Co-estimation of state of charge and state of health for 48 V battery system based on cubature Kalman filter and H-infinity. (10th December 2022)

• Record Type: Journal Article
• Authors: Ning, Zhansheng; Deng, Zhongwei; Li, Jinwen; Liu, Hongao; Guo, Wenchao
• Is Part Of: Journal of energy storage. Volume 56:Part B(2022)
• Publication Date: 2022-12-10
• Subjects: State of charge -- State of health -- Multi-timescale -- Cubature Kalman filter -- H-infinity -- Parameters identification; Energy storage -- Periodicals; Energy storage -- Research -- Periodicals; 621.3126
• DOI: 10.1016/j.est.2022.106052
• Languages: English
• ISSNs: 2352-152X

Abstract

Abstract: As a cost-effective vehicle electrification solution, 48 V system can achieve a 15 %–20 % reduction in CO2 emission without major changes to the traditional automotive architecture. The aggressive charging and discharging conditions of this system put forward higher requirements for battery management system (BMS), especially for accurate state estimation. In this paper, a multi-scale co-estimation method is proposed to obtain the state of charge (SOC) and state of health (SOH) for 48 V battery system. First, a multi-scale estimation framework is developed according to the natural characteristics of the battery parameters and states. Second, based on the equivalent circuit model of battery, the recursive least squares is used to online identify the internal resistance of battery, advanced filtering algorithms including cubature Kalman filter and H-infinity are employed to estimate battery SOC and capacity. Through the multi-timescale framework, the computing burden of co-estimation can be reduced to enable its application on an embedded BMS. Finally, a series of tests based on typical driving cycles are carried out to verify the proposed method, and the results are also compared with other existing methods. The results show that the mean absolute errors of SOC and capacity estimation are less than 0.88 %, and 0.64 %, respectively, which verifies the superiority of the proposed method. Highlights: A method is proposed to achieve the co-estimation of battery states. A multi-timescale framework is designed to reduce the computing burden. The cubature Kalman filter is used to strengthen the nonlinear estimation ability. The H-infinity algorithm is employed to achieve high-precision capacity estimation.