Fault diagnosis methods of transformer based on dissolved gas detection in oil by gas chromatography

Large transformer is the core equipment in power system, the diagnosis and prevention of equipment fault is very important to the safe operation of power system. At high temperature and high voltage, dissolved gas produced by the decomposition of transformer oil is an important indicator of transformer operation status. However, due to the low content of dissolved gas and the complexity of the measurement processes, it is easy to produce errors, which brings great challenges to the accurate detection of dissolved gas. In addition, how to establish the correct relationship between the content of dissolved gas components with the types and degrees of transformer fault also needs to be studied. Therefore, this paper first clarified the measurement processes of dissolved gas in transformer oil, then analysed the possible error sources of each link, then introduced common error assessment methods and proposed feasible methods to reduce dissolved gas test errors, and finally introduced the application of artificial intelligence to fault diagnosis of transformers based on dissolved gas content. This paper will provide some feasible theoretical support for reducing the measurement error of dissolved gas in transformer oil and accurately diagnosing transformer faults.