Title :
Prediction intervals for Weibull observations, based on early-failure data
Author_Institution :
Dept. of Math. & Stat., Massachusetts Univ., Amherst, MA, USA
fDate :
12/1/1996 12:00:00 AM
Abstract :
Let n similar components with Weibull lives be put on life-test simultaneously, and let the first r failures be observed. A prediction interval for failure-time Xk, 0<r<k⩽n, is constructed. An explicit formula is obtained for computing quantiles of a relevant pivotal statistic. Prediction intervals based on various estimators of parameters and methods are compared. The results can be used to predict the failure-time of a k-out-of-n:F system
Keywords :
Weibull distribution; consecutive system reliability; failure analysis; life testing; Weibull observations; early-failure data; failure-time prediction interval; k-out-of-n:F system; life-test; parameter estimation; pivotal statistic quantiles; Estimation theory; Exponential distribution; Life testing; Maximum likelihood estimation; Parameter estimation; Shape; Statistical distributions; Weibull distribution;
Journal_Title :
Reliability, IEEE Transactions on