Chronological Monte Carlo Simulation for Evaluating Spare Transformer Requirements in Distribution Substations

Abstract

This paper deals with the sizing of stocks of spare transformers and mobile units, necessary to maintain the reliability of a set of substations of electric energy distribution. Two types of failure are considered: (i) catastrophic, in which the transformer suffers unrepairable damage and must be replaced by a new one and (ii) random, when the transformer can be repaired in the substation itself, not requiring its replacement by a spare. Two probabilistic methods are considered based on: Markov process and chronological Monte Carlo simulation (MCS), both implemented in Matlab environment. Initially, the methods are applied to the same set of transformer class and a comparison with results from the literature ensures the validation of the tested models. Because it is more flexible and capable of incorporating operational conditions not supported by the Markov model, the chronological MCS method is chosen to perform a series of numerical applications with a real system considering factors such as: load transfer, increase in the number of transformers in operation, inclusion of reinforcements to the stock of spare transformers and mobile units. Several sensitivity analyses are carried out to identify, from the reliability indices and operating costs, the factors that most influence the system performance.