Research On Performance Prediction And Fault Diagnosis Of Electric Vehicle Power Battery

New energy vehicles can effectively help to improve the environmental pollution caused by fuel vehicle emissions,and can effectively alleviate some of the pressure brought by energy shortage.Power battery system is the core component of electric vehicle.Power battery technology directly affects the safety,convenience,economy and life of electric vehicle.The batteries used in electric vehicles must meet certain requirements,such as safety,economy and long service life.At present,the lithium-ion power battery has opened an era of environmental protection and energy saving for China’s automotive engineering industry.However,the lithium-ion power battery storage energy storage system is not fully guaranteed in technology and service life,which limits its wide operation.Combined with the current development status and trend of electric vehicles,it is necessary to study the performance analysis and fault diagnosis of automotive power battery.This paper mainly studies the lithium-ion battery pack of hybrid electric vehicle.The main research work is as follows:(1)Based on a conventional method,the state of charge(SOC)of power battery is calculated,By comparing with a series of other algorithms,the optimized least squares support vector regression machine is proposed to predict SOC,which can greatly reduce the estimation and error of power battery SOC,and improve the efficiency and safety of hybrid vehicle.(2)Charging and discharging overtime often occurs in the power battery pack,which will greatly affect its performance and reduce the remaining useful life(RUL),which will shorten the battery range of electric vehicles.Another criterion of battery performance is remaining useful life RUL).In this paper,we propose a hybrid battery RUL prediction model based on ARIMA and SVM.Two basic models are proposed:(1)parallel combination model,which is optimized by combination prediction.(2)The series combination model is used to fit and preprocess the residual data after the prediction results.(3)In the process of using lithium-ion power battery,it may be difficult to detect the fault,and the sensor can not directly detect the fault type.Its fault characteristics need to be extracted by some algorithms.Therefore,this paper mainly uses EEMD-SVM method to analyze and extract the fault characteristics from the voltage signal of lithium-ion power battery,so as to improve the accuracy of power battery fault diagnosis.