INTELLIGENT AND HYBRID AI-BASED FAULT DETECTION FRAMEWORK FOR REAL-TIME MONITORING IN INVERTERDOMINATED MICROGRIDS

Abstract

The evolution of microgrids toward renewable energy integration has led to increased deployment of inverterdominated and low-inertia systems. Traditional fault detection mechanisms, tailored for conventional grids, struggle under these new dynamics. This paper proposes an intelligent and hybrid AI-based framework incorporating Wavelet Transform and Convolutional Neural Networks (WT-CNN) alongside Support Vector Machine–Fuzzy Logic (SVM-FL) models. The framework aims to achieve high-speed, adaptive, and accurate fault detection and classification in real time. Implemented using MATLAB/Simulink and tested under varying operating conditions, the model demonstrates superior performance in detection latency, classification accuracy, and robustness against system disturbances.