteristics in the whole battery life.The simulation results show that the estimation error based on the ensemble model is the smallest and the prediction effect is the best.Based on the above work,this paper has made some progress,which has certain reference significance for future research.