Study On Modeling And Simulation Of Dynamic Characteristics Of LiFePO₄ Battery For Electric Vehicle

Power battery has a Significant impact on the mileage and safety of electric vehicle(EV), LiFePO4 Li-ion battery has become the prior choice for the EV power source, due to its excellent safety, non-pollution, low cost and long cycle life. The modeling of power model is a vital part of the development of battery management system(BMS), concerning that whether BMS can implement the accurate prediction of the state of chare and the state of health. Therefore based on battery dynamic characteristics, establishing battery model of high accuracy and low complexity is essential to BMS.This paper made tests and analysis on the battery dynamic characteristics, including the terminal voltage characteristic and the capacity characteristic under different rate discharge currents, the open voltage characteristic under different SOC, as well as the Electrochemical impedance spectroscopy(EIS) characteristic under different frequency electrical excitation signals, to supply certain test foundation for the subsequent modeling and simulation of battery.Due to the shortage of current Randles equivalent circuit model, by introducing the complex parameter inductance element, the complex parameter capacitance element as well as constant phase element and comparing them with the elements of Randles model, the equivalent circuit correction model was established. By simulated annealing algorithm, based on fitting EIS, the identification for correction model parameters were conducted. Then the fitting result was validated and analyzed, and was compared with the fitting result of Randles model, then it was showed that the absolute error of correction model was less than that of Randles model, compared with Randles model, the relative error was 3.83%, the accuracy was improved by about 4%. The EIS under different SOC was fitted by correction model, the relation between model parameters and SOC changing was analyzed, deriving the parameter Q has monotonous relationship with SOC, which can be referred to the SOC prediction of BMS.Considered the demand that BMS need battery time domain model to describe the battery volt-ampere characteristic, by combining the actual operation condition on EV, the correction model was transferred from frequency domain to time domain. Based on the OCV-SOC characteristic curve, the open voltage was identified; Based on the terminal voltage responses after removing the discharge current, the fitting function formula between SOC and equivalent circuit element parameters was acquired. Based on Simulink, the dynamic simulation model of LiFePO4 power battery was build, and the validations were conducted under Constant current discharge test and FUDS test. The result of simulation showed the model can greatly simulate the volt-ampere dynamic characteristic, the accuracy of model was validated, further supplying certain theoretical foundation for the application of battery on BMS.