Adaptive energy management of a battery-supercapacitor energy storage system for electric vehicles based on flexible perception and neural network fitting. (15th June 2021)

• Record Type: Journal Article
• Authors: Zhu, Tao; Wills, Richard G.A.; Lot, Roberto; Ruan, Haijun; Jiang, Zhihao
• Is Part Of: Applied energy. Volume 292(2021)
• Publication Date: 2021-06-15
• Subjects: Electric vehicle -- Hybrid energy storage system -- Energy management -- Cost optimisation -- Variable perception horizon -- Neural network fitting; Power (Mechanics) -- Periodicals; Energy conservation -- Periodicals; Energy conversion -- Periodicals; 621.042
• DOI: 10.1016/j.apenergy.2021.116932
• Languages: English
• ISSNs: 0306-2619

Abstract

Graphical abstract: Highlights: Proposed online EMS imitates the optimal offline EMS to reduce online complexity. Variable perception horizon considers micro-trip singularity and EMS update frequency. EV driving states that impact EMS design are analysed and refined for online prediction. Neural network offers new rules from real-time driving states and predefined rules. Proposed online EMS can realise more than 97% cost optimisation of offline benchmark. Abstract: The hybrid energy storage system (HESS) composed of batteries and supercapacitors (SCs) is a dual energy storage technology that can compensate for the shortcomings of a single energy storage technology acting alone. The energy management of HESS splits the power and energy demands from the electric vehicle (EV) to the battery and SC and thus is vital to EV propulsion. This paper presents an online energy management strategy (EMS) that optimises the operating costs of battery-SC HESS and can be adaptive to real-time EV driving conditions. We analyse the optimal offline benchmarks to guide online EMS design and propose the adaptive online EMS with variable perception horizon based on both neural network and rule-based techniques. Compared with existing research, the proposed EMS features reduced complexity, flexible perception and intelligent rulemaking. Case study results show that the proposed variable perception horizon and neural network fitting can improve EMS optimality compared with the conventional methods in existing research. The proposed EMS can realise more than 97% cost optimisation efficacy of offline benchmarks. By the proposed EMS, this paper is expected to provide a practical and effective energy management approach for the battery-SC HESS to reduce costs in EV applications.