Research On Design Of SOC Prediction Model Of NewEnergy Vehicle Power Battery

In recent years,the energy crisis and environmental pollution have increasingly become the focus of public attention,and new energy vehicles have been rapidly developed as a means to effectively alleviate the crisis.The power system is an indispensable energy storage system for new energy vehicles.To a certain extent,it determines the performance of new energy vehicles.The state of charge(SOC)of the power battery is one of the key problems that the power system of the new energy vehicle needs to solve.Accurate SOC prediction can effectively judge the cruising range of the car,can provide users with reasonable vehicle control opinions,and ensure the reliable operation of the battery management system.At the same time,for power batteries,the quasi-deficiency SOC prediction can effectively avoid operations such as overcharging and overdischarging of power batteries that will permanently damage the battery.Therefore,the SOC prediction of power batteries,as a key link in the development of new energy vehicles,has very important research significance.In this paper,the lithium iron phosphate battery was selected as the research object through comparison.In order to make the research more realistic and reflect the overall performance of power batteries in industrial applications,this paper selects the lithium iron phosphate battery pack used in real vehicles as the experimental object.First,after understanding the battery characteristics and performance parameters of lithium iron phosphate batteries,we designed and conducted experiments for the selfdischarge,discharge voltage,discharge rate,ambient temperature,and internal resistance of the battery pack.The characteristics of various battery performance parameters and their influence on the SOC of the lithium iron phosphate battery pack are analyzed through experiments.Finally,the discharge voltage,discharge rate,ambient temperature and battery internal resistance are selected as the main parameters of SOC prediction.Secondly,this paper designs and builds a SOC prediction model based on BP neural network on the MATLAB platform.The discharge data of lithium iron phosphate battery packs in various environments were collected through experiments,and these data were used to train the SOC prediction model.The experimental research on the trained model is carried out,the influence of the number of hidden layers on the SOC prediction accuracy is analyzed,and finally the conclusion that the prediction accuracy of the model built only using BP neural network is insufficient and the stability needs to be improved is obtained.Finally,in view of the problems existing in the SOC prediction model built by BP neural network,this paper uses genetic algorithm to optimize the model weight threshold selection.Through research,design and use MATLAB code to write genetic optimization algorithm for BP neural network SOC prediction,the optimization algorithm and prediction model fusion.Use the data obtained from previous experiments to train and analyze the optimized model.The experimental results show that the genetic algorithm has obvious optimization effects on the accuracy and stability of the prediction model,and under the premise that the hidden layer is small,a model that meets the accuracy of the actual application can be obtained,which has a great impact on the model topology The simplified effect shortens the training and prediction time of the model.