Data Requisites for Transformer Statistical Lifetime Modelling Part II: Combination of Random and Aging-Related Failures

Statistical lifetime modeling is of importance for replacement management of aged power transformers. Survival data are recognized as important as failure data in improving the accuracy level of the lifetime models since transformer failures are rare events and most of the units are still in operating condition. This paper argues that differentiating random failures and aging-related failures is also important. Different data requisites for modeling random failures and aging-related failures are analyzed and compared through Monte Carlo simulations. The transformer life-cycle failure model can be built by combining the random and aging-related failure models. A case study is presented to show that through postmortem analysis, the two failure modes can be distinguished and, hence, it helps to improve the accuracy of the combined model.