A real-time PMP energy management strategy for fuel cell hybrid buses based on driving segment feature recognition. (18th November 2021)

• Record Type: Journal Article
• Authors: Sun, Xiaoxiao; Zhou, Yafu; Huang, Lijian; Lian, Jing
• Is Part Of: International journal of hydrogen energy. Volume 46:Number 80(2021)
• Page Start: 39983
• Page End: 40000
• Publication Date: 2021-11-18
• Subjects: Fuel cells -- Fuel cell hybrid bus -- Energy management strategy -- Pontryagin's minimum principle -- Driving feature recognition; Hydrogen as fuel -- Periodicals; Hydrogène (Combustible) -- Périodiques; Hydrogen as fuel; Periodicals; 665.81
• DOI: 10.1016/j.ijhydene.2021.09.204
• Languages: English
• ISSNs: 0360-3199

Abstract

Abstract: Establishing a reasonable energy management strategy (EMS) is the key to improve the service durability, power performance and fuel economy of the fuel cell hybrid electric vehicle (FCHEV). This paper obtains energy distribution optimal solution for the fuel cell hybrid bus (FCHB) based on Pontryagin's minimum principle (PMP) algorithm, and the problems of inaccurate estimation of motor power and difficult real-time application are solved. Firstly, the driving feature recognition is completed by collecting the motor output power directly when the FCHB stops at the station. On the basis of it, the sub-optimal co-state value is chosen. Secondly, the sub-optimal co-state is used to complete the real-time application of PMP algorithm in the driving segment. The results are acquired through the simulation and the actual comparison experiment, compared with rule-based simulation and rule-based actual experiment, the hydrogen consumption of the proposed strategy decreases by 20.3% and 28.9% on average. Moreover, the online computation time per step of the proposed strategy is 3.64 ms averagely, less than sampling time interval 1s. It is shown that the proposed method has lower hydrogen consumption rate and excellent real-time performance. Graphical abstract: Image 1 Highlights: A method of driving feature recognition based on DBSCAN algorithm is proposed. A real-time PMP EMS based on bus driving feature recognition is proposed. The proposed EMS can realize real-time optimization without known road condition. Hydrogen consumption is close to that of known road condition PMP strategy.