Design And Implementation Of Battery SOC Estimator Based On Double Kalman Filtering Algorithm

With the continuous development of new energy vehicles, electric vehicles also receivemore and more attention. As the key power devices in electric vehicles, power batteriesbecame a research hotspot gradually. In order to enhance and improve battery perfor-mance, battery management system is required to manage the batteries with full rangeof protection and monitoring. Battery state of charge(SOC) is an important state of bat-tery, which provides important reference basis for the management of the batteries. Inthe same time, SOC afects the vehicle performance control.Firstly, the importance of SOC estimation is expounded, main current power bat-teries are introduced, also the advantages and disadvantages of various SOC estimationalgorithm are compared with.Then the basic structure and working principle of lithium-ion batteries are intro-duced. Current lithium ion batteries research model of electric car are compared with.And corresponding battery state equations are given according to the diferent batteryequivalent circuit model.After built the battery model, some related battery parameter identifcation experi-ment are carried out to get battery model parameters. The relation curve of open circuitvoltage and battery SOC is obtained through calibration experiment. Battery capaci-tance and resistance parameters are obtained through parameter ftting experiment, andthe parameters are identifed in the model validation test. In this test, the outputs ofbattery model are compared with the outputs of real battery with same inputs in orderto verify the accuracy of battery model.Based on Kalman Filtering principle and Extended Kalman Filtering algorithm, Dou-ble Kalman Filtering algorithm is put forward to estimate lithium ion power batterySOC. Firstly, EKF method could get the frst SOC estimates, then Ampere-Hour Integralmethod could get the second SOC estimates, at last the two SOC estimates are weightedcombined in Double Kamlan Filtering to get an fnal DKF SOC estimate. Through manyexperiments, DKF SOC estimate is compared with AH SOC estimate and EKF SOC estimate to verify the accuracy and the advantages of DKF SOC estimation algorithm.In the end, the DKF estimation algorithm is embedded in embedded system basedon ARM platform for physical verifcation experiment. The embedded system includesupper computer and lower computer. Lower computer is mainly responsible for data ac-quisition, and the data acquired from lower computer, like battery current and voltage,are transmitted to upper computer for data analysis and processing. Also upper com-puter is in charge of protection control to ensure the operation of the batteries. DKFSOC estimation is running in lower computer, the lower computer could get DKF SOCestimate through the data collected from lower computer. At last, the efectiveness ofDKF estimation algorithm is verifed through the analysis of DKF SOC estimate fromvarious experiments.