Research On Estimation Technology Of Lithium Battery SOC For Electric Vehicles

In the current global advocacy of "low carbon emissions",the electric car attractsthe attention of the people with its pollution-free exhaust emission, new energy drivingetc.Our government has issued a number of policies to support the development ofelectric vehicle industry. The accurate prediction of battery states of charge(SOC) inelectric cars is a key technology of battery management system,and it is an importanttechnology for electric vehicle. Therefore the research of it is of great significance.Lithium iron phosphate battery is the most common used battery for electric cars.ithas good stability, specific heat energy, long circle life, etc. In this paper,we take lithiumiron phosphate battery as experimental object, begin to battery remaining poweraccurate prediction research. First analyzes the characteristic of lithium batteries,including temperature characteristics, charging and discharging characteristics,self-discharge and characteristics of aging.Based on these characteristics, I put forwardseveral major factors that affect battery’s SOC, which guide the process on theprediction the battery’s SOC.Second, a simple battery management system hardwareplatform is constructed, though which we can do requirements according to instructions,such as charging and discharging the battery under series of correspondingrates,monitoring some necessary parameters,gathering and transmission datainformation.At last, considering about the traditional SOC forecasting methods usuallylack precision,in this paper I analyzed the neural network algorithm, finally I putforward the BP neural network algorithm to solve remaining power calculation. Matlabsoftware is used to construct the neural network model, first import the sample data intomodel for training, then test model with other set of sample data. Through theverification, this method error is within8%and the prediction accuracy has metdemanded.According to the influence of temperature on the SOC, test the monomer battery indifferent temperature by charging and discharging experiment a lot of times in different temperatures, and find the relationship between temperature and battery capacity, revisethe model with establish compensation, so as to further improve the predictionprecision.After further modified, in the temperatures that battery can withstand, thismodel can be used for single battery and the whole battery pack charge or discharge atany rate in power simulation, after verified, the measured predicted further close to thereal value, revised model is more accurate.