Scholarlink

This study designs a simplified photoacoustic spectroscopy system for detecting methane, ethane, and ethylene gas concentrations in transformer oil. Unlike traditional devices, it uses only an infrared heat source and a single optical filter, reducing complexity and cost. The filter wavelength is chosen based on the absorption lines of the target gases, while others show weak or no absorption. In the experiment, the light source frequency ranged from 20 to 90 Hz (excluding 50 Hz) with 10 Hz steps, and four optical power levels were tested: 7.55 mW, 9.44 mW, 11.64 mW, and 13.81 mW. The optimal performance was achieved at 20 Hz and 13.81 mW, with a sensitivity of 0.68 uV/ppm and a minimum detection limit of 2.35 ppm. The signal-to-noise ratio (SNR) decreased with higher frequencies, with the highest SNR observed at 7.55 mW below 60 Hz and lower SNR at higher power levels above 60 Hz, indicating the interaction between frequency and power. The results show that the system offers high accuracy, reliability, and stability, providing an efficient, cost-effective solution for real-time gas monitoring in transformer oil, with strong potential for industrial predictive maintenance applications.

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Research on Single-Filter Photoacoustic Spectroscopy for Detecting Total Concentration of Dissolved Methane, Ethane, and Ethylene in Transformer Oil

  • 会议论文

  • 发表日期：

    2026年

查看全文
  • Xiaofang Yuan,
  • Zien Liu,
  • Beibei Tang,
  • Jun Cao,
  • Tao Lin,
  • Fengxiang Ma,
  • Yue Zhao

摘要

This study designs a simplified photoacoustic spectroscopy system for detecting methane, ethane, and ethylene gas concentrations in transformer oil. Unlike traditional devices, it uses only an infrared heat source and a single optical filter, reducing complexity and cost. The filter wavelength is chosen based on the absorption lines of the target gases, while others show weak or no absorption. In the experiment, the light source frequency ranged from 20 to 90 Hz (excluding 50 Hz) with 10 Hz steps, and four optical power levels were tested: 7.55 mW, 9.44 mW, 11.64 mW, and 13.81 mW. The optimal performance was achieved at 20 Hz and 13.81 mW, with a sensitivity of 0.68 uV/ppm and a minimum detection limit of 2.35 ppm. The signal-to-noise ratio (SNR) decreased with higher frequencies, with the highest SNR observed at 7.55 mW below 60 Hz and lower SNR at higher power levels above 60 Hz, indicating the interaction between frequency and power. The results show that the system offers high accuracy, reliability, and stability, providing an efficient, cost-effective solution for real-time gas monitoring in transformer oil, with strong potential for industrial predictive maintenance applications.