Computer aided recognition of defects in GIS

Author

Ziomek, W. ; Schlemper, H.-D. ; Feser, K.

Author_Institution

Inst. of Electr. Power Eng., Poznan Univ. of Technol., Poland

Volume

1

fYear

1996

fDate

16-19 Jun 1996

Firstpage

91

Abstract

A computerized pattern recognition system based on phase resolved partial discharge (PRPD) measurements has been developed. The paper presents results of defect classification obtained for various GIS installations. The system is trained to distinguish between seven basic defect types, such as voids in spacers, moving metallic particles, protrusions on electrodes and floating electrodes. The defect classification is based on 60 parameters extracted from PRPD patterns and utilizes different recognition algorithms. Tests with a database of more than 600 individual measurements gave satisfactory results. Tests with multiple defects and disturbed measurements were done. The database was collected during laboratory experiments and on-site tests of GIS installations

Keywords

electrodes; gas insulated substations; pattern recognition; power engineering computing; voids (solid); GIS; computerized pattern recognition system; defect classification; disturbed measurements; electrode protrusions; floating electrodes; moving metallic particles; multiple defects; on-site tests; phase resolved partial discharge; recognition algorithms; spacers; voids; Classification algorithms; Databases; Geographic Information Systems; Partial discharges; Pattern analysis; Pattern recognition; Power engineering computing; Pulse measurements; Testing; Voltage;

fLanguage

English

Publisher

ieee

Conference_Titel

Electrical Insulation, 1996., Conference Record of the 1996 IEEE International Symposium on

Conference_Location

Montreal, Que.

ISSN

1089-084X

Print_ISBN

0-7803-3531-7

Type

conf

DOI

10.1109/ELINSL.1996.549290

Filename

549290