Simultaneous Optimization of Real-Time Control Strategies and Powertrain Design for Fuel Cell Hybrid Vehicles

The successful introduction of low-carbon footprint and highly efficient fuel cell vehicles represents nowadays a key action to achieve sustainable mobility worldwide. The main technological barriers (i.e., market price, lifetime and performance) to be overcome justifies an increasing attention towards the development of mathematical tools featuring co-optimization capabilities, so as to adequately account for the strong interactions and mutual influence between design criteria and selected control strategies. This paper thus presents and discusses the integration of a comprehensive model of a generic FCHV architecture with a specifications independent control strategy within a modular constrained optimization algorithm, the latter conceived in such a way to simultaneously find the optimal FCHV powertrain design and real-time applicable control strategies. Suitable design and energy management criteria, accounting for also the impact of driving mission on proper management of available power sources, were selected. The proposed co-optimization procedure aims at determining the main powertrain design parameters (i.e., nominal fuel cell system power and battery pack energy density), as well as some key driving cycle-related information (i.e., power prediction time horizon), so as to enable effective adaptation of the specifications independent control strategy to the current design. The latter aspect is particularly contributing as it allows accounting for the impact of most frequent driving path on vehicle optimization, as well as ensuring proper cycle dependent energy management be performed in each driving mission. When applying the methodology to a specific driving cycle, it was possible demonstrating that the selected optimization variables allow exploiting to the maximum extent the known positive characteristics of fuel cell technology in conjunction with driving cycle features. Moreover, optimization outcomes indicate that range-extender designs, with degree of hybridization as high as 0.8, shall be preferred to maximize the benefits associated to easy-to-apply rule-based control strategies developed for efficient FCHV powertrain destined to a wide range of end-users.