Data-driven hierarchical control for online energy management of plug-in hybrid electric city bus

The pre-determined city bus routes and the availability of partial-trip information obtained through vehicular connectivity provides new opportunities for plug-in vehicles to plan electric energy reasonably. This paper presents a data-driven hierarchical control method for online energy management of plug-in hybrid electric city buses, which can learn from globally optimal solutions based on historical accumulated cycles while taking advantage of connectivity-enabled partial-trip information. The devised scheme comprises two levels of control modules. The upper battery state-of-charge planner trained using historical optimal data is employed for deriving a reference state-of-charge based on the current battery state, remaining trip length, and low/high speed ratios. The lower powertrain controller is then applied to regulate the engine operation according to the reference state-of-charge and powertrain states. This article presents two contributions: (1) both accumulated historical optimal data and partial-trip information are assimilated to augment the applicability of the control hierarchy, thus achieving better resilience to “unseen” driving patterns; (2) given limited resources of micro-controllers, the control strategy is proven to be a real-time implementable, close-to-optimal solution. A variety of results show that the proposed approach can achieve significant fuel savings (4.99%-14.80%) as compared to the charge depleting and charge sustaining strategy.