Sensing materials for optical fiber-based selective detection of acetylene

Early fault detection in oil-filled power transformers is an important factor in improving the stability and reliability of the electrical grid. Faults typically result from high-temperature degradation of the mineral oil, either by operation above temperature specification or through localized heating due to electrical discharge and arcing. The standard strategy for diagnosis of fault conditions is to periodically sample the mineral oil and perform dissolved gas analysis (DGA). Varying concentrations of hydrogen (H₂), methane (CH₄), acetylene (C₂H₂), and other hydrocarbons are generated as the oil degrades and can be indicative of fault type. The development of optical fiber-based sensors for dissolved gas detection within transformer oil may provide important new advantages above ex-situ DGA, such as real-time monitoring and spatially resolved (distributed) sensing within the transformer oil. Several different material systems are investigated for detection of acetylene, as well as other hydrocarbons relevant to mineral oil degradation within transformers. Inspired by materials highlighted in the literature for selective acetylene catalysis, several nanostructured nickel / silica oxide-based films are investigated for evanescent field-based sensor materials. In particular, materials are investigated for sensitivity to acetylene, cross-sensitivity to other relevant gas species, and operation at elevated temperature (up to 80°C). Machine learning tools are applied to UV-vis transmission data to enhance gas discrimination and guide sensor design.