Adaptive Embedded Control System for Real-Time Power Quality Improvement Using Hybrid Ai Techniques

Abstract

ABSTRACT: Power quality degradation—including voltage fluctuations, harmonic distortions, and transient events—poses significant operational challenges in modern electrical networks. This paper proposes an adaptive embedded control system that utilizes hybrid AI techniques to monitor and improve power quality in real time. The system integrates convolutional neural networks (CNNs) for pattern recognition, genetic algorithms for optimization, and adaptive filtering algorithms for control actuation. These components are embedded within a compact microcontroller-based platform capable of high-frequency sampling and ultrafast response. Experimental validation using a real-time digital simulator (RTDS) demonstrates that the system efficiently identifies disturbances and executes corrective strategies within milliseconds. The hybrid AI approach enhances adaptability, allowing the system to maintain performance even when encountering previously unseen disturbance patterns. Results indicate substantial reductions in total harmonic distortion and improved voltage regulation across various load conditions.

KEYWORDS: Power quality, Adaptive control, Hybrid AI, Embedded systems, Harmonic mitigation