Estimation of some stochastic models used in reliability engineering

Tapio Huovinen

Research output: Book/ReportReport

Abstract

The work aims to study the estimation of some stochastic models used in reliability engineering. In reliability engineering continuous probability distributions have been used as models for the lifetime of technical components. We consider here the following distributions: exponential, 2-mixture exponential, conditional exponential, Weibull, lognormal and gamma. Maximum likelihood method is used to estimate distributions from observed data which may be either complete or censored. We consider models based on homogeneous Poisson processes such as gamma-poisson and lognormal-poisson models for analysis of failure intensity. We study also a betabinomial model for analysis of failure probability. The estimators of the parameters for three models are estimated by the matching moments method and in the case of gammapoisson and beta-binomial models also by maximum likelihood method. A great deal of mathematical or statistical problems that arise in reliability engineering can be solved by utilising point processes. Here we consider the statistical analysis of nonhomogeneous Poisson processes to describe the failing phenomena of a set of components with a Weibull intensity function. We use the method of maximum likelihood to estimate the parameters of the Weibull model. A common cause failure can seriously reduce the reliability of a system. We consider a binomial failure rate (BFR) model as an application of the marked point processes for modelling common cause failure in a system. The parameters of the binomial failure rate model are estimated with the maximum likelihood method.
Original languageEnglish
Place of PublicationEspoo
PublisherVTT Technical Research Centre of Finland
Number of pages148
ISBN (Print)951-38-3332-1
Publication statusPublished - 1989
MoE publication typeD4 Published development or research report or study

Publication series

SeriesValtion teknillinen tutkimuskeskus. Tutkimuksia - Research Reports
Number598
ISSN0358-5077

Keywords

  • reliability
  • estimation
  • modelling
  • failure analysis

Fingerprint Dive into the research topics of 'Estimation of some stochastic models used in reliability engineering'. Together they form a unique fingerprint.

Cite this