Flicker Source Detection Including Fixed Speed Wind Turbines Using Empirical Mode Decomposition

This study describes an approach to identify multiple flicker sources at the point of common coupling (PCC). The voltage signals different flicker such as electrical arc furnace, fixed-speed wind turbine, and diesel-engine driven generator were recorded PCC. For this purpose, various aerodynamic mechanical faults a turbine shear tower shadow, gearbox tooth-breaking, blade crash, pitch angle error misfiring, exciter, governor error, are considered. After acquiring faults, empirical mode decomposition (EMD) robust signal processing technique for extracting useful features was used. Then, reducing required memory space computational burden, minimal-redundancy-maximal-relevance (MRMR) symmetric uncertainty (SU) feature selection methods applied. Also, increasing efficiency methods, cooperative game-theoretic method utilized. Afterward, two classifiers based on Naive-Bayes support vector machine (SVM) used detect faults. Simulation results presented validate effectiveness proposed method.