基于工况识别的燃料电池公交车能量管理策略

为解决燃料电池混合动力公交车中基于优化的能量管理策略难以实车应用的问题,在分析燃料电池公交车(Fuel cell hybrid bus,FCHB)行驶路线的固定性和片段性的基础上,提出了一种基于SOM-K-means(Self-organized mapping K-means)工况识别的能量管理策略。首先,根据公交车站点将行驶路线划分为多个行驶片段,在车辆停站时,运用SOM-K-means二阶聚类模型完成工况识别,获取车辆下一行驶片段的识别协态变量；当车辆在下一个行驶片段运行时,运用识别协态变量完成基于庞特里亚金极值原理(Pontryagins maximum principle,PMP)求解的能量管理策略的实时应用。其次,建立基于公交车实际运行数据的仿真实验,最后建立硬件在环实验,将所提出的策略移植入整车控制器(Vehicle control unit,VCU)中进行实验。实验结果表明,与基于规则的能量管理策略相比,本研究提出的能量管理策略降低了19.77%的平均等效氢气消耗。且该策略在VCU中每一步的计算时间大约为30 ms,计算结果与仿真结果完全一致,满足车辆对能量管理策略的时效性和准确性的要求。

Energy management strategy of fuel cell bus based on working condition identification

To solve the problem that energy management strategy based on optimization in fuel cell hybrid electric buses is difficult to apply to real life vehicles, an energy management strategy based on SOM-K-means driving condition identification is proposed with reference to the analysis of the fixedness and fragmentation of the fuel cell bus (FCHB) driving route. Firstly, the driving route is divided into driving segments according to bus stops. When the vehicle stops, the SOM-K-means second-order clustering model is used to identify the driving condition, and obtain the predictive co-state of the next driving segment. When the vehicle runs in the next driving segment, a predictive co-state is used to complete the real-time application of the minimum fuel equivalent fuel consumption strategy based on the PMP solution. Secondly, the simulation experiments based on the actual driving data of the bus are established. Finally, the proposed strategy is applied to the vehicle control unit (VCU). The results show that compared with the rule-based strategy, the proposed strategy reduces hydrogen consumption by 19.77%. The calculation time of each step in the VCU is about 30 ms, and the calculation results prove to be completely consistent with the simulation results, meeting the requirements of vehicle for the timeliness and accuracy of energy management strategy.