Application of discrete wavelet transform and back-propagation neural network for internal and external fault classification in transformer

Simple item page
dc.contributor.authorNgaopitakkul, Atthapol
dc.contributor.authorJettanasen, Chaiyan
dc.contributor.authorAsfani, Dimas Anton
dc.contributor.authorNegara, Yulistya
dc.date.accessioned2026-08-06T10:23:16Z
dc.date.available2026-08-06T10:23:16Z
dc.date.issued2019-01-01
dc.description.abstractThis paper proposes an algorithm for internal and external fault discrimination in the three-phase two-winding power transformer based on a combination of discrete wavelet transform (DWT) and back-propagation neural network (BPNN). The maximum ratio obtained from division algorithm between DWT coefficient value of differential current and zero sequence component in post-fault condition differential current signals is employed as an input for the training pattern for BPNN in order to discriminate between internal fault and external short circuit. The proposed algorithm performance has been test using various cases studies based on Thailand electricity transmission and distribution systems data. Results show that the proposed technique can achieved satisfy accuracy for internal and external fault detection and discrimination in the considered system. This methodology and result can be used to further improve protection system of power transformer in the future.
dc.identifier.citationInternational Journal of Circuits Systems and Signal Processing, 13, 458-463, 2019
dc.identifier.issn19984464
dc.identifier.other2-s2.0-85070714415
dc.identifier.urihttps://dspace.kmitl.ac.th/handle/123456789/9509
dc.sourceInternational Journal of Circuits Systems and Signal Processing
dc.subjectBack-propagation neural network
dc.subjectExternal short circuit
dc.subjectInternal winding fault
dc.subjectPower transformer
dc.subjectWavelet transform
dc.titleApplication of discrete wavelet transform and back-propagation neural network for internal and external fault classification in transformer
dc.typeArticle

Files

Collections

Research