Publication:
Partial Discharge Pattern Classification Using Frequency-Domain Statistical Descriptors

Simple item page
dc.citedby2
dc.contributor.authorMd Thayoob Y.H.en_US
dc.contributor.authorGhani A.B.Abd.en_US
dc.contributor.authorGhosh P.S.en_US
dc.contributor.authorid6505876050en_US
dc.contributor.authorid24469638000en_US
dc.contributor.authorid55427760300en_US
dc.date.accessioned2023-12-28T08:57:59Z
dc.date.available2023-12-28T08:57:59Z
dc.date.issued2003
dc.description.abstractInternal partial discharges taking place inside an insulation has been identified as one of the major contributors of the high voltage equipment failures. Recently, with the rapid development of computer based signal processing, it is possible to identify the source of partial discharge (PD) signals originating from complex insulation structures under varied operating conditions. The research on PD pattern classification is mostly performed in the time-domain but there is also an increasing trend to undertake frequency-domain analysis. In this research work, an experiment has been carried out on an 11KV, single-core 240mm2 XLPE cable. The operating conditions are varied using two parameters which are the soil condition in terms of soil thermal resistivity and cable loading in terms of core temperature. The time-domain electrical PD signals obtained from a PD measurement system are transformed into frequency spectrum by short-time fast Fourier transform (STFFT) using Matlab toolbox. In order to carry out the PD pattern classification, nine frequency-domain statistical descriptors are identified for the present work. A database for the range of values of the nine descriptors have been developed using PD signals obtained from one of the soil thermal resistivity at normal full load working temperature that is identified as standard. The values of the descriptors obtained from signals corresponding to other operating conditions stated earlier are then compared against the standard database and the number of successes is measured in terms of the commonly used recognition rate. The analysis of the results has clearly indicated that the proposed methodology has the ability to classify PD patterns originating from the cable under varied operating conditions.en_US
dc.description.natureFinalen_US
dc.identifier.epage175
dc.identifier.scopus2-s2.0-1542747992
dc.identifier.spage171
dc.identifier.urihttps://www.scopus.com/inward/record.uri?eid=2-s2.0-1542747992&partnerID=40&md5=59d347f64f495538ef5b7fb39ec76204
dc.identifier.urihttps://irepository.uniten.edu.my/handle/123456789/29861
dc.pagecount4
dc.sourceScopus
dc.sourcetitleElectrical Insulation Conference and Electrical Manufacturing and Coil Winding Conference and Exhibition
dc.subjectClassification (of information)
dc.subjectDigital signal processing
dc.subjectElectric cables
dc.subjectElectric loads
dc.subjectElectric transformers
dc.subjectFast Fourier transforms
dc.subjectFrequency domain analysis
dc.subjectPattern recognition
dc.subjectSoil testing
dc.subjectSpectrum analysis
dc.subjectStatistical tests
dc.subjectThermal conductivity
dc.subjectTime domain analysis
dc.subjectVoltage distribution measurement
dc.subjectPartial discharges (PD)
dc.subjectShort time fast fourier transforms (STFFT)
dc.subjectElectric breakdown
dc.titlePartial Discharge Pattern Classification Using Frequency-Domain Statistical Descriptorsen_US
dc.typeConference paperen_US
dspace.entity.typePublication
Files
Collections
SCOPUS