Research On Battery Fault Diagnosis Of Pure Electric Vehicle Based On Android

In recent years,the two global problems of environmental pollution and energy crisis have become increasingly serious,and it is urgent to seek clean and environmentally friendly energy to replace fossil fuels.Because pure electric vehicles have outstanding advantages in terms of energy saving and environmental protection,electric vehicles gradually replace fuel vehicles as people’s means of transportation.Two factors,vehicle mileage and power battery safety,restrict the development of electric vehicles,which impose higher requirements on the performance of power batteries.Voltage faults are the most common fault phenomenon in battery faults.It is of great significance to quickly and accurately detect the possible faults of batteries to improve battery safety and prolong service life.This article takes 18650 lithium-ion battery as the research object,analyzes the battery polarization phenomenon and inconsistency by studying the working principle and operating characteristics of lithium ion battery.The research results show that the battery ambient temperature,battery over-charge and over-discharge,charge-discharge rate and self-discharge phenomenon will have a greater impact on battery internal resistance,capacity,voltage and other parameters.Secondly,build a fuzzy neural network battery fault diagnosis model,use fuzzy system and BP neural network to solve the uncertainty problem of battery fault,mainly based on the battery voltage caused by the battery fault characteristics and the cause of the fault as the model input Variables,establish the fault membership function and process the original battery voltage data to obtain the respective membership,and use fuzzy sets to represent.Through MATLAB simulation,the target output and actual output of the model are compared,and it is concluded that the diagnostic model meets the accuracy requirements of the system.Finally,the battery fault diagnosis model is mounted on the battery fault diagnosis APP based on the Android system.The Android client mainly implements communication and data transmission through Bluetooth,displays the important battery and other battery information collected by the lower computer on the interface,and processes the battery voltage parameters according to the model membership function.Imports the diagnosis model output results,according to the model,Theoutput result is used for early warning of the cause of failure.Use CANoe to simulate fault data,and establish a communication channel with the Android client via Bluetooth to CAN module(CANBlue)to verify the practicability of the model on the Android platform.The results prove that the APP can output accurate warning information for battery fault diagnosis.