Loading...

A Consensus Optimization Mechanism in Distributed Energy Systems

Babaeinejad, Omid | 2025

102 Viewed
  1. Type of Document: M.Sc. Thesis
  2. Language: Farsi
  3. Document No: 58817 (05)
  4. University: Sharif University of Technology
  5. Department: Electrical Engineering
  6. Advisor(s): Farhadi, Alireza; Sharifzadeh, Mahdi
  7. Abstract:
  8. This thesis investigates a consensus-based optimization process in distributed energy systems, focusing on solving the optimal power flow problem. In this study, a comprehensive framework is designed using a decentralized approach to improve the efficiency of energy distribution while maintaining security and economic stability. This method employs federated learning to preserve data privacy across regions and reinforcement learning to achieve guaranteed global optimality, as opposed to optimality limited to specific regions. The proposed algorithm is designed such that all components of the system, despite operating independently, reach a globally optimal solution with minimum cost. The system also considers data privacy and ensures that each agent shares only the essential information located at the boundaries between regions. The IEEE 30-Bus test network is used to evaluate this process. The obtained results indicate that the combination of consensus mechanisms with federated reinforcement learning provides a scalable and efficient approach for managing energy systems with a high penetration of distributed energy resources. The proposed framework, while reducing the need for extensive data exchange and preserving the decision-making autonomy of regions, is capable of adapting to multi-agent and decentralized structures of modern energy networks. Therefore, the proposed method can be used as a practical solution for the development of intelligent, secure, and economically efficient energy systems
  9. Keywords:
  10. Consensus-Based Optimization ; Optimal Power Flow ; Reinforcement Learning ; Federated Learning ; Energy Management ; Distributed Energy Systems

 Digital Object List

 Bookmark

No TOC