A new signal processing technique for power system disturbance detection and classification

A prototype multiwavelet-based recognition and DS-technique for classifying power quality disturbances is implemented and tested under various events. First, the m ultiresolution -analysis of multiwavelet transform technique is used for the analysis of the power system disturbances. The features are extracted from multiwavelet transform at the different resolution levels make up knowledge base which is then compared to that of IEEE standards in Heuristic Classifier and Chi-square distribution factor is derived from the last resolution level in Statistical Classifier. DS-technique has been used for the decision making between the results of Heuristic Classifier and Statistical Classifier; for the recognition the power system disturbances. Various events are tested, such as, voltage sag, voltage swell, outage, interruption, oscillatory-impulsive, oscillatory-transient, noise and notching show that the classifier can detect and classify different power disturbance types efficiently. Five hundred samples are tested and their results are shown later on with help of the confusion table.