Design of an Experiment to Examine Repair Process Errors of Military Vehicle Mechanics

摘要

This thesis develops a method and concept for analysis of errors made by US Army mechanics. A process model is developed to describe the hierarchy of actions accomplished by soldier mechanics to complete a diagnosis and repair effort on a disabled vehicle. From this process model an error classification scheme is developed. An error list is derived from the error classification scheme and used in combination with a list of factors that contribute to soldier mechanic's errors to determine shortcomings in the US Army system that selects, trains, employs, and provisions soldier mechanics. An experiment is developed which allows non-intervening observers to collect information regarding the incidence of error types with their associated contributing factors. This information is used in a statistical analysis. The analytical method used is canonical correlation. Canonical correlation analysis produces a rank ordering and relative scaling of the factors that contribute to soldier mechanic's errors. This analytical result may then be used by top-level US Army decision-makers when deciding the allocation of research and development funds to reduce the frequency of errors made by soldier mechanics, thus improving the overall effectiveness of the US Army maintenance effort.