Accession Number : ADA136567

Title :   Effects of Assuming Independent Component Failure Times, if They Are Actually Dependent, in a Series System.

Descriptive Note : Annual rept. 1 Sep 82-30 Sep 83,


Personal Author(s) : Moeschberger,M L ; Klein,J P

PDF Url : ADA136567

Report Date : 26 Oct 1983

Pagination or Media Count : 91

Abstract : The overall objective of this proposal is to investigate the robustness to departures from independence of methods currently in use in reliability studies when competing failure modes or competing causes of failure associated with a single mode are present in a series system. The first specific aim is to examine the error one makes in modeling a series system by a model which assumes statistically independent component lifetimes when in fact the component lifetimes follow some multivariate distribution. The second specific aim is to assess the effects of the independence assumption on the error in estimating component parameters from life tests on series systems. In both cases, estimates of such errors will be determined via mathematical analysis. A graphical display of the errors for representative distributions will be made available to researchers who wish to assess the possible erroneous assumption of independent competing risks. A third aim is to tighten the bounds on estimates of component reliability when the risks belong to a general dependence class of distributions (for example, positive quadrant dependence, positive regression dependence, etc.). Major decisions involving reliability studies, based on competing risk methodology, have been made in the past and will continue to be made in the future. This study will provide the user of such techniques with a clearer understanding of the robustness of the analyses to departures from independent risks, an assumption commonly made by the methods currently in use. (Author)

Descriptors :   *Statistical analysis, *Systems analysis, *Parts, *Reliability, Failure, Mathematical analysis, Life expectancy(Service Life), Distribution, Multivariate analysis, Life tests, Parametric analysis, Nonparametric statistics, Risk, Bivariate analysis, Exponential functions

Subject Categories : Administration and Management
      Statistics and Probability

Distribution Statement : APPROVED FOR PUBLIC RELEASE