Online identification of battery model parameters and joint state of charge and state of health estimation using dual particle filter algorithms
Yonghong Xu,Xia Chen,Hongguang Zhang,Fubin Yang,Liang Tong,Yifan Yang,Dong Yan,Anren Yang,Mingzhe Yu,Zhuxian Liu,Yan Wang
DOI: https://doi.org/10.1002/er.8541
IF: 4.6
2022-08-19
International Journal of Energy Research
Abstract:Summary Aiming at the problems of time‐varying battery parameters and inaccurate estimations of state of charge (SOC) and state of health (SOH), a joint estimation algorithm of SOC and SOH is proposed. A particle filter algorithm is used to identify the parameters online on the basis of a second‐order equivalent circuit model. The algorithm feasibility is verified through the terminal voltage estimation accuracy. Considering that an accurate SOH is one of the foundations to achieve an accurate SOC estimation, a dual particle filter is used to jointly estimate SOC and SOH. Under different test conditions, the effect of different initial values (initial SOC and capacity), temperatures, operation conditions, particle number, and model parameters on the estimation accuracy and robustness is compared and analyzed. The effectiveness of the proposed algorithm is validated by experimental data under different operation conditions. Experimental results show that the online particle filter algorithm can well predict the dynamic battery model parameters. The proposed algorithm has high robustness and a good tracking effect when estimating SOC with a mean absolute error of less than 1.3%, a root mean square error of less than 1%, and a tracking terminal voltage.
energy & fuels,nuclear science & technology