Research On State-of-charge Estimation Of Life PO₄Battery

Because the requirement of energy saving and low-carbon, electric vehicle as aclean energy consumption has become a hot issue on current research. Battery as thepower source of electric vehicles which is directly restricted its development. Thelithium iron phosphate battery have many advantages such as security, high specificpower, long cycle life, environment-friendly and so on, so it become the best choice forelectric vehicle power supply. Accurate estimate the state of charge (SOC) is thepremise to improve battery efficiency and increase the battery life, and also thekeystone and difficulty in the study of battery’s management system. For the purpose ofaccurately estimate SOC, this thesis is mainly do the following work.This paper start from the principle and operating character of the lithium ironphosphate battery and analysis battery’s voltage character, internal resistance character,capacity character and analysis battery’s various capacity affecting factors in detail.Based on that, aim at the problem the traditional SOC definition can’t adapt to thedynamic condition, and SOC-OCV curve based on the tradition SOC definition can’tuse in battery’s whole cycle life, so adapt the dynamic SOC definition. After all thispaper use test to prove the accuracy of the dynamic SOC definition.At the same SOC point the open circuit voltage (OCV) of battery is different whilebattery’s rest time is different. If the battery’s rest time is not long enough, the voltagehysteresis does not disappear completely, so it will bring on such phenomenon that theSOC-OCV curve under the charge condition and discharge condition does not coincide.If not consider the influence of rest time it will cause a great SOC estimation error. Inview of this situation, a new SOC-OCV relationship with rest time was used in thispaper. So it can realize the fast and accurate correction of SOC, no matter battery atwhich rest time.The estimation accuracy of the extended Kalman filter (EKF) algorithm dependson the accuracy of the battery model. On the basis of the analysis and comparison ofdifferent battery models, the second-order RC model which has universal applicabilityby changing the parameters was chosen as the paper used equivalent circuit model. Andthen the improved hybrid pulse test was used to calculate the parameters of the batterymodel, then the accuracy of the second-order RC model was proved though the contrastwith the result of simulation and experiments. When estimate SOC, this paper use a method which combine the extended Kalman filter, open-circuit voltage method andcurrent integral method and make a further research to this method, and in the endverify the feasibility of this method.Knowing the operating character of the battery and parameter identification of thebattery model require a lot of preliminary experiments, according to these needs abattery experimental platform was built, and this platform can also be seen as a simplebattery management system. In the end compile the flow chart of extended Kalmanfilter and use full-cycle test to verification this algorithm can track battery SOCpreferable.