A Direct Method for Determining Design and Support Parameters to Meet an Availability Requirement

Abstract

Customers of systems with high availability requirements are in some cases interested in migrating from buying the actual system to buying the availability of the system through availability- contracts. Availability contracts are a subset of outcome-based contracting mechanisms that include: Power-by-the-Hour and Performance-Based Logistics (PBL). Discrete event simulation is usually a preferred approach to model and predict the life-cycle characteristics (cost and availability) of large populations of complex real systems managed over long periods of time with significant uncertainties. However, while using discrete event simulation to predict the availability of a system or a population of systems based on known or predicted system design and support parameters is relatively straightforward; determining the design and support parameters that result in a desired availability is generally performed using search-based methods that can become impractical for designing systems with more than a few variables and when significant uncertainties are present. This paper presents a direct method that is not search based and uses an availability requirement to predict the required logistics, design (including reliability) and operation parameters using discrete event simulation in a time forward direction. This paper also addresses managing availability requirements for systems that include prognostics and health management (PHM) strategies. PHM methods are incorporated into systems to avoid unanticipated failures that can potentially impact system safety and operation, result in additional life-cycle cost, and/or adversely affect the system availability.