INVESTIGATION OF EFFECT DUE TO CORRELATION BETWEEN COMPONENTS ON SYSTEM RELIABILITY

Abstract

System reliability estimates are generally made using a model which assumes independence between components, and results are often claimed to be conservative. For a two component serial system bivariate distributions are developed for three cases: (1) bivariate exponential, (2) bivariate geometric, and (3) a composite exponential, geometric bivariate. These distributions are then utilized to investigate the reliability of a two component serial system when an estimate of the correlation coefficient is available. The estimate of the reliability thus obtained is then compared with the corresponding estimate obtained by use of the model which assumes that the system reliability is the product of the component reliabilities. The difference between these two estimates is tabulated for values of the correlation coefficient between -.25 and +.25, for each of the three bivariate distributions. Conditions under which the effect is maximum are explored and a method of approximating the reliability difference is suggested.

Open PDF

Document Details

Document Type
Technical Report
Publication Date
Jan 01, 1963
Accession Number
AD0479709

Entities

People

  • David R. Ayres
  • Ira N. Schwarz

Organizations

  • Naval Postgraduate School

Tags

Communities of Interest

  • C4I
  • Energy and Power Technologies

DTIC Thesaurus Topics

  • Bivariate Density Functions
  • Confidence Limits
  • Data Science
  • Discrete Distribution
  • Distribution Functions
  • Equations
  • Estimators
  • Information Science
  • New York
  • Probability
  • Probability Density Functions
  • Probability Distributions
  • Random Variables
  • Statistical Algorithms
  • Statistical Analysis
  • Statistical Estimation
  • United States

Fields of Study

  • Engineering
  • Mathematics

Readers

  • Regression Analysis.