Weighted Estimation of Component Reliability in Series Systems

作者：Luz Gamiz M

来源：IEEE Transactions on Reliability, 2012, 61(3): 779-786.

DOI：10.1109/TR.2012.2194192

• Heavy censoring
• Kaplan-Meier
• mean squared error
• statistically independent competing risks
• survivor function

The characteristics of the failure times of components in a series system are estimated from observed system failure times and causes of failure. Usually this situation is also approached from a competing risks model viewpoint where, in case of statistical independence, the lifetime of each component may be estimated by the Kaplan-Meier method (KM). For simplicity, we consider only two components in the system. If we are interested in estimating the reliability or survivor function of each component, we may construct for each case an estimator by conveniently interpreting the data set. In this paper, we suggest a modified version of the KM method for estimating the survivor function of a component that uses the information about the lifetime of the other component (which is not being analysed at the moment) to get a better accuracy with heavily censored sampling information, which in fact is the case of the reliability context addressed here.

• 出版日期2012-9