{"title":"Partial discharges in high-voltage transformers","authors":"F. A. Sakarya, M. Hayes","doi":"10.1109/SSAP.1992.246887","DOIUrl":null,"url":null,"abstract":"A novel approach to on-line localization of partial discharges in a power transformer is presented. A 2-D technique based on subspace rotation is given. By far-field source and narrowband signal assumptions, the localization problem is reduced to an arrival angle estimation problem. The first step uses two eigenvalue decompositions to clean the data and compute the time delays. The second step converts time delays into source locations using the geometrical relationships between them. The new localization technique computes 2-D arrival angles directly from the normalized eigenvectors of a newly derived signal space matrix. The method eliminates the constraint on subarray spacing and the need for a search procedure inherent in most 2-D eigenanalysis techniques, and does not require pairing of angles.<<ETX>>","PeriodicalId":309407,"journal":{"name":"[1992] IEEE Sixth SP Workshop on Statistical Signal and Array Processing","volume":"11 4 1","pages":"0"},"PeriodicalIF":0.0000,"publicationDate":"1992-10-07","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"[1992] IEEE Sixth SP Workshop on Statistical Signal and Array Processing","FirstCategoryId":"1085","ListUrlMain":"https://doi.org/10.1109/SSAP.1992.246887","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"","JCRName":"","Score":null,"Total":0}
引用次数: 0
Abstract
A novel approach to on-line localization of partial discharges in a power transformer is presented. A 2-D technique based on subspace rotation is given. By far-field source and narrowband signal assumptions, the localization problem is reduced to an arrival angle estimation problem. The first step uses two eigenvalue decompositions to clean the data and compute the time delays. The second step converts time delays into source locations using the geometrical relationships between them. The new localization technique computes 2-D arrival angles directly from the normalized eigenvectors of a newly derived signal space matrix. The method eliminates the constraint on subarray spacing and the need for a search procedure inherent in most 2-D eigenanalysis techniques, and does not require pairing of angles.<>