Prediction of Battery Life and Fault Inspection of New Energy Vehicles using Big Data

Abstract

New energy vehicles have gradually become the preferred means of transportation for people to travel greenly. Lithium batteries, as batteries for new energy vehicles, its quality directly affects the safety of vehicles and mileage is also the core data that people consider when choosing vehicles one. Therefore, the research uses big data to predict and test the battery life and failure of new energy vehicles. When predicting the battery life, the improved P-GN model has a good prediction effect and the model reaches the convergence state only three iterations and converged to 0.83. The optimal fitness converges at the beginning of iteration, reaching the optimal value of 1.946 after only four iterations. The error between the predicted value and the actual value of the remaining cruising range is within an acceptable range. When the weather condition is good, the prediction effect of the remaining cruising range is excellent and the fluctuation of the prediction difference is small. When the weather conditions are severe, the model can still predict the cruising range of the battery pack normally. When performing fault detection on the battery pack, the fault detection system can accurately and quickly detect the type of fault and effectively analyse the inconsistency of the battery and be accurate to the single faulty battery. When analysing the single faulty battery, it is proposed that the fault detection system can accurately diagnose the fault in the test battery, which not only takes a short time but also has good feasibility.