Skip to Main Content
A new method based on Generalized S-transform (GST) time-frequency analysis and decision-making tree support vector machines (DMT SVMs) classifier for identification of power quality disturbances (PQDs) is presented. Firstly, GST is introduced to analyze the typical PQDs, including the inter-harmonics, where a set of useful characteristics are extracted. Then 50 disturbance training samples are employed to construct the characteristics sets which are applied to train a multi-lay SVMs classifier. Finally, 500 testing PQDs samples are identified using the SVMs classifier, in which N kinds of PQDs are classified by N-1 turns. Results show that the proposed method could detect and classify the PQDs effectively. The classifier has an excellent performance on training speed and correct ratio.