Optimization of Inspection, Testing and Maintenance (ITM) of Transformers

Inspection/Testing/Maintenance (ITM) (both fixed-interval and condition monitoring-based) of large power transformers helps detect existing or impending failures of components and improve availability of the equipment. Traditionally, ITM operations are not optimized based on desired system reliability/availability and costs of risk reduction. This work has developed a framework that combines knowledge gained from historical data (e.g., related to outages, failure modes, causes, and consequences) with Physics-of-Failure (PoF)-based models (built with consideration of credible failure modes, causes, reliability and ITM parameters such as fixed-interval preventive maintenance frequency, and ITM costs) to predict system failure probabilities as a function of age and ITM costs. Optimal ITM requirements for a sample Generator Step up Unit (GSU) were evaluated through the trade-off between ‘forced outage probabilities’ (of duration $\gt 1$ day) and ‘ITM costs’ for a limited scope focusing on winding failures. It is found that oil analysis is the one testing methodology most effective in reducing the forced outage probabilities for winding failures. On the other hand, increasing the frequency of offline electrical tests has a much lesser impact while significantly increasing the overall cost of the ITM strategy.