Identifying the a priori time-to-failure distribution of unique highly vital elements using the expert method

Abstract

Abstract. Aim. Systematic failures, unlike random hardware failures, cannot be described using the mathematics of the probability theory and the dependability theory. However, such failures are the biggest problem due to their unpredictability. In the case of describing systematic failures of unique highly vital systems, a solution is presented by an approach that involves taking into account the quantitative criteria of functional performance of a facility in time that are defined, for example, by prescribing a set of parameters for each function that characterize its ability to perform, as well as acceptable limits for such parameters’ variation. The paper aims to develop an approach to the use of expert evaluation for the purpose of identifying the type and parameters of the distribution of the time to failure of unique highly vital elements. The author examined an approach to determining the a priori distribution of the time to failure of unique highly vital elements by pairwise comparison that would be useful for improving the accuracy of their dependability indicators.Methods. Hierarchy analysis, fuzzy logic and permutation theory were used. Fuzzy variables were introduced, the degrees of belonging to which are interpreted as subjective probabilities of the time to failure and its characteristics within different time intervals. Methods were proposed for accounting for the accuracy of expert evaluation and for solving the cluster sampling problem.