Comparative Analysis of New Methods for Defect Type Recognition by Dissolved Gas Analysis

Shutenko, Оleg; Kulik, Alexei

doi:10.1109/khpiweek57572.2022.9916319

Cited by 2 publications

(1 citation statement)

References 16 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Over the more than 50-year history of development and use of the DGA method, many different standards and techniques for interpreting DGA results have been proposed [2][3][4], which use one or another diagnostic criterion to recognise the fault type. These criteria are the value of gas ratios, the percentage gas content as well as the value of gas-to-gas ratios with maximum gas content.…”

Section: Introductionmentioning

confidence: 99%

Method of fault‐type recognition based on the dissolved gas analysis using a set of diagnostic criteria

Shutenko,

Kulyk

2023

IET Generation Trans & Dist

Self Cite

View full text Add to dashboard Cite

This paper proposes a method for recognising the fault type using a set of diagnostic criteria. Existing approaches to the interpretation of dissolved gas analysis (DGA) results usually rely on single diagnostic criteria. The proposed method solves the problems of incomplete use of diagnostic information and discrepancies in the norms and criteria for recognising the same fault inherent in the existing methods. The feature of the method is the analysis of the compliance of the DGA results of the diagnosed equipment with the ones of the equipment with a confirmed diagnosis from the fault database. This approach outperforms the existing methods in the number of faults recognised, allowing a wider range of faults to be detected, including combined faults. In addition, the method provides an opportunity to identify possible causes and localise the fault. This feature is unique and provides a significant advantage when interpreting DGA results. The proposed method demonstrates significantly higher accuracy (100%) in recognising various fault types, including combined faults, in comparison to existing methods (about 50%). The proposed method has a 60% success rate in determining the cause and location of a fault, but this figure depends on the database with faults of various types used.

show abstract

Section: Introductionmentioning

confidence: 99%