LAUSR

This paper proposes a new signal segmentation method, reduced-sample empirical mode decomposition, to extract the highly correlated monocomponent mode of oscillations. The two efficacious power quality indices are extracted from the Hilbert transformed (HT) array of the first three intrinsic mode functions. A novel class-specific weighted random vector functional link network (CSWRVFLN) classifier is proposed to recognize the complex power quality disturbances (PQDs). The performance of reduced sample Hilbert–Huang transform (RSHHT) combined with CSWRVFLN (RSHHT-CSWRVFLN) method is tested and compared with tunable-Q Wavelet transform associated with HT and CSWRVFLN and empirical wavelet transform along with HT and CSWRVFLN methods. The short event detection, lesser computational complexity, superior classification accuracy, and robust antinoise performance are the major advantages of the proposed RSHHT-CSWRVFLN method. Furthermore, a field-programmable gate array embedded processor is used to test and validate the feasibility of the proposed method for online monitoring the PQDs.