Data-Driven Diagnostic Analysis of an Oil Leakage Incident in a Utility-Scale Distribution Transformer

Oil leakage in power distribution transformers is a critical reliability concern, often leading to insulation degradation, overheating, and potential failure if left unaddressed. Identifying the underlying causes of oil leakage and its impact on transformer performance is essential for maintaining grid stability and operational safety. This paper presents a data-driven diagnostic analysis of an oil leakage event. This event happened from a four winding distribution transformer connecting a utility scale Battery Energy Storage System (BESS) part of a PV generation plant to the grid. The analysis employed a systematic approach deploying the historic data collected from a transformer edge processing device. The analysis included high-sampling-rate voltage and current data from both the lowvoltage (LV) and high-voltage (HV) sides of the transformer. Additionally, electrical parameters such as the severity factor and excitation current trends were calculated. A root cause analysis was performed to identify anomalies influencing transformer performance. The study examines the correlation between the transformer operating conditions as measured from the monitoring device and the oil leakage occurrence, offering valuable insights into the implications of the oil leakage on the transformer reliability. Furthermore, the diagnostic analysis findings were validated using post-event maintenance reports, reinforcing the accuracy and effectiveness of the proposed methodology. The findings underscore the importance of data-driven diagnostics in boosting predictive maintenance and enhancing the reliability of distribution transformers.