A novel multi-model probability battery state of charge estimation approach for electric vehicles using H-infinity algorithm. (15th March 2016)

• Record Type: Journal Article
• Authors: Lin, Cheng; Mu, Hao; Xiong, Rui; Shen, Weixiang
• Is Part Of: Applied energy. Volume 166(2016)
• Page Start: 76
• Page End: 83
• Publication Date: 2016-03-15
• Subjects: Electric vehicles -- Batteries -- State of charge estimation -- Multi-model probability -- H-infinity; Power (Mechanics) -- Periodicals; Energy conservation -- Periodicals; Energy conversion -- Periodicals; 621.042
• DOI: 10.1016/j.apenergy.2016.01.010
• Languages: English
• ISSNs: 0306-2619

Abstract

Highlights: A novel multi-model probability battery SOC fusion estimation approach was proposed. The linear matrix inequality-based H∞ technique is employed to estimate the SOC. The Bayes theorem has been employed to realize the optimal weight for the fusion. The robustness of the proposed approach is verified by different batteries. The results show that the proposed method can promote global estimation accuracy. Abstract: Due to the strong nonlinearity and complex time-variant property of batteries, the existing state of charge (SOC) estimation approaches based on a single equivalent circuit model (ECM) cannot provide the accurate SOC for the entire discharging period. This paper aims to present a novel SOC estimation approach based on a multiple ECMs fusion method for improving the practical application performance. In the proposed approach, three battery ECMs, namely the Thevenin model, the double polarization model and the 3rd order RC model, are selected to describe the dynamic voltage of lithium-ion batteries and the genetic algorithm is then used to determine the model parameters. The linear matrix inequality-based H-infinity technique is employed to estimate the SOC from the three models and the Bayes theorem-based probability method is employed to determine the optimal weights for synthesizing the SOCs estimated from the three models. Two types of lithium-ion batteries are used to verify the feasibility and robustness of the proposed approach. The results indicate that the proposed approach can improve the accuracy and reliability of the SOC estimation against uncertain battery materials and inaccurate initial states.