Characterizing System Failure Curves with Vitality

Abstract

Identifying the causes of failure is fundamental to improving the survivability of Army systems whether they are remote sensors, complex vehicles, or networks. The shape of a repair frequency curve of a complex system is determined by variability in the initial quality of its components, variability in the system's wear and maintenance, and random failures independent of the condition of the components. The contributions of these sources of failure are modeled using vitality theory, which describes population survival curves in terms of the stochastic decline of a hidden Markov process (vitality) to an absorbing boundary representing death. By reinterpreting the vitality model, the population becomes the complex system, the individuals become the system's components, vitality becomes the remaining amount of component wear prior to failure, and absorption into the boundary represents component failure and replacement. Fitting the model to the repair frequency curve of a 180 ton rear dump truck, the shape of the repair curve is quantitatively partitioned into three factors: the operating environment, quality control in manufacture, and variability in maintenance. This information may be useful in identifying causes of system failure.

Open PDF

Document Details

Document Type
Technical Report
Publication Date
Dec 01, 2008
Accession Number
ADA505887

Entities

People

  • J. J. Anderson

Organizations

  • University of Washington

Tags

Communities of Interest

  • Materials and Manufacturing Processes

DTIC Thesaurus Topics

  • Absorption
  • Abstracts
  • Algorithms
  • Boundaries
  • Complex Systems
  • Differential Equations
  • Dump Trucks
  • Environment
  • Equations
  • Gaussian Distributions
  • Maintenance
  • Markov Processes
  • Probability
  • Quality Control
  • Standards
  • Survival
  • Vehicles

Fields of Study

  • Engineering

Readers

  • Economics
  • Logistics and Supply Chain Management.
  • Statistical inference.