AI DRIVEN POWER QUALITY ANALYSIS IN DISTRIBUTION GRIDS

The increasing integration of distributed generation (DG) and power-electronics-based equipment has intensified power quality (PQ) disturbances such as harmonics, voltage fluctuations, and transients in distribution grids. While artificial intelligence (AI) has demonstrated strong potential for PQ monitoring and classification, existing research remains focused mainly on signal analysis rather than intelligent control of renewable energy converters for active PQ enhancement. This review provides a comprehensive synthesis of AI-driven approaches for PQ detection, classification, prediction, and mitigation. A practical low-cost measurement framework is also presented, utilizing an ATmega328 microcontroller interfaced with current and voltage transformers (CT, VT) for PQ data acquisition. The processed parameters are transmitted via ESP8266 Wi-Fi to a cloud platform such as ThingSpeak for real-time visualization and storage. MATLAB further supports data analytics, AI-based classification, and harmonic analysis, enabling improved decision-making for PQ management. By comparing machine learning, deep learning, and hybrid AI models, this study evaluates accuracy, response time, and total harmonic distortion (THD) reduction. The findings highlight a major research gap: the limited use of AI for adaptive control of renewable energy converters to directly enhance PQ. Future directions are proposed toward self-optimizing and intelligent PQ management in evolving smart grids. The system uses a low-cost microcontroller and an open-source cloud platform, reducing system cost by over 60% compared to commercial PQ analyzers. The framework also integrates reinforcement learning for autonomous control actions and explores digital twin technology for virtual PQ prediction before faults occur. This study bridges the gap between theoretical AI models and practical embedded deployment for PQ monitoring. Unlike existing survey papers, this work integrates hardware implementation with intelligent analytics.