A new method to detect dissolved gases in transformer oil using NIR-IR spectroscopy

Abstract

Dissolved gas analysis (DGA) is widely accepted as an effective technique to detect incipient faults within power transformers. Gases such as hydrogen, methane, acetylene, ethylene and ethane are generated as a result of insulation oil decomposition, while carbon monoxide and carbon dioxide are generated as a result of insulation paper decomposition. Laboratory-based gas chromatography (GC) in accordance to ASTM D3612 standard is globally accepted as a reliable technique to quantify dissolved gases in transformer oil samples. However, this technique incurs running cost and requires an expert to conduct the test. Furthermore, due to the complexity of the equipment, measurement is only performed in a laboratory environment and takes long time to get the results since the extraction of the oil sample from the transformer. To overcome these issues, various online DGA techniques have been developed and currently used by industry. However, these techniques are still considered in development stage with some limitations that have to be addressed to fully accept its results. This paper proposes a new method to estimate the concentration of various dissolved gases in transformer oil using near infrared-to-infrared (NIR-IR) spectroscopy that can be conducted instantly onsite with relatively inexpensive equipment. The proposed measuring technique does not call for expert person to conduct the measurements and has the potential to be implemented online. Results show a good correlation between oil spectral response and the concentration of each dissolved gas in the oil at particular wavelength range. Fuzzy logic model approach is employed to model this correlation and estimate the concentration of each dissolved gas.