Abstract: A detection technology for dissolved fault characteristic gases in insulating oil based on photoacoustic spectroscopy is proposed to meet the real-time monitoring requirement of transformer operation status. Considering the characteristics of low total amount and difficult separation of fault characteristic gases dissolved in transformer insulation oil, a high-performance miniaturized non-resonant photoacoustic cell is designed. By constructing a two-way excitation structure in the photoacoustic cell with a coaxial high-reflection mirror, the amplitude of the photoacoustic signal is effectively enhanced. By connecting a vacuum degassing device with the high-performance non-resonant photoacoustic cell through a gas circuit, an automated detachment and analysis system for dissolved fault characteristic gases in transformer oil is formed. By optimizing the selection of absorption spectral lines, the crossinterference of water vapour on the characteristic fault gas acetylene is suppressed and the detection accuracy is improved. In addition, by combining wavelength modulation technology with second harmonic detection technology, the background noise is reduced and the detection sensitivity of acetylene is further improved. The experimental results show that this proposed scheme can achieve a detection limit of 0.05 μL/L for acetylene in transformer insulation oil.

Key words: spectroscopy, photoacoustic spectroscopy, vacuum degassing, transformer, dissolved acetylene in oil