Reliability Analysis Based on Incompletely Specified Data

Abstract

In this paper new algorithm for the reliability analysis of system is considered. The reliability analysis consists of 2 essential steps - creation of mathematical model and quantitative analysis. The system mathematical model is constructed depending on the specifics of analysis and properties of the investigated system. In this paper the mathematical model in form of the structure function of non-coherent Multi-State System (MSS) is considered. The non-coherent system is specific group of systems in context of reliability analysis for which for system component degradation does not always lead to a degradation or failure of the system. MSS is mathematical model that allows analyzing of some (not only two) states/performance levels of system reliability. The structure function is type of mathematical model, that express dependency between system behavior and behavior of its components. Structure function can be represented in different forms, for example, as minimal cuts/paths, fault tree, reliability bloc diagrams. One of them used in this paper is Multi-Valued Decision Diagram (MDD). MDD is typical used for representation of data of large dimension. Created mathematical model is then used for quantitative analysis. This analysis includes calculus of different system characteristics such as availability/unavailability and other indices and measures. New methods for the creating of structure function in form of MDD and calculation of some indices for quantitative analysis is proposed in the paper. Important advantage of this method is possibility to use for MDD construction based on incompletely specified data and analysis of non-coherent MSS. Usage of the method is demonstrated on biker crashes survival evaluation.