ABSTRACT
In the literature related to reliability and maintainability analysis on reparable equipments, a great number of studies have been published about modelling and analyzing their failure times. However, there are no so much studies which analyze the behavior of these equipments when failures happen at demand. In this case, models are completely different since the failure distribution is discrete. In this work, the model is performed from the real behavior of equipments without any a priori probability distribution. From this model, discrete probability function and cumulative distribution function, needed for the subsequent parameters estimation, are obtained. By combination of these functions and imperfect maintenance models, likelihood function for reliability analysis is constructed to jointly estimate its parameters, that is, the failure at demand probability and the maintenance effectiveness, and their variability. Then, an application case performs the estimation procedure applied to a database of a safety equipment of a Nuclear Power Plant, where the obtained estimations together with their variability can be used to plan optimal test and maintenance intervals.
