Modeling And State-of-charge Estimation Of LiFePO₄ Battery In Electric Vehicles

LiFePO₄as positive materials of lithium ion batteries is paid more and moreattention by its good performance and low cost, and it is expected to become the powersource of scale application on electric vehicles in the future. This paper first does somestudy on the charge and discharge characteristics ofLiFePO 4battery, analyzes thevoltage response characteristic of different discharge rate and temperatures. Finally getsthe curve of the Open circuit voltage（OCV） and terminal voltage with the varying ofState-of-Charge（SOC）, the relationship curve between the terminal voltage and SOC withdifferent discharge rate and the curve of the temperature characteristic when the battery ischarging and discharging. And it provides the testing support for the entirelyunderstanding about the charge and discharge characteristics of theLiFePO 4battery.This paper established the Thevenin equivalent circuit model which is suitable forelectric vehicleLiFePO₄battery. And it includes a large number of experiments in atypical environment temperature, and obtains the model parameters under various ambienttemperatures through the Hybrid Pulse Power Characterization Test（HPPC）, and revealsthe law of the model parameters with temperature variation. The research on the influenceabout the temperature to the model parameters can provide a theoretical basis for thedevelopment and design of Battery Management System（BMS）. At the same time, it usesthe Beijing Bus Dynamic Stress Test（BBDST） to identify the Thevenin model, andanalyzes the reliability of the model. The results show that, the Thevenin model can meetthe requirements of practical applications with relatively high precision .The SOC estimation of electric vehicles is one of the core issues needed to work outon BMS studies, the accuracy degree of the estimation is important to the battery’s fullperformance. In this paper, Extended Kalman Filter algorithm is used to estimate the SOCon the basis of theLiFePO 4model which has been established already. The researchshows that, this method can estimate the remaining power which is almost the same withthe result based on Thevenin model, it can achieve the precision of the SOC estimation.