Artificial Intelligence for Smart Grid Optimization: Review on Stability, Renewables, Forecasting, Power Quality, and Fault-Tolerant Operations
Year: 2025
Type: preprint
Abstract: The rapid growth of renewable energy integration introduces challenges of stability, reliability, and power quality in modern power systems. Artificial Intelligence (AI), including Machine Learning (ML) and Deep Learning (DL), has emerged as a key enabler of predictive stability assessment, renewable dispatch, and fault-tolerant power electronics. This paper presents a comprehensive review across five critical domains: smart grid stability, renewable energy integration, deep learning-based forecasting, fault-tolerant multilevel inverters, and machine learning for power quality enhancement. A taxonomy is developed to categorize AI applications, supported by comparative evaluation tables, bibliometric analysis, and maturity mapping. Cross-domain synthesis reveals that while hybrid AI models consistently outperform singlemethod approaches, real-world deployment remains constrained by scalability, computational complexity, and lack of standardized datasets. Current research is fragmented, with forecasting, control, and cybersecurity often studied in isolation rather than within unified frameworks. Future opportunities lie in explainable and adaptive AI architectures, hardware-in-the-loop validation, and integrated frameworks combining forecasting, fault tolerance, and cybersecurity. This review highlights research gaps and provides a roadmap for developing resilient, industry-ready AI solutions for renewable-dominated smart grids.
Author: Mehmil Zahra
Institution: University of Kashmir
Language: English
Topic: Smart Grid and Power Systems
Subfield: Electrical and Electronic Engineering
Field: Engineering
Domain: Physical Sciences
Open Access status: gold
OpenAlex ID: W4414520132
Cites:
Cited by:
Related to: