A Diagnostic Study of Dissolved Gases in Transformers based on Fuzzy Logic Approach

Stray gassing of transformer oil is a convoluted physical phenomenon, in which several parameters act simultaneously thus making the understanding of Dissolved Gas Analysis (DGA) more daunting. Historically, prudent maintenance strategies are better conceived by studying the condition of units in service. To alleviate transformer management and decision matrix, transformer condition assessment is critical using unfailing, noninvasive diagnostics and monitoring mechanisms. The diagnosis of stray gassing of oil can be performed by distinguishing and embedding transformer severities using Fuzzy logic (FL) mapping. Despite the fact that the IEC FL approach is handy in the appraisal of dissolved gases, on occasion the DGA results are impossible to match by current codes and consequently providing inconclusive results. This work reveals a supplemental gas ratio C
2
H
6
/CH
4
based upon trial and error on a fleet of transformers. These supplemental ratio provide more detailed information about the thermal decomposition of transformer oil from various infancy faults and associated degradation gases transforming from Methane (CH
4
) to Ethane (C
2
H
6
) to Ethylene (C
2
H
4
). Additionally, in the IEC FL approach, there is no quantitative measure for the possibility of distinct faults. Further, this work provides incipient fault diagnosis mechanism and some guidelines to allow effective planning and maintenance practice.