期刊名称：International Journal on Electrical Engineering and Informatics
印刷版ISSN：2085-6830
出版年度：2018
卷号：10
期号：2
页码：203-219
DOI：10.15676/ijeei.2018.10.2.1
出版社：School of Electrical Engineering and Informatics
摘要：This article presents an advanced continuous wavelet transform (CWT) basedapproach for fault detection and localization in distribution systems using the artificial neuralnetwork (ANN). In this study, CWT extracts distinct features from the transient signalscaptured from the bus. The derived features are utilized to train and test appropriate ANNarchitecture in different stages to detect and localize the faults. The proposed schemeprovides an optimum method for classification as well as localization of the various kinds offault with different source short circuit (SSC) level in different locations. The whole detectionand localization process consists of several stages. In the first stage, it detects faulty feeder.The faulty line is identified in the second stage. Finally, in the third stage, fault type and faultlocation are being calculated from the relaying point. The performance of the proposed CWTANNbased approach is quite promising as compared to traditionally used algorithms.However, a correlation-based feature selection technique is also implemented to reducetraining time and improve accuracy. This algorithm is tested in 11 kV radial Indiandistribution network but can be applied in other distribution networks also.
关键词：Continuous Wavelet Transform; Artificial Neural Network; Fault localization;Fault Detection; Unsymmetrical fault; Distribution system
