A novel parameters identification method of Lithium-ion battery equivalent circuit model under dynamic stress test

This paper selects the second-order equivalent circuit model (2
^nd
ECM) as the battery model and proposes decoupled recursive least squares with adaptive moving window forgetting factor (AMWFF-DRLS), and the algorithm decouples different dynamic impedances in the 2
^nd
ECM. The objective function solves the interference between each other and successfully combines the proposed AMWFF technology to make adaptive weight adjustments through historically predicted voltage response error data points, effectively reducing the amount of algorithm calculation and improving parameter identification accuracy. Finally, this paper establishes a mathematical model of Lithium-ion batteries on the MATLAB platform and selects a dynamic stress test (DST), to verify the model parameter estimation ability of the AMWFF-DRLS. The estimation performance is compared with traditional RLS methods. Under the DST working condition, the average absolute error (MAE) of AMWFF-DRLS is improved by 16.67 % compared to the traditional RLS, while the root mean square error (RMSE) is improved by 14.93 %.