Artificial Intelligence Solution For Incipient Faults Diagnosis Of Oil-Filled Power Transformers

Abstract

Power transformers are high cost important
equipment used in the transmission and distribution of
the electric energy. A power transformer in operation is
subjected to different stresses such as electrical stress
and thermal stress due to natural ageing and loading
regime which lead to liberation of gases from the
hydrocarbon mineral oil. Dissolved gas analysis (DGA)
has been widely used as an effective technique to detect
the incipient fault of the transformer. There are different
conventional DGA methods developed for analyzing
these gases such as key Gas, Rogers Ratio,
Doernenburg, International Electrotechnical
Commission (IEC) Ratio, and Duval triangle. Artificial
Intelligence (AI) can be also used to detect power
transformers incipient faults. In this paper, the
applications of two AI approaches have been presented
based on IEC standard which is fuzzy logic approach
(FLA) and artificial neural network approach (ANNA).
Each approach is used to get the correct diagnosis of the
incipient faults and the accuracy is then calculated.