Advanced signal processing techniques for multiclass disturbance detection and classification in microgrids

This study proposes the application of fuzzy assessment tree (FAT)-based short-time modified Hilbert transform (STMHT) as a new multiclass detection and classification technique, for a distributed generation (DG)-based microgrid. The time varying non-stationary power signal samples extracted near the target DG are initially de-noised by passing through the morphological median filter and then processed through the proposed STMHT technique for disturbance detection. Further based on the overlapping in the target attribute values, an FAT has been incorporated, which significantly classifies the different multiclass disturbances on a standard IEC microgrid model simulated in MATLAB/Simulink environment with highest precision in accuracy.