Planning Accelerated Destructive Degradation Test with Competing Risks

Date
2010-11-01
Authors
Shi, Ying
Meeker, William
Meeker, William
Major Professor
Advisor
Committee Member
Journal Title
Journal ISSN
Volume Title
Publisher
Altmetrics
Authors
Research Projects
Organizational Units
Statistics
Organizational Unit
Journal Issue
Series
Department
Statistics
Abstract

Accelerated destructive degradation tests (ADDTs) are widely used in manufacturing industries to obtain timely product reliability information, especially in applications where few or no failures are expected under use conditions in tests of practical length. An ADDT plan specifies the test conditions of accelerating variables, running time, and the corresponding allocation of test units to each condition. Usually, variables such as temperature, voltage, or pressure can be used as accelerating variables to accelerate degradation of a product. For some applications, however, tests at high-stress test conditions would result in more than one type of failure for test units, called competing risk problems. Careful test planning is important for efficient use of limited resources: test time, test units, and test facilities. This paper describes methods to find unconstrained and constrained optimum test plans for competing risk applications under a given test optimization criterion, such as minimizing the large-sample approximate variance of a failure-time distribution quantile at use conditions. A modified general equivalence theorem (GET) is used to verify the optimality of a given ADDT plan. Generally, an optimum test plan provides insight for constructing a good compromise test plan which tends to be more robust and practical. Monte Carlo simulations are used to provide visualization of the results that might be obtained from given test plans. The methods are illustrated with an application for an adhesive bond.

Comments

This preprint was published as Ying Shi and William Q. Meeker, "Planning Accelerated Destructive Degradation Test with Competing Risks", Statistical Models and Methods for Reliability and Survival Analysis (2013): doi: 10.1002/9781118826805.ch22.

Description
Keywords
Citation
DOI
Source
Collections