Abstract (EN):
This paper presents the effect of environmental factors, such as wind speed change, on the classification of power quality (PQ) disturbances in grid-connected wind energy systems. Initially, based on the selection of suitable features and 3-Dimensional feature plots, the PQ disturbances are classified. Further, the disturbances are accurately classified using S-transform based feature extraction followed by classification by modular probabilistic neural network (MPNN), support vector machines (SVMs) and least square support vector machines (LS-SVMs). Different types of sag and swell disturbances due to the change in load and wind speed are created using MATLAB/Simulink and an experimental prototype setup for the classification problem. The results reveal that S-transform based extracted feature data, when trained with MPNN, SVMs and LS-SVM, can effectively classify the PQ disturbances. The accuracy and reliability of the proposed classifier are also validated on signals with noise content. A comparative study is also carried out to determine the robustness of the techniques used. Finally, conclusions are duly drawn.
Language:
English
Type (Professor's evaluation):
Scientific