City University of Hong Kong

CityU Institutional Repository >
3_CityU Electronic Theses and Dissertations >
ETD - Dept. of Management Sciences  >
MS - Doctor of Philosophy  >

Please use this identifier to cite or link to this item:

Title: Designing accelerated life test plans under progressive interval censoring with random removals
Other Titles: Dai you sui ji yi zou de zhu bu shan shi mo xing xia de jia su shou ming shi yan she ji
Authors: Ding, Chang (丁昌)
Department: Department of Management Sciences
Degree: Doctor of Philosophy
Issue Date: 2010
Publisher: City University of Hong Kong
Subjects: Accelerated life testing.
Industrial engineering.
Quality assurance.
Notes: CityU Call Number: TA169.3 .D56 2010
ix, 173 leaves 30 cm.
Thesis (Ph.D.)--City University of Hong Kong, 2010.
Includes bibliographical references (leaves 153-165)
Type: thesis
Abstract: Accelerated life tests (ALTs) are widely used to assess the quality of high reliability products. However, in industrial applications, life tests are often being censored for various practical considerations. The Type I and Type II censoring schemes, which terminate a test at a specified time or after a specified number of failures are observed, are frequently applied to save experiment time and/or cost. In some studies, it is not uncommon for some non-failed units to be randomly removed from the test because of their dangerous status and/or the reduction of experiment funding and facility. Furthermore, interval inspection schemes are frequently employed by experimenters due to their convenience in implementation to avoid the need to continuously monitor the test. In this study, the life tests under consideration integrate the features of ALT, interval inspection and progressive censoring with random removals. Type I and Type II censoring schemes are considered. In particular, our study focused on the derivation of optimal ALT plans and the design of ALT sampling plans. For these two problems, the lifetimes of units are assumed to follow a Weibull distribution while the number of removals at each inspection is assumed to follow a binomial distribution. The optimal ALT plans are determined, which minimize the asymptotic variance of an estimated quantile at use condition. A numerical study was conducted to investigate the properties of derived ALT plans under different parameter values. Based on the patterns found from the numerical results, some useful suggestions are provided for experimenters in designing ALT plans under similar conditions. Several numerical examples are provided for illustrative purposes. For the design of ALT sampling plans under similar settings, two levels of constant stresses higher than the use condition are employed. The optimal low stress level and the allocation proportion, are obtained by minimizing the generalized asymptotic variance of the maximum likelihood estimation of model parameters. The required sample size and acceptability constant which satisfy given levels of producer's and consumer' risks are found. For validation purposes, the true acceptance probabilities of the derived ALT sampling plan are evaluated using a Monte Carlo simulation. Some numerical examples are also provided.
Online Catalog Link:
Appears in Collections:MS - Doctor of Philosophy

Files in This Item:

File Description SizeFormat
abstract.html132 BHTMLView/Open
fulltext.html132 BHTMLView/Open

Items in CityU IR are protected by copyright, with all rights reserved, unless otherwise indicated.


Valid XHTML 1.0!
DSpace Software © 2013 CityU Library - Send feedback to Library Systems
Privacy Policy · Copyright · Disclaimer