Peer-to-peer multi-energy sharing for home microgrids: An integration of data-driven and model-driven approaches

To form the economic and sustainable communities, peer-to-peer energy sharing is becoming one of the most promising ways for coordinating home microgrids. This paper presents a double auction-based peer-to-peer multi-energy sharing mechanism, to achieve the coordination of the intelligent, self-interested, and privacy-conscious home microgrids, and conduct the electricity sharing and heat sharing simultaneously. Based on a combination of model-driven optimization and data-driven prediction, an integrated model is developed, which can analyze historical transaction data, excavate the hybrid market trading rules, accurately and stably predict energy sharing strategies, and efficiently optimize the operation strategies of uncontrollable and controllable equipment. A hybrid genetic algorithm-extreme learning machine approach is proposed to realize data-driven prediction. According to the cost minimization objective, a joint optimization of the home microgrid elec-tricity and heat energy sharing strategies and distributed energy operation strategy is conducted, to facilitate the energy interaction among microgrids and enhance the coordination between electricity and heat systems. In addition, we have evaluated the accuracy and stability of the genetic algorithm-extreme learning machine model and proved the validity of the model. Numerical results demonstrate that the proposed model can efficiently solve the multi-energy sharing problem, obtain high cost savings, improve the independence of microgrids, realize complementary advantages of multi-energy, and address the complicated relationships among the peers, which are regarded as stakeholders with separate privacy and interests.

Automated summary

The paper introduces a double auction-based peer-to-peer multi-energy sharing mechanism, which utilizes an integrated model and data-driven prediction to optimize coordination of home microgrids, efficiently solve multi-energy sharing problems, and enhance the independence of microgrids.