Lithium-ion battery state of charge estimation with model parameters adaptation using H∞ extended Kalman filter. (December 2018)

• Record Type: Journal Article
• Authors: Zhao, Linhui; Liu, Zhiyuan; Ji, Guohuang
• Is Part Of: Control engineering practice. Volume 81(2018)
• Page Start: 114
• Page End: 128
• Publication Date: 2018-12
• Subjects: Electric vehicle -- Lithium-ion battery -- State of charge -- Model uncertainties -- H∞ extended Kalman filter -- Model parameters adaptation; Automatic control -- Periodicals; 629.89
• DOI: 10.1016/j.conengprac.2018.09.010
• Languages: English
• ISSNs: 0967-0661

Abstract

Abstract: Model-based methods can effectively improve the estimation accuracy of state of charge (SOC) for lithium-ion batteries in electric vehicles. Due to the influence of complex electrochemical mechanism and other factors such as temperature, model uncertainties, including unmodeled dynamics and varying parameters, result in that it is difficult to obtain accurate estimation of the SOC using an equivalent circuit model with fixed parameter values under various working conditions of battery. In this paper, two nonlinear models based on single and dual RC models are established, and observability of the nonlinear models is discussed. To bound the influence of model uncertainties, an H ∞ extended Kalman filter is proposed based on robust control theory to estimate the SOC, Ohmic and polarization resistances simultaneously. The performance and robustness of the proposed method are evaluated and compared with a standard extended Kalman filter using multi-temperature datasets. The experimental results show that the proposed method is capable to estimate the SOC more accurately over a large operating range of battery. Furthermore, the validation results of datasets from a battery management system confirm that the proposed method can achieve good performance for real life conditions in a battery pack of electric vehicles. Highlights: An HEKF is proposed to estimate battery SOC and model parameters simultaneously. Performance of proposed method is evaluated based on multi-temperature datasets. Robustness of the proposed method against different disturbances is validated. The methods get good performance for real life conditions of actual battery pack.