Research On Power Battery SOC Estimation Based On LSTM Neural Network Adapting To Small Sample Data

As the power source for electric vehicles,batteries are the power source for electric vehicles.Battery charge status(Stage of Charge,SOC)is a parameter that reflects the remaining battery level.The range of values is 0 to 1.The SOC is concerned with monitoring the operating state and battery life and is an important indicator of the battery capacity.Usually,the power battery SOC is not directly available and can be estimated.By linking the battery SOC to the battery state,power,including voltage,current,or internal resistance.The research methods selected in this article are neural network backup methods.The condition of the study is that the battery SOC must be assessed on condition that it is appropriate for the data of small samples.Evaluating the state of charge of a battery is excellent research and is worthwhile in practice.The specific research content of this article is as follows.1.First,clarify the structure and function of the battery,analyze the current state of research on the methods of assessing the battery SOC,classifying the SOC assessment method,explaining the advantages and disadvantages of the various methods,and the research used in this article identifies the methods..2.The selected single cell 18650 lithium battery has been tested and analyzed for performance.The test platform was built and tested in accordance with GB/T 31485-2015 and GB/T 38661-2020 standard requirements for experimental battery testing.Diagram for generating and processing experimental data.The characteristics of charged and discharged batteries are analyzed.This includes the effects of voltage,current,charging and discharging speed,and battery temperature.In addition,to determine the relationship between the above parameters and the battery SOC,a sensitivity is analyzed to determine the parameters related to the battery SOC.3.Consider the internal battery SOC estimation model,the standard LSTM neural network model,construct the battery SOC estimation model based on the LSTM neural network,and analyze the results of the SOC estimation model for the LSTM neural network from This article presents a model of battery SOC estimation based on the LSTM neural network that optimizes the update rules of the LSTM neural network and adapts it to small sample data and SOC estimation of Accurate and fast battery using small data can help you achieve.We will compare the precision sample data of battery SOC estimation in various neural networks and benefit from the research methods used in this article.The conclusion shows that the battery SOC estimation method based on the LSTM neural network,which is suitable for small sample data compared to other approximation methods,has an accurate and stable battery SOC estimation.During the charging process,the mean square error of the MSE,the mean absolute error MAE error and the SE standard error of the estimation using the LSTM adjustable neural network,small samples reached 0.0029%,0.12%,and 0.18%.The MSE,MAE and SE errors in this network estimation during the emission process reached 0.0015%,0.15%,and 0.22%,respectively.