Reliability-growth analysis of locomotive electrical equipment

This paper presents a stochastic model developed to describe the reliability-growth of advanced technological systems during a planned on-service period, at the end of which the supplier has to demonstrate the conformity of system reliability to the contractual target. Purchase contracts of such systems generally regard a fleet of units (e.g., a fleet of new traction units for railway or subway service, airplanes, buses,...) so that the development programs are arried out on several system’s copies, often put on service in batches during the development program. Thus, heterogeneity over units’ reliability often arises, as a consequence of different scheduling of the improving activities. Hence,a stochastic reliability-growth model able to reveal heterogeneity was developed. This model was applied to the electrical equipment of a fleet of locomotives built by Ansaldo Breda for the Italian railway transport company. A graphical goodness-of-fit method shows the adequacy of the model to describe the observed reliability-growth process of a single locomotive. Again, a testing procedure was developed in order to reveal possible heterogeneity over the locomotives of the whole fleet and, hence, on the basis of the revealed heterogeneity, a mixed Poisson reliability-growth (MPRG) model was proposed to describe the reliability-growth process of a randomly selected locomotive of the fleet. Then, the probability of demonstrating the observance of the contractual obligations at the end of the development period was predicted on the basis of the failure data collected during the whole development program. Moreover, a procedure was proposed to test, before the development program is terminated, whether the planned remaining time is sufficiently long to assure a high probability of demonstrating the attainment of the reliability target at the end of the program.