Detection of Flicker Caused by Wind Farm Based on Mathematical Morphology Filter and Hilbert-Huang Transform

Abstract

Voltage fluctuation and flicker are among the most serious impacts of wind power generation on power quality. To assess and suppress voltage fluctuation and flicker, an accurate detection method is necessary. According to characteristics of flicker caused by wind farm, a systematic, accurate and fast flicker detection method was introduced. First, with Mathematical Morphology (MM) filter, noises contained in flicker signal were filtered out. Then to the filtered flicker curves, a track method based on Hilbert transform was adopted to pick out the flickers envelope which is the base of attaining relevant parameters of flicker. At last, the envelope was simultaneously analyzed in both time-domain and frequency-domain by a novel non-stationary signal processing method, the Hilbert-Huang transform (HHT). Simulations show that with this method, time, frequency and amplitude information of non-stationary flicker signals can be accurately detected and the proposed method is effective in the analysis of flickers caused by wind farm.