Predictive maintenance of railway transformer oil based on periodic content analysis

The high frequency of operation of commuter trains, due to passenger demand as well as the selection of railway as the mode of daily transportation for commuting on weekdays, increases the usage of on-board power, especially for a train's traction system. As maintenance is rarely performed on transformer oil, it deteriorates and negatively affects transformer performance, increases heat, and may damage the transformer as well. This will result in significantly costly maintenance expenses for train operators. Therefore, this paper proposes a predictive maintenance schedule for transformer oil. The recommendations are based upon an analysis of transformer oil contents and its properties over a 90-month period of operation. A linear correlation between the properties of the oil and the train's period of operation yielded a predictive maintenance schedule, primarily reclamation and filtration, for the oil at the threshold of each property. Major oil changes are to be considered when all properties are approaching their thresholds. As oil deterioration increases over time, a specific maintenance schedule was suggested. This was tested and observed on several transformer units. The content analysis of each oil is also discussed. Based on the results, this predictive maintenance schedule can be used on other trains with the same transformer model or other trains using the same type of insulating oil.