Privacy-preserving economic dispatch for multi-energy virtual power plants: A distributed approach against Sybil attacks

Driven by advances in multi-energy coordination technologies, virtual power plants are transitioning toward multi-energy virtual power plants (MEVPP) that integrate diverse distributed energy resources. In terms of data communication architectures within MEVPP systems, the hierarchical structure imposes high computational demands on the system operator, whereas the peer-to-peer architecture exposes energy agents to risks of privacy leakage and cyber-attacks, particularly Sybil attacks at the information layer. This paper initially designs a fully distributed MEVPP framework that incorporates distributed heterogeneous multi-energy resources and develops a multi-agent day-ahead economic dispatch framework. We introduce a novel integration of the distributed collaborative neurodynamic optimization scheme with a state-decomposition-based communication protocol, and provide a reputation-based defense mechanism for the potential Sybil attacks in the MEVPPs. The proposed method ensures both computational efficiency and communication security, and its effectiveness is validated through rigorous mathematical analysis and comprehensive case studies.