Privacy-Preserving Distributed Economic Dispatch Based on State Decomposition in Smart Grids

Consensus-based distributed economic dispatch (ED) algorithms are promising solutions to the smart grid ED problem. However, distributed algorithms necessitate data sharing among system nodes, potentially compromising node privacy. This study proposes a privacy-preserving method based on state decomposition. By introducing random noises, the shared data of a node is decomposed into several parts, and each neighbor of the node receives only a random subset of the original data, thereby increasing the difficulty of inferring the node’s privacy from its shared data. It is theoretically proved that the method can preserve privacy without compromising the algorithm optimality. Compared to existing works, the method simultaneously achieves privacy preservation, optimality, low cost, and does not require extra communication. Moreover, the method is developed in a challenging scenario, where adversaries know the connection weights between each pair of nodes, ensuring no privacy is disclosed even if weight information is leaked. The effectiveness is validated on the IEEE 57-bus and IEEE 300-bus systems through comparisons with existing privacy-preserving methods.