Distributed Voltage Regulation for Distribution Networks With Privacy-Preserving Under the Framework of VPP

Abstract

Global concern about climate change has accelerated the integration of renewable energy. To accommodate the high penetration of renewables at the distribution level and maintain system flexibility under a fully distributed architecture, this paper develops a voltage control strategy based on federated learning coordinated by a virtual power plant. A dynamic network partitioning method is introduced using a comprehensive performance index, along with an adaptive genetic algorithm featuring elite retention. An enhanced alternating direction method of multipliers with adaptive penalty modulation is employed to improve the convergence efficiency. Additionally, a two-stage encryption mechanism is applied to protect user privacy and ensure cybersecurity during distributed coordination. The effectiveness and feasibility of the proposed method are validated on a modified IEEE 33-bus system.