Research On SOC And Capacity Estimation Of Lithium-ion Battery For Electric Vehicle

In the context of energy saving, electric vehicles have become an inevitable trend in the development of automobile industry. The battery state of charge and battery capacity are two important parameters of energy management of electric vehicles, and the estimates are key technologies of it. In this paper, we choose the lithium batteries of electric vehicles as the object, the aging state of the SOC estimation and capacitance decay estimates have been studied, and designed multiple time scales with adaptive dual unscented Kalman algorithm to simultaneously estimate SOC and capacitance attenuation.Firstly, we had tested the performance characteristics of lithium-ion batteries on experimental platform. And analyzed the temperature, the rates of charge and discharge and other factors that impacting on the SOC estimates, used charge-discharge cycles testing battery to find the change of capacity fade and internal resistance.Then create a first-order RC equivalent circuit model with capacity attenuation and use pulse discharge testing the battery to obtained experimental data, the recursive least squares identification the battery model parameters, using pulse testing and UDDS conditions verification the cell model with good accuracy.Finally, having designed the adaptive multi-scale dual UKF algorithm which uses two UKF filter, one used to estimate the battery states, and other UKF to battery parameters that in order to update battery parameters in real-time, and the micro-UKF output SOC as a measurement input for macro to estimate the battery capacity, using two time scales to reduce computation, adding adaptive aspects to deal with measurement noise and system noise to get a more accurate estimation accuracy. In the MATLAB/Simulink to simulate and analyze algorithms compare the measured data, the results show that the algorithm can obtain a higher SOC estimation accuracy, and can be traced to rapid changes in capacitance estimate.