文章提出了一种自底向上的虚拟电厂(virtual power plant,VPP)自组织聚合运行调度方法,旨在通过分布式能源(distributed energy resources,DER)间的动态自组织聚合,降低调度过程中的调控量。首先,针对分布式能源的出力特性,以跟踪系数量化评估其与负荷的一致性水平,并给出相关优化调控模型,同时考虑DER具备的基础智能,作为自组织聚合的基础。其次,引入联盟博弈的框架,给出了聚合效用函数及其对应的利益分配机制,并以其作为依据确立了DER的自组织聚合条件。最后,基于Pareto规则提出“聚合-分裂”机制,促进虚拟电厂的有序进化,形成虚拟电厂自组织聚合运行策略。算例表明,所提方法可根据DER出力特性动态组合,其总体调控量与集中优化时相当,相比各DER独立优化则显著下降;其计算量和计算时间,与各DER独立优化时相当,较集中优化显著下降。

This paper presented a bottom-up self-organizing aggregation operation scheduling method for virtual power plants, aiming at reducing the amount of regulation in the process of scheduling by dynamic self-organization aggregation among distributed energy resources (DER). Firstly, according to the output characteristics of DER, the consistency between which and the load is evaluated with the tracking coefficient, and the relevant optimal control model is given. As the basis of self-organizing aggregation, the basic intelligence of DER is considered. Secondly, the framework of coalition formation games is introduced, and the aggregating utility function and its corresponding benefit sharing mechanism are proposed, on the basis of which the self-organizing aggregation conditions of DER are established. Finally, according to Pareto rules, the "merge-split" mechanism is proposed, which can promote the orderly evolution of virtual power plants and form a self-organizing aggregation operation scheduling(SAOS) for virtual power plants. The results show that SAOS can be constituted dynamically according to the output characteristics of DER. Its overall scheduling quantity is equivalent to that of centralized optimization, and is significantly reduced compared with that of independent optimization. The calculation amount and calculation time of SAOS are equivalent to those of independent optimization, and significantly lower than centralized optimization.