A Comprehensive Review of Detection and Classification Techniques for Power Quality Disturbances in Micro-Grid Systems

Abstract

The quality upsets of power quality are major operation concerns of Micro-Grid systems due to the increasing use of renewable sources of power, power electronic converters, and dynamic loads. despite the variety of signal-processing, machine-learning, and deep-learning methods that have been suggested regarding disturbance detection and classification, the literature is still disjointed in terms of datasets, evaluation criteria, and real-time implementation. The review is a scientific study of the literature on the subject of power quality disturbance monitoring in micro-grids through a prisma-guided methodology, and the reviews were refined according to the established inclusion and exclusion criteria. The selected works are compared in terms of the nature of disturbance, the tool of extracting the features, the classifier, the dataset, the performance evaluation, the computer complexities, and the capability to implement it. The review shows that hybrid and deep-learning-based models are highly classified, yet limited by the problem of data imbalance, unavailability of real-life three-phase datasets, low noise tolerance, and low interpretability. It is on this synthesis that the paper reflects the key gaps in the research and recommends the roadmap of the future of robust, real-time, and scalable micro-grid power quality monitoring.