Application of Polycrystalline SnO2 Sensor Chromatographic System to Detect Dissolved Gases in Transformer Oil

Dissolved gas analysis (DGA) has an essential role to transformer status. A gas sensor is fabricated for DGA by using the developed one-dimensional polycrystalline SnO2 fiber material. On basis of this, a portable gas chromatographic device (PGCD) is developed for DGA. A wavelet-genetic algorithm (GA) threshold denoising method is proposed for noise reduction and applied to chromatogram to detect the weak peaks of the PGCD for latent transformer faults. Then, an improved filter matching method based on gray incidence degrees for peaks detection is presented to overcome the limitations of the derivative method. Results indicate that the proposed signal processing method demonstrates competitive performance, thereby providing a favorable foundation for quantitative analysis. Tests were conducted to measure the repeatability and accuracy of the developed polycrystalline SnO2 sensor and PGCD. Comparison with the commonly used flame ionization detector is performed, and the effectiveness of the developed sensor and the corresponding PGCD are verified.