Unavailability evaluation in non-Markovian systems with spare components

Abstract

Markovian models for repairable systems assume constant state transition rates. If the operating components are in their useful life phase then the assumption of a constant failure rate is valid. This assumption may not extend, however, to the repair and installation process. the inclusion of non-exponential residence times will make the process non-Markovian. This paper presents a range of studies for a repairable system containing an increasing number of spare components. The effect on system unavailability of various distributions is compared for a wide range of failure, repair and installation rates using Monte-Carlo simulation. The failure rate is assumed to be constant and the distributional variation is applied to the repair and installation processes. In order to facilitate comparison, the mean values associated with the repair and installation times are held constant. The change in unavailability is therefore due entirely to the change in the shape of the distribution.