Electric Vehicle Safety System Development And Battery Life Prediction

With the increasing attention to electric vehicles in China,the ownership of electric vehicles is growing rapidly,but while it is convenient for transportation,the safety of electric vehicles is also being questioned.In recent years,there are frequent safety accidents such as lithium battery spontaneous combustion,uncontrolled motor and brake failure in electric vehicles,most of which are caused by electric vehicles not being maintained for a long time and owners not knowing enough about the status of electric vehicles,resulting in overuse of lithium batteries.For this reason,this paper designs a complete set of electric vehicle on-board intelligent safety monitoring systems,which uses automotive CAN bus,Beidou positioning,4G communication,Android cell phone client,and deep learning technologies to achieve data collection,fault diagnosis,single lithium-ion battery prediction,and other functions.The details are as follows.(1)According to the functional requirements of the system,the overall design of the electric vehicle safety system is proposed,and the system consists of three parts: on-board terminal,cloud server,and Android cell phone client.The onboard terminal reads key data from the motor control unit,vehicle controller,and battery management system on the electric vehicle through the CAN bus and uploads them to the cloud server through the 4G module.The cloud server receives the data from the vehicle terminal and saves it in the My SQL database as viewable historical data.Vehicle-related data and vehicle fault information are viewed through the Android cell phone client.The expanded function of the system is to propose the remaining life prediction of parts through the public data set on the Internet and use the public data set to build a suitable prediction model to predict the remaining life of key locations,which is described in detail in this paper with a lithium-ion battery as an example.(2)Hardware design of the vehicle-mounted terminal and its software development.The vehicle-mounted terminal uses STM32F105 RB as the main controller and integrates the power conversion module,CAN bus collection module,4G communication module,and Beidou positioning system module.For software development,the software design of data acquisition,4G communication,and Bei Dou positioning is carried out in C language based on the embedded real-time operating system.(3)In order to build a prediction model for the remaining life of multiple single lithium batteries,a prediction model combining an attention-based convolutional neural network with a long and short-term memory network is proposed and trained and validated using a publicly available dataset of cyclic charging and discharging of lithium batteries.(4)Remote monitoring platform design.Based on Aliyun,the server is developed to realize the functions of data reception,processing,storage,and transmission,etc.The app design thinks from the user’s point of view,designs a refreshing interface and simple operation functions,and realizes the display of information such as real-time data,historical data,and historical faults.(5)Experimental testing.The test results show that the hardware design of each module of the vehicle terminal is reasonable,the remote transmission is reliable,the App can accurately read and display the data of the electric vehicle and prompt the fault information of the vehicle.The system designed in this paper realizes the functions of real-time data collection,historical data storage,fault diagnosis,and fault information reminding,which can grasp the safety condition of electric vehicles in real-time and judge whether the vehicle is malfunctioning in time,which helps the maintenance of electric vehicles and has certain engineering application value.