Lithium Battery SOC Estimation Based On Unscented Kalman Filter

This paper uses 3.2V/10 Ah lithium iron phosphate battery as experimental research object,aiming at the difficulty of SOC estimation of power lithium-ion battery under complex working conditions and complex environmental conditions,relevant researches are carried out.This paper analyzes the basic characteristics and working principle of lithium iron phosphate batteries,and studied the battery charge and discharge rate,environmental temperature,recycle times influence on the working performance of the battery.On the basis of the PNGV equivalent circuit model,based on the unscented Kalman filter algorithm,the battery state of charge is estimated in real time.Aiming at the numerical instability and filter divergence problem of UKF estimation SOC,an improved method of adding noise adaptive covariance matching is adopted.At the same time,the threshold adjustment factor is introduced to automatically select the window size and the noise matrix is corrected in real time.The simulation experiments were carried out under constant current discharge conditions,subsection constant current discharge conditions,typical DST conditions,and verified the improved algorithm under different initial values.The simulation results show that the improved algorithm can not only avoid the matrix negative and unsteady values of the unscented Kalman filter,but also greatly improve the estimation accuracy of the state of charge.In the case of current changes constantly,the maximum error of the improved algorithm is only about 0.5%.