Research On Fault Diagnosis Methods Of Electric Vehicle Power Battery

The rapid development of world industry has led to many problems such as energy shortage and environmental pollution.New energy vehicles have become the main development direction of the world automobile industry,and its main force is electric vehicles.The power source of electric vehicle is battery,which plays a very important role.Nowadays lithium ion power batteries are used by most electric vehicles.The performance of lithium batteries is affected by complex working conditions and various factors,and minor faults may reduce the driving range of electric vehicles and cause inconvenience for people to travel.However,serious faults have the risk that the battery will catch fire and even explode,endangering people’s lives and property.Therefore,it is necessary to diagnose the battery faults,which can avoid these situations.The fault diagnosis method of electric vehicle power battery is studied in this paper.Because lithium battery is a very complex nonlinear system,it is very difficult to judge the cause of the fault,so we choose to build a fault diagnosis system of lithium battery based on fuzzy neural network.This paper analyzes the importance of battery fault diagnosis system,and briefly describes the current situation of electric vehicle and battery management system at home and abroad and the development of fault diagnosis technology.After that,the working principle of pure electric vehicle is analyzed,the power battery and battery management system are also analyzed,A battery fault diagnosis system attached to the battery management system is designed to diagnose the faults that BMS is difficult to diagnose,such as small battery capacity,excessive internal resistance,insufficient charge and large self-discharge.This paper introduces the working principle of lithium battery,analyzes the mechanism of battery failure,and analyzes the relationship between battery failure and battery failure symptoms.Due to the limitations of the experiment and considering personal safety,besides the necessary experiments,a battery model is built to simulate the experiment.Select the second-order RC model as the equivalent model of lithium ion battery,identify the parameters,build the battery model based on MATLAB platform,and verify the accuracy of the model.Using Simulink to simulate the charge and discharge of four kinds of faulty batteries,which are smaller battery capacity,increased internal resistance,insufficient charge and large self-discharge,respectively,and get the charge and discharge curves and voltage historical data of faulty and normal batteries in MATLAB.According to fuzzy set knowledge and battery fault mechanism,the battery fault diagnosis rules are determined.The membership degree is calculated by the membership function of each symptom,and the training sample library and test samples of neural network are established.Two kinds of neural networks,BP(Back Propagation)and RBF(Radial Basis Function),are used to train the training sample database,and the simulation results show that these two methods can accurately diagnose battery faults.Compared with BP neural network,RBF neural network is simple and the training speed is fast.A classification model of Support Vector Machine(SVM)is created in MATLAB,which labels the battery faults,trains and classifies and predicts them.The predicted fault categories are the same as the actual fault categories,and the prediction results are accurate.These three methods can accurately diagnose the battery fault.