Parameter Identification And SOC Estimation Of Power Battery For Electric Vehicle

In recent years, electric vehicles have been developing rapidly, so, as the corecomponent, the power battery becomes the focus of research throughout the world,and how to estimate the battery’s State of Charge (SOC) accurately has always beenthe emphasis and difficulty in this research field.Aiming to improve the accuracy of SOC estimation, the main work of this papershows as follows:First of all, this paper explains the significance of studying the power battery, andpoints out the superiorities of li-ion battery in comparing with several other kinds ofbatteries commonly used; then studies the definition of SOC and its affecting factors,also a comprehensive interpretation and analysis of the estimation method is given.Next, discussing the model of the li-ion battery, and in contrast withelectrochemical models and equivalent circuit models, establishing the second-orderRC network equivalent circuit model which facilitating to realize in engineeringconditions, then deducing the state-space model. Afterwards, the Open Circuit Voltage(OCV)-SOC relationship of the battery is measured, then identifying the parametersof the equivalent circuit model with an offline method, and verifying the rationality ofthe model at the same time.Online identifying the parameters of the equivalent circuit model with theForgetting Factor Recursive Least Squares (FFRLS) method, and programming it inan S-Function form in the circumstance of Matlab/Simulink, then conducting theidentification experiment under the simulating conductions, analyzing the results andthe feature of li-ion battery, the correctness and rationality of the identification methodis also proved.Proposing an OCV method which estimating both the parameters and the states,that is, identifying and revising the parameters of the model with FFRLS method, andestimating the OCV with Kalman filtering theory, then getting the SOC according to the OCV-SOC relationship. In this part, firstly, the reference value of SOC isdiscussed; secondly, the Simulink model of Parameter and State estimation method isbuilt; and at last, the accuracy of this method is proved under the simulate conditions.