| Unlike many industries that have their equipment in one location with consistent usage patterns, armies move their vehicles between different geographic locations with varying environmental, and usage conditions. This creates interesting conditions for study, as those geographic changes can be studied to detect their effect on system reliability.;Unfortunately, this is not being fully exploited, due in part to the poor capture and storage of information, a problem faced by many operators of maintenance databases.;This thesis develops a method to characterize failure data contained in a maintenance database using a standardized naming system, and applies a proportional hazards model for each geographic location using covariates to represent the conditions.;In addition to understanding how a system has performed, the proportional hazards model will allow geographic location factors to be used in predicting system reliability and spares parts requirements in a new location. |
