Data Quality and Reliability Analysis of U.S. Marine Corps Ground Vehicle Maintenance Records

Abstract

We evaluate data quality issues present in Marine Corps maintenance records and develop statistical models to identify the most influential predictor variables to estimate the expected number of failures that cause a vehicle to be non-operational. When a vehicle becomes non-operational, we refer to it as a deadlining event. We analyze data collected from 3,154 Medium Tactical Vehicle Replacement (MTVR) vehicles between January 1, 2011 and December 31, 2013. Data quality issues are present in vehicle serial numbers, maintenance defect codes, regional code, and odometer readings. Due to the high level of inaccuracy in odometer meter readings, vehicle mileage cannot be used as a metric for usage. We build Poisson generalized linear regression models to estimate the expected number of vehicle deadlining events. Without the presence of a true measurement of vehicle usage, the insight gained from fitting regression models to the maintenance data is limited. The number of unscheduled maintenance events acts as a surrogate usage measure within the model. In our model, more than one scheduled maintenance event per year shows evidence of reducing the number of deadlining events. We recommend the improvement of odometer meter reading accuracy in order to provide an effective usage measurement for future studies.

Open PDF

Document Details

Document Type
Technical Report
Publication Date
Jun 01, 2015
Accession Number
ADA632454

Entities

People

  • Adam T. Foley

Organizations

  • Naval Postgraduate School

Tags

DTIC Thesaurus Topics

  • Accuracy
  • Data Sets
  • Failure Mode And Effect Analysis
  • Ground Vehicles
  • Iraqi-War
  • Life Cycle Management
  • Logistics
  • Maintenance
  • Maintenance Management
  • Maintenance Personnel
  • Marine Corps
  • Measurement
  • Reliability
  • Tactical Vehicles
  • United States
  • United States Naval Academy
  • Vehicles

Fields of Study

  • Engineering

Readers

  • Logistics and Supply Chain Management.
  • Personnel Management and Statistics in the Military and Department of Defense
  • Regression Analysis.