Cloud-supported energy management strategy for a hybrid electric heavy-duty commercial vehicle considering dynamic mass and upper-part load

The power demand of heavy-duty commercial vehicles varies significantly with dynamic mass and upper-part load, which should be considered in energy management strategy (EMS) development. This paper thus proposes a two-level energy optimization framework that incorporates these factors. The cloud level generates EMSs from multiple vehicles’ operational data, while the vehicle level conducts energy management and power distribution. The periodic transfer, strategy verification, and updating between them enhance EMS effectiveness and reduce communication dependence. By designing a mass identifier and analyzing the upper-part load power demand, the two factors are incorporated into the EMS based on a twin delayed deep deterministic policy gradient algorithm. Simulations show the proposed EMS attains an average 98.26% fuel optimality of dynamic programming, and an average 19.59% fuel economy improvement against the rule-based EMS. Bench test indicates the power-following capability of the auxiliary power unit, thereby grounding the practicality of the proposed EMS.