A Novel Method of Fault Prediction in Transformer Oil using Infrared Spectroscopy

Abstract

Dissolved gas analysis (DGA) is the most widely used technique for predicting incipient faults in transformer oil. In order to perform DGA, the concentration of dissolved gases such as Methane (CH4), Ethylene (C2H4) and Acetylene (C2H2) are required. The concentration of these gases can be determined by means of Gas Chromatography (GC) which is the most globally used technique. However, this technique requires an experienced operator, has high maintenance and running costs, and time consuming as well. In view of these issues, a new method of fault prediction is proposed in this paper that makes use of the absorption phenomena of transformer oil. Fourier Transform Infrared (FTIR) Spectroscopy was performed on oil samples that have been degraded due to the accumulative and repetitive impact of high voltage impulses. The FTIR spectra gave the peak absorbance values of the dissolved gases viz; CH4, C2H4 and C2H2. Finally, the Duval triangle method was implemented for fault prediction. Also, other properties of oil such as breakdown voltage and dielectric constant have also been evaluated.