A combinatorial auction energy trading approach for VPPs consisting of interconnected microgrids in demand-side ancillary services market

This work presents a two-stage optimization model for virtual power plants (VPPs) consisting of interconnected microgrids (IMGs) with different load profiles, to provide demand-side frequency control ancillary services (D-FCAS). In the day-ahead scheduling stage, the hourly power consumption baseline and regulation capacity of the next day are predicted and rewarded by the market operator. In the real-time control stage, the dynamic regulation (RegD) signal from the grid is distributed according to the electricity unit price of each microgrid (MG), followed by a novel peer-to-peer (P2P) energy trading mechanism among the MGs to reduce RegD-following violation of the VPP. Particularly, three P2P methods are designed and compared for P2P energy trading: (i) a multi-leader and multi-follower (MLMF) Stackelberg game where the buyers are regarded as the leaders; (ii) a MLMF game where the sellers are the leaders; and (iii) a combinatorial auction mechanism (CAM) where buyers and sellers bid simultaneously. After the P2P energy trading is settled, the real-time power consumption control is coordinated locally inside each MG, to maximize the gross revenue of VPP. Particle swarm optimization (PSO) is innovatively employed to obtain the optimal energy exchange solution in the P2P energy trading process, which is imported to each MG for local power control. Extensive simulations and comparative studies are conducted, and the numerical results show that the VPP with CAM P2P energy trading can reduce operating costs through superior RegD tracking performance compared to the VPP with other or no energy trading methods.