Bi-level optimal sizing and energy management of hybrid electric propulsion systems. (15th February 2020)

• Record Type: Journal Article
• Authors: Zhu, Jianyun; Chen, Li; Wang, Xuefeng; Yu, Long
• Is Part Of: Applied energy. Volume 260(2020)
• Publication Date: 2020-02-15
• Subjects: Multiobjective bi-level optimization -- Hybrid electric propulsion system -- Modified adaptive equivalent consumption minimization strategy -- Fuel consumption -- Greenhouse gas emission -- Net present cost; Power (Mechanics) -- Periodicals; Energy conservation -- Periodicals; Energy conversion -- Periodicals; 621.042
• DOI: 10.1016/j.apenergy.2019.114134
• Languages: English
• ISSNs: 0306-2619

Abstract

Highlights: A bi-level method is proposed for the optimization of a hybrid electric propulsion system. Optimal component sizing and energy management are implemented simultaneously. A modified adaptive equivalent consumption minimization strategy is developed. The proposed bi-level method outperforms single-level optimizations. Abstract: Hybrid electric propulsion systems attract considerable research interest because of their potential to reduce fuel consumption, greenhouse gas emission, and net present cost. However, independent optimization for component sizing or energy management may lead to performance degradation. The present study proposes a multiobjective bi-level optimization that performs component sizing at the upper level and energy management at the lower level simultaneously. Multiobjective particle swarm optimization is developed for the upper level because of its merits in computational time and generational distance. An adaptive equivalent consumption minimization strategy, which has a light computational load, has been modified for the lower level by updating the equivalence factor based on the battery stage of charge and engine efficiency. Real-time hardware-in-the-loop experiments are carried out to validate the effectiveness of the optimization. The results of the proposed bi-level optimization are compared with two independent single-level optimizations. The optimal solution of the proposed method is significantly superior to the single-level optimizations. Furthermore, the result of the single-lower-level optimization is closer to that of the bi-level optimization than that of the single-upper-level optimization.