Determining the Reliability of Power Transformers Considering a Five States Markov Model

This paper presents a methodology to calculate power transformers reliability indicators using a five state Markov model, which contemplates the probabilities of failure, maintenance, overload, and shutdown. The objective of including more states that lead to the abnormal operation of the equipment than in other works aims to obtain a better estimate of the risk of failure, as well as direct more asserive maintenance actions. With these analyses, we will illustrate the impacts on the availability of transformer variations in failure rates. The methodology is verified through a case study using real data from 10 substations and 25 power transformers belonging to a Brazilian electricity company. The presented results demonstrate the impacts on the avaliability of power transformer variations in failure rates, showing that not considering all the possible causes of abnormal transformer operation may underestimate the Severity index.