Effect of MTTF and MTTR statistical uncertainty on mission dependability for power electronics equipped systems

In order to predict and assess the RAM (reliability, availability, maintainability) or dependability characteristics of high reliability systems, suitable models, implemented and solved by means of appropriate software tools, have usually to be utilised. The results deriving from such procedure are obviously function of the dependability parameters (mean time to failure MTTF and mean time to repair MTTR) of each basic component, which are usually known or can be assessed by on-the-field reliability demonstration tests with an associate statistical confidence limit. The aim of this paper is to evaluate the impact of such statistical uncertainty on the dependability measures carried out at the system level. A modelling procedure, previously developed by the authors and based on the Monte Carlo method, is adopted for simulating the performances of the system as a function of the dependability characteristics of its components. The results of a real case study, represented by a multivoltage vehicle for railway applications, are shown and discussed.