Enhanced online model identification and state of charge estimation for lithium-ion battery with a FBCRLS based observer. (1st November 2016)

• Record Type: Journal Article
• Authors: Wei, Zhongbao; Meng, Shujuan; Xiong, Binyu; Ji, Dongxu; Tseng, King Jet
• Is Part Of: Applied energy. Volume 181(2016)
• Page Start: 332
• Page End: 341
• Publication Date: 2016-11-01
• Subjects: Model identification -- State of charge -- Online estimation -- Noise variances estimate -- Bias compensation -- Lithium-ion battery; Power (Mechanics) -- Periodicals; Energy conservation -- Periodicals; Energy conversion -- Periodicals; 621.042
• DOI: 10.1016/j.apenergy.2016.08.103
• Languages: English
• ISSNs: 0306-2619

Abstract

Highlights: Integrated online model identification and SOC estimate is explored. Noise variances are online estimated in a data-driven way. Identification bias caused by noise corruption is attenuated. SOC is online estimated with high accuracy and fast convergence. Algorithm comparison shows the superiority of proposed method. Abstract: State of charge (SOC) estimators with online identified battery model have proven to have high accuracy and better robustness due to the timely adaption of time varying model parameters. In this paper, we show that the common methods for model identification are intrinsically biased if both the current and voltage sensors are corrupted with noises. The uncertainties in battery model further degrade the accuracy and robustness of SOC estimate. To address this problem, this paper proposes a novel technique which integrates the Frisch scheme based bias compensating recursive least squares (FBCRLS) with a SOC observer for enhanced model identification and SOC estimate. The proposed method online estimates the noise statistics and compensates the noise effect so that the model parameters can be extracted without bias. The SOC is further estimated in real time with the online updated and unbiased battery model. Simulation and experimental studies show that the proposed FBCRLS based observer effectively attenuates the bias on model identification caused by noise contamination and as a consequence provides more reliable estimate on SOC. The proposed method is also compared with other existing methods to highlight its superiority in terms of accuracy and convergence speed.