Availability modelling and analysis of a two-component parallel system under stochastic dependence

Abstract

In real-world settings, machines are not available all the time. They can undergo different collapses and malfunctions. This may increase costs and sometimes gravely threaten safety. To face this challenge, it is important to assess the availability based on the different dependencies between their components. The purpose of this paper is to compute exactly the availability of a two-component parallel system considering stochastic dependence. We propose an efficient and user-friendly model, based on Cox proportional hazards model using the generalised Weibull distribution. A calculation framework is presented to compute more realistic system availability even for real systems provided with a history of failures. A numerical example is given to assess the stochastic dependence effect on the availability of the system and to illustrate the model. A managerial insight is provided to allow the practitioners to better estimate this latter in order to develop adequate maintenance strategies.

Keywords Availability modelling, Cox proportional-hazards model, Multi-components systems, Stochastic dependence