Efficient reliability analysis of an axial compressor in consideration of epistemic uncertainty

The reliability of complex systems, e.g. infrastructure networks or turbines, and further its proper prediction, is of paramount importance to our modern society. The concept of survival signature enables to entirely separate the structure of a network and its components probabilistic properties. As a consequence, once survival signature of a network has been computed, a reliability analysis can be performed by evaluating only the probabilistic part of the network. It is precisely this advantage that makes the analysis particularly efficient and clearly distinguishes the concept from traditional approaches. In reality, epistemic uncertainties, as vague or varying expert knowledge on a component's future behavior, impede a proper reliability analysis. Considering and quantifying these in a plausible manner is therefore of major interest to decision-makers. Fuzzy probabilities offer a naturally fitting approach for this purpose. Despite the advantageous efficiency of the survival signature, analyzing complex systems under considertaion of epistemic uncertainties implies high computational effort. In order to counteract these efforts, two extended Monte Carlo simulation methods, originally utilized for imprecise structural reliability problems, are adapted into the context of system reliability. By integrating fuzzy probabilities into the probability structure of the survival signature and employing advanced Monte Carlo methods, this paper provides an efficient approach to quantify the reliability of a complex system taking into account epistemic uncertainties. Beyond the merging of the theoretical aspects, the approach is applied to a functional model of an axial compressor, in order to demonstrate its usability.