Date of Award

3-1996

Document Type

Thesis

Degree Name

Master of Science

Department

Department of Operational Sciences

First Advisor

Lee J. Lehmkuhl, PhD

Abstract

This study used the Poisson regression technique, bootstrap estimates, and the Kaplan-Meier estimates for survivor curves to determine the impact of Programmed Depot Maintenance (PDM) on weapon system availability. More specifically, these techniques estimated the effect in the number failures per year due to PDM, but the bootstrap technique also estimated the effect in the amount of downtime experienced by a weapon system due to PDM. Although the Poisson regression model did not pass the rigors of statistical testing, the Poisson regression suggested differences in the number of failures per year between PDM and no-PDM weapon systems. The results of the bootstrap estimates for the number of failures per year and amount of downtime per year showed that the no-PDM weapon systems experienced approximately 1.3 failures per year while remaining unserviceable approximately 120 days per year, and PDM weapon systems experienced approximately .3 failures per year while remaining unserviceable approximately 30 days per year. PDM reduced the number of failures per year and drastically reduced weapon system downtime per year. PDM increased weapon system availability from approximately 67 percent to approximately 92 percent.

AFIT Designator

AFIT-GOR-ENS-96M-03

DTIC Accession Number

ADA324256

Download

Share

COinS