Research On Reliability Modeling And Evaluation Of Powertrain Of Electric Vehicle

Under the concepts of energy saving,emission reduction,and environmental protection,electric vehicles have developed rapidly,and the global trend of vehicle electrification has become clear.However,in the process of rapid development,the development quality of electric vehicles has always attracted attention,and reliability problems of electric vehicles have occurred from time to time.However,as the key power system of electric vehicles,its reliability determines the safety performance of the vehicle to a large extent.The current research on its reliability has not been fully carried out,and the related research results are rarely reported.For this reason,The thesis takes the electric vehicle power system as the research object,and "research on system structure characteristics,working characteristics and failure analysiscore subsystem reliability evaluation research-power system overall reliability evaluation research-the degradation of power system vulnerable components under different working conditions Characteristic and Reliability Evaluation Research” is the main line,which conducts research on the reliability of electric vehicle power system and related issues.The main content includes:(1)Aiming at the unclear correlation strength between the failure of each component of the electric vehicle power system and the failure of the entire system,and the unclear logical relationship between component failure,sub-system failure and the failure of the entire system.The thesis uses the FMEA(Failure Mode and Effects Analysis)method to analyze the components of the power system,and builds the drive motor,motor controller,PDU(Power Distribution Unit)and battery system fault tree models;finally,using the constructed theoretical model,the failure rate,importance and reliability of each component of the power system are simulated and calculated,and the influence value of each component failure on the failure of the entire system and the reliability of the power system under different service periods are obtained.The research results show that the battery module is the most prone to failure component.(2)In the service process of battery modules,the service life is mainly affected by factors such as temperature,charge and discharge current,and state of charge,but the primary and secondary relationship of these factors on the battery service life is not clear.In response to this problem,the paper proposes a lithium-ion battery life prediction and reliability evaluation method that considers stress acceleration and stress compensation.First,through orthogonal experiment,main effect analysis and variance analysis,the impact of different stress factors on battery life is obtained.The primary and secondary relationships are temperature,current,and state of charge in order;secondly,according to the primary and secondary relationship,the corresponding compensation function is introduced to construct the battery life prediction model under different working conditions;finally,the discharge current is calculated at 1.3C at room temperature.The test conditions and the 0.5C discharge current working condition at an ambient temperature of 40℃ are used as the test conditions.The life prediction model constructed is used to predict and evaluate the life and reliability of the studied battery,and the battery life interval and Corresponding reliability change curve.(3)During the service process of lithium-ion batteries,there are many factors that affect their performance degradation,and it is difficult to predict and evaluate their remaining life and reliability with accurate physical models.In response to this problem,based on the Wiener process combined with Bayesian theory,the paper proposes a lithium-ion battery remaining life prediction and real-time reliability evaluation model.In this model,the drift coefficient is used to describe the individual difference in product degradation,and the diffusion coefficient is used to describe the overall difference in the product.Obtain the initial parameters of the model based on historical data and Gibbs sampling algorithm,obtain realtime data during the test,and apply the construction model to fuse the historical data and realtime data to realize the prediction of the remaining life of the lithium-ion battery and the realtime evaluation of the reliability.The research results show that,compared with the evaluation data of other Wiener models,the dual random parameter Wiener model used in the paper has a better predictive effect on the reliability of batch lithium-ion batteries.