Real-Time interaction of active distribution network and virtual microgrids: Market paradigm and data-driven stakeholder behavior analysis

In order to incorporate the independent Virtual Microgrids (VMGs) to the real-time operation of upstream active distribution network (ADN), an interactive dispatching model of VMGs and ADN is proposed, in which the downstream VMGs perform self-dispatching while trading both energy and ancillary service procurement to the Distribution System Operator (DSO). The bi-level bidding and market clearing model is modelled as a data-driven Multi-Agent Reinforcement Learning (MARL) with the solution of Win-or-Learn-Fast Policy Hill-Climbing (WoLF-PHC) algorithm, which is an online and fully-distributed training, enabling VMGs to dynamically update their bidding strategies based on previous market clearing results. VMGs would thereafter conduct the economic dispatching considering the conditional value-at-risk (CVaR) of penalties caused by the curtailment of renewables, load loss, and failure of providing energy or ancillary service to DSO. Finally, the evolutionary game theory (EGT) with replication dynamic equations (RDEs) is adopted to analyze the inherent dynamics of the proposed MARL driven by WoLF-PHC, revealing the relation between VMGs’ bidding strategy convergence and the trading paradigm. The case study validates the advancement of computational performance of WoLF-PHC compared with conventional Q-learning in the aspects of convergence and computation speed, and the impact of risk coefficient on the VMGs’ real-time dispatching strategies is also demonstrated.