Real-Time Anomaly Detection in Smart Grids Using Graph Neural Networks

Abstract

The increasing complexity and scale of smart grids necessitate efficient and accurate real-time anomaly detection mechanisms to ensure grid reliability and security. Traditional detection methods often fall short in capturing the complex spatial and temporal dependencies inherent in smart grid data. This paper proposes a novel approach leveraging Graph Neural Networks (GNNs) to detect anomalies in smart grids by modeling the grid as a graph where nodes represent measurement points and edges represent electrical or communication connections. Our method exploits the graph structure and temporal dynamics to identify anomalies such as faults and cyber-attacks with high accuracy and low latency. Experimental evaluations on real and synthetic datasets demonstrate that the proposed GNN-based framework outperforms conventional machine learning models, offering a scalable and effective solution for real-time anomaly detection in smart grids.