Computing small-fleet aircraft availabilities including redundancy and spares

Abstract

Logistics support is a key element of aircraft transportation systems. This paper is concerned with the impact of aircraft spares provisioning decisions on the availability of aircraft. Spares provisioning in this context is complicated by the fact that spares may be shared across aircraft and that aircraft may have redundant systems. In addition, decisions concerning aircraft spares support require a rapid response for safety reasons. Analytical models have proven to provide a quicker response time than corresponding simulation models. There is an existing analytical model that includes the effect of redundancy and spares, but the underlying assumption is that a large number of aircraft are being modeled. In many applications, predictions of the number of times an aircraft can fly each day and the number of aircraft that are ready at any time are applied to a small fleet of aircraft. This paper demonstrates the improvement in computational accuracy that is achieved by reflecting the impact of small numbers of aircraft on availability projections. The approach used is to extend existing finite queuing spares models to including redundancy. Further, the method is used to optimize spares provision with respect to a user specified availability goal. Although the case study for this work is a military combat aircraft application from the Gulf War, the method is applicable to any small system of vehicles or machines where components may be redundant, demand and repairs may be approximated as following an exponential distribution, and limited access to spare parts is the rule. Scope and purpose In many situations in exploration, mining, rescue, and defense, it is necessary to dispatch a small fleet of machines to a remote area to perform an important function. Since the location is difficult to access, resupply is often difficult or impossible. In anticipation of this situation, the machines often include redundant parts to allow for some component failures in the field that do not eliminate a machine from further use. Decisions have to be made, in advance, as to which backup components (spares) should be included with the machines at the remote site to be used when failures exceed the built-in redundancy level. This turns out to be an interesting modeling problem. Since the number of machines is relatively small, exact counting of machines and their components (both redundant and spares) must be done in order to make a decision. This leads to a combinatorial problem that cannot be solved for practical-sized problems with a reasonable computational effort by, say, discrete simulation. In this paper, we develop a methodology, based on finite queuing theory, which addresses this need. It is applied and validated on an availability situation similar to the deployment of aircraft to the Middle East during the Gulf War.