Energy Flows Balance of Electric‐Heating Interconnected Microgrids Under Grid‐Forming Converter Integration Based on Heterogeneous Graph Convolutional Network

ABSTRACT The combination of electric‐heating interconnected microgrids (EHIMs) with grid‐forming converters (GFMCs) offers a highly promising approach to improving energy management within decentralised energy systems. Despite its potential, precisely modelling and analysing energy flow in such systems remains a considerable challenge due to the variety of energy sources, storage technologies and control strategies involved. This paper presents a novel methodology for the unified calculation of energy flow within EHIMs, utilising heterogeneous graph convolutional networks (HGCNs). In this framework, the microgrid components are represented as a heterogeneous graph, where nodes correspond to various system elements, such as photovoltaic panels, wind generators, solar thermal systems and energy storage units, whereas the edges represent the energy interactions and exchanges between these elements. By adopting this graph‐based structure, the model is better equipped to handle the variability of renewable energy sources and the adaptable nature of grid‐forming converters. Extensive simulation results show that the HGCN‐based approach surpasses traditional methods in terms of both prediction precision and computational efficiency, providing a more scalable solution for analysing real‐time energy flows in EHIMs. Ultimately, the proposed framework holds great promise for significantly improving the stability, operational efficiency and resilience of future smart grid infrastructures.