Feature vector extraction by using empirical mode decomposition for power quality disturbances

This work presents a relatively new method known as empirical mode decomposition (EMD) for power quality disturbances. In a comprehensive and wider range of approaches and engineering activities, there is a increasing concern for power system disturbances monitoring techniques. The need of increasing performances in terms of accuracy and computation speed is permanently demanding new efficient processing techniques on power system visualization. For system monitoring, feature extraction of a disturbed power signal provides information that helps to detect and diagnose the responsible fault for power quality disturbance. Traditionally, monitoring spectral and harmonic analysis of dynamic systems is based on Fourier based transforms and the wavelets. The Fourier transform usually has been used in the past for analysis of stationary and periodic signals. Qualification to providing a more accurate `real-time' demonstration of a signal without any artifacts imposed by the non-locally adaptive limitations of the fast Fourier transform (FFT) and wavelet processing. In this work, the first step of Hilbert-Huang transform (HHT), EMD, has been regarded as a powerful tool for adaptive analysis of non-linear and non-stationary signals.