Predictive Battery State of Charge Reference Generation Using Basic Route Information for Optimal Energy Management of Heavy-Duty Fuel Cell Vehicles

One of the main challenges for the energy management of heavy-duty fuel cell/battery hybrid vehicles is to control the battery charge within safe operating levels without hindering fuel efficiency. This article proposes a novel method that uses basic route information from navigation systems to implement an optimal and predictive energy management scheme to improve battery charge control, fuel economy, and fuel cell system lifetime. Indeed, it is possible to obtain a rough estimate of the electric load profile using elevation and speed forecasts and optimize the power split before starting the trip. The result of this offline optimization is saved as a location-based battery state of charge (SoC) reference for the online power-split strategy. The novelty of this work is related to the formulation of a quadratic program to generate the predictive SoC reference by optimizing the hydrogen economy while respecting the battery charge constraints. Afterward, a simple but effective online energy management strategy is used to track the predictive references during the simulations. The use of real-world driving cycles is essential to demonstrate the benefits of the proposed method. In particular, it is critical to consider the road elevation because it highly influences the power demand of heavy-duty vehicles. The simulation results prove that using the predictive SoC reference significantly improves the hydrogen economy and battery charge control compared to a non-predictive energy management scheme. In addition, the high-power operation time of the fuel cell system is reduced, resulting in slower system degradation compared to the non-predictive strategy. The effectiveness and robustness of the proposed method are tested against 1750 hours of real-world driving scenarios, considering uncertain speed and vehicle mass information for the reference generation. Statistical and global analyses of the results prove that the implementation of the predictive SoC reference generation method is highly beneficial for the performance of heavy-duty fuel cell vehicles.