数据驱动的插电式混合动力客车能量管理策略

Data-Driven Optimal Energy Management Strategy for Plug-in Hybrid Electric Buses

针对插电式混合动力公交客车的能量管理优化问题,提出了基于庞特里亚金极小值原理的两类工况片段新特征参数,其中非平衡状态特征参数与转速-扭矩分布相关,平衡状态特征参数与最优控制相关,并由此构建了行驶工况数据库.针对影响该类车辆能耗的关键阶段——电能平衡阶段,通过工况间的相似性,提出了含修正的数据驱动的最优等价因子估计方法,构建了数据驱动的能量优化管理策略.结果表明,该策略能使不同实际工况采用的等价因子和电池荷电状态SOC轨迹接近于理论最优,较目前常用的规则策略能显著地提升燃油经济性约5.72%,较同类研究能克服准确预测车辆未来行驶工况的难题.

Two classes of new characteristic parameters based on Pontryagin＇s minimum principle were proposed for the energy management optimization of plug-in hybrid electric buses and were used to build a driving condition database,in which non-equilibrium state characteristic parameters were related to the speed-torque distribution and equilibrium state characteristic parameters were related to the optimal control.A data-driven optimal equivalence factor estimation method with correction capability was also proposed and a data-driven energy management strategy was built through the driving condition similarity in the stage of charge sustaining,which is a key stage for the fuel consumption of buses.The results show that the strategy can make the equivalence factor and SOC（state of change）trajectory of different trips be close to the optimal.The new strategy improves the fuel economy by about 5.72% comparing with the rule-based strategy,and it can overcome the difficulty in accurately predicting future driving conditions.