Safety Prediction Of Vehicle Power Battery

Recently,power battery industry develops rapidly in China,at the same time,the safety problems such as fire and explosion caused by thermal runaway frequently occur in the use of power battery,which have aroused widespread concern of the whole society.This paper focuses on the study of thermal runaway of battery,and analyzes the main factors affecting the thermal runaway of battery,Using battery information platform to collect data of electric vehicle charging in operation,combined with the improved outlier detection method,a battery safety prediction method is proposed.The main research content of this article was as follow:Firstly,based on searching a large number of domestic and foreign literature on the study of fire and explosion of power batteries,this paper summarized the prediction of thermal runaway and the safety problems caused by thermal runaway,it is found that there is a close relationship between the difference of internal resistance,the difference of battery health status and the temperature on battery safety,so this paper focused on this lack.Secondly,this paper built a power battery monitoring information platform composed of vehicle information collection terminal and remote monitoring system.Vehicle information collection terminal was the terminal equipment responsible for the collection and transmission of vehicle battery information,and the remote monitoring system was responsible for receiving the battery data,analyzing the current status of the battery and storing the relevant data in the database.Then,the internal resistance and the health status of the battery are estimated by the collected battery data.Aiming at the internal resistance of battery,this paper presented a recursive least square method to identify the internal resistance of battery.The recursive least square method can track the parameters quickly and in real time,which was more conducive to the real-time monitoring of battery.In view of the battery health state,this paper proposes a battery health state estimation method based on the extreme learning machine with equal voltage interval charging time.The new health indicators selected were easier to obtain.Combined with the characteristics of the extreme learning machine,such as fast calculation speed and high generalization,the battery health state can be obtained quickly and accurately.Finally,this paper proposed an improved outlier detection method to detect the temperature,internal resistance and health status of the power battery,and directly obtained the safety index of the power battery.The results are combined with the actual detection results of the experimental vehicle.It is proved that the safety index of power battery analyzed in this paper was helpful to reduce the possibility of accidents caused by thermal runaway of power battery.