Research on the Application of Federated Learning with Privacy Preservation for Inter-Regional Grid-Storage Optimization
Summary
No summary yet. VIRAT can read this paper and write a plain-language summary of its aim, methods, key findings and limitations.
Summarise with VIRAT Free to try with your Science Catalogs account.Abstract
The increasing adoption of renewable energy and large-scale energy storage systems necessitates cross-regional grid and storage coordination, presenting significant challenges due to sensitive operational data and limitations of centralized optimization approaches. Federated learning (FL) offers a promising alternative by avoiding direct data exchange. However, traditional methods suffer from slow convergence, deteriorating fairness when data is not independent and identically distributed (iid), and gradient-based privacy leakage risks. To address these challenges, this paper proposes a privacy-preserving federated optimization framework. It combines federated learning with an ADMM-based decomposition to ensure physical feasibility, employs clustered client sampling and gradient projection (FedCSGP) to mitigate conflicts arising from data heterogeneity, and incorporates secure aggregation and differential privacy to provide formal privacy guarantees. Experiments on multi-region IEEE testbeds demonstrate that the proposed framework achieves performance close to centralized optimization while reducing communication overhead and improving fairness. It also exhibits strong resistance to adversarial inference, offering a practical and theoretically feasible pathway for secure, efficient, and equitable federated optimization in next-generation smart grids.
Fields of Study & Subjects
Privacy-Preserving Technologies in Data
Smart Grid Energy Management
Integrated Energy Systems Optimization
Publication details
- Journal / source
- International Conference on Electrical Automation and Artificial Intelligence (ICEAAI)
- Pages
- 1273–1277
- Publication date
- 16 January 2026
- Language
- EN