Posbist Reliability Theory for Typical Systems with Multicomponents

Abstract

This paper investigates Posbist reliability theory for typical systems with multicomponents. This theory is based on possibility and binary-state assumptions. It works well under situations of insufficient data and epistemic uncertainty. When the lifetime of every component of a typical system is assumed unrelated and a normal, strictly convex fuzzy variable in some possibility space, the Posbist reliability is concretely derived where the exactly value can be obtained at any moment, unlike existing ones only given by a lower and upper bound. We numerically compare the Posbist reliability between typical systems of four types: series, parallel, series-parallel, and parallel-series. Moreover, numerical safety evaluations are given based on possibilistic and probabilistic models.