VoltSecure: A Secure Federated Learning Model for Decentralized Energy Management Systems

The evolution of decentralized energy management systems (EMS) has been influenced by the integration of smart devices along with distributed energy resources (DERs) into modern power grids. Although these systems improve flexibility and efficiency, they create large volumes of sensitive operational data, which are of primary concern for privacy, security, and scalability. Managing and forecasting energy consumption using traditional centralized machine learning approaches tend to encounter latency, data siloing, and increased vulnerability to cyber threats. In this paper, we propose VoltSecure, a decentralized smart grid EMS optimized for weakly connected environments, which implements FL in a privacy-preserving context. VoltSecure permits the training of collaborative models at different EMS nodes without transferring raw data, thus enhancing privacy while diminishing the communication burden. The framework provides local model updates containing sensitive information with benchmarks that make them vulnerable to detection and collapse claims when combined with nimble participant data. It is also designed with strong aggregation methods that improve resistance against these advanced targeted attacks. The system designed for edge deployment achieves real-time responsiveness and scalability to heterogeneous power domains. The performance evaluation of VoltSecure is detailed based on extensive real-world energy consumption datasets. The findings show that VoltSecure balances high prediction accuracy and privacy guarantees while also showing considerable strength against different adversarial attacks and data fragmentation degrees. Also, the analysis of communication costs confirms the model's efficiency, particularly concerning limited bandwidth. VoltSecure significantly enhances the capabilities of next-generation smart grids by enabling secure, privacy-preserving, and scalable energy intelligence. Future research will focus on the association of blockchain technology with auditability, participatory incentives for client retention, and adaptation during dynamic shifts in real-time grid conditions. Through the integration of federated learning with strong security measures, VoltSecure provides the foundation for reliable and intelligent decentralized energy systems.