Year: 2026
Type: article
Abstract: The fast incorporation of renewable energy into the modern power systems like solar and wind has posed an enormous challenge on the management of smart grid like intermittency in energy, as well as, demand variability, and grid stability problems. All of these dynamic conditions can be particularly difficult to respond to with the traditional centralized energy management approaches, which have significant latency, cannot scale effectively, and have communication overhead. The proposed study aims to overcome these issues by introducing a distributed Edge-Artificial Intelligence (Edge-AI) system that can be used to control renewable energy production in smart grids through reinforcement learning (RL). The framework proposed uses smart edge nodes that can conduct local data analysis and real-time decision making on optimal allocation of energy. The system architecture incorporates renewable energy sources, energy storage systems, smart metres, edge computing units, and a grid control interface, which can be used to coordinate the distributed resources of energy in a decentralised manner. Reinforcement learning model is used to optimise dynamically energy distribution, load balancing and storage utilisation under different generation and demand conditions. The efficiency of the suggested framework will be measured by the most important energy management indicators such as the rate of energy efficiency improvement, the rate of using renewable energy, and the rate of reducing power losses. Experimental findings prove that the distributed Edge-AI system brings multiple benefits of higher use of renewable energy, less transmission losses, and better stability of the entire grid. Moreover, the reinforcement-reward learning-controlled mechanism creates adaptive real-time decision making which suggests the possibility of the Edge-AI technologies in the development of the efficient, resilient and autonomous smart grid energy management system.
Author: T M Sathish Kumar
Language: English
Topic: Smart Grid Energy Management
Subfield: Electrical and Electronic Engineering
Field: Engineering
Domain: Physical Sciences
Open Access status: diamond
OpenAlex ID: W7220225156