A switching Kalman filter for SoC estimation of ultracapacitor balancing systems

• Published in: Journal of energy storage 2024-07, Vol.93, p.112170, Article 112170
• Main Authors: Li, Heng, He, Peinan, Tang, Yihan, Chang, Taozhen, Li, Shuo, Fan, Yunsheng
• Format: Article
• Language: English
• Subjects: Cell balance; Kalman filter; State of charge; Switched system; Ultracapacitors
• ISSN: 2352-152X; 2352-1538
• DOI: 10.1016/j.est.2024.112170

The Kalman filter is a key method for the State-of-Charge (SoC) estimation in ultracapacitors. However, most existing studies limit themselves to the SoC estimation of individual ultracapacitor cells, overlooking the practical scenario where ultracapacitors are often coupled with balancing circuits. In this paper, a switching Kalman filtering method is proposed for the SoC estimation of ultracapacitor with balancing circuits. We build a linear switched system model by combining the equivalent circuit model of a ultracapacitor and a switching resistor balance circuit. The observability of the switched system is shown and its discrete-time form is given. Finally, a switching Kalman filter is designed to estimate the SoC of the switched system, in which the observer parameters can be dynamically switched according to the system model. The experimental platform of the ultracapacitor balance system is built by using the control board and the environment chamber. Extensive experimental validation is performed on the experimental platform. It is demonstrated that this method significantly improves the accuracy of SoC estimation when the ultracapacitor is operated in the balancing mode, and the accuracy is improved by at least 3.756% compared to the normal method. •We introduce a linear switched system model incorporating the equivalent circuit of an ultracapacitor, to effectively characterize its dynamic behavior when integrated with a balancing circuit.•A novel switching Kalman filter approach is proposed to estimate the ultracapacitor’s SoC. This method accounts for changes in the switching states within the balancing circuit, enhancing the accuracy of SoC estimations.•Extensive experiments are conducted across various charging and discharging currents, as well as different temperatures, utilizing a physical platform. The reliability and accuracy of the proposed observer inSoC estimation is also critically examined.