Power quality disturbance classification based on adaptive neuro-fuzzy system

This paper presents an adaptive neuro-fuzzy inference system and a set of novel features for classification of power quality disturbances. The most common types of disturbances including flickers, harmonics, impulses, notches, outages, sags, swells, and switching transients are considered in this research. The proposed method employs voltage waveforms for analysis. The features are extracted utilizing the signal processing techniques such as the windowed discrete Fourier transform and S-transform. Evaluation studies based on both simulated and field data are reported.