Power transformer lifetime modeling

Abstract

As large proportions of power transformers are approaching or have exceeded their design life, concerns have aroused at their impact on the reliability of power networks, and forward replacement planning/budgeting is therefore required. In this paper, an appropriate transformer lifetime model is recognized as the key for the accurate replacement volume prediction. Since transformer failures are rare events in most of the electric utilities, industry-wide reliability data reflecting global perspective on transformer lifetime are recognized as good sources for formulating baseline models. A Bayesian Updating procedure is then proposed to incorporate the prior knowledge on the distribution of transformer lifetime (the baseline model) with available field failure data. Sequentially updating the model whenever new failure occurs allows the existing lifetime model to be improved in a progressive manner.