Estimating The State Of Charge For Li-ion Battery By UKF

As its high terminal voltage, low self-discharge rate, the energy volume ratio andno memory effect, excellent performance, Lithium battery has been widely applied. Thispaper mainly research on lithium battery, and focus on the research of the battery’s workstatus, and described the work process by the charged of state (SOC).Battery status isone of the important indicators of battery management system (BMS). There isimportant significance for BMS. Taking control measures according to the battery SOC,it can protect the battery pack in the process of charging and discharging, and avoidbeing overcharged or discharged and causing damage to the battery pack. Meanwhile itis also an important predictor of limited driving distance. Generally, the current is stablein the discharge, and the remaining energy is calculated easily and accurately. But thereis a complex condition in the pure electric vehicle, because of the great change ofcurrent and temperature. The great change of voltage leads to calculate the SOCdifficultly.There are many kinds of methods to estimate the battery SOC, each has itsadvantages and disadvantages. In order to improve the estimation accuracy, theselection of filter method in the process of estimating has become an important branchin the research of the battery SOC.The current domestic and foreign research shows that obtaining accurate SOCexpression, establishing a proper cell model and selecting the appropriate SOCestimation algorithm are the three key factors to predict the battery SOC accurately.Based on the analysis above, this article’s main research work of SOC estimation are asfollows:①Research on a dynamical battery model (KiBaM model),the model can describethe cell available charge and bound charge, through the simulation of the model, toverify the nonlinear characteristic of the battery capacity and battery’s recoverycharacteristics②Study and compare several kinds of commonly used battery equivalent model,this paper select the Thevenin model by the comprehensive consideration of the modelcalculation, precision, applicable scope. We get a comprehensive model throughcombining the Thevenin model and KiBaM model, verify the model’s precision basedon the model parameter identification simulation.③Study the advantages and disadvantages of some commonly-used methods that estimate the battery SOC, through the analysis of the influence factors of the batterySOC, respectively study the relationship with environment temperature, charge anddischarge rate, cycle times, battery internal resistance and the battery SOC. Then obtainthe revised Ah counting method.④Study a unscented kalman filer method, and its core ideology UT transform,then establish the SOC estimation model by using UKF filter, and simulation in Matlab.Then compared with the estimation results based on the revised Ah counting method,the conclusion verify that the UKF filter is effective in estimation battery SOCThis paper based on a comprehensive battery model which combined the Theveninmodel with the KiBaM model, using UKF filter to estimate the battery SOC, andcompared it with the revised Ah counting method, the simulation results show that UKFis effective in estimation battery SOC...