NEURO-EDGE AI FOR SMART GRIDS: A LOW-LATENCY FRAMEWORK FOR REAL-TIME FAULT DETECTION, LOAD FORECASTING, AND ENERGY OPTIMIZATION IN DISTRIBUTED POWER SYSTEMS

The evolution of distributed power systems necessitates intelligent, low-latency frameworks for optimizing operations in real-time. This study introduces a Neuro-Edge AI framework that integrates neuromorphic computing, federated learning, and explainable diagnostics to enable real-time fault detection, accurate load forecasting, and efficient energy optimization in smart grids. The system is designed for edge-controlled environments, enhancing the autonomy and adaptability of electrical infrastructure. Our proposed architecture prioritizes data locality and minimal latency, allowing decentralized intelligence and predictive capabilities to be embedded at the grid edge. Experimental evaluations demonstrate improved fault detection accuracy (94.6%), forecasting precision (92.1%), and energy optimization (89.8%) compared to traditional models. The framework promotes resilient, sustainable, and secure operations across distributed energy networks.