Test-To-Failure Reliability Validation Test Design

View at publisher
Authors: Jiliang Zhang
Cited by: 0
References: 4

Abstract

SUMMARY & CONCLUSIONSReliability validation test is the test for validating that the product reliability target is met with the required statistical confidence. Success-Run method is commonly used for the test design, in which no failure is allowed in the test for concluding the reliability target is met with the given statistical confidence. However, the time-to-failure distribution cannot be estimated if no failure is observed in the test. Alternatively, test-to-failure method allows some failures occur in the test and still can validate the product reliability. It is a much better approach to overcome or mitigate some limitations in Success-Run test design method.In this paper, the test design method of test-to-failure for the Weibull time-to-failure distribution with a known Weibull shape parameter is studied. After summarizing an existing method with Binomial distribution [1] - [2] in Section 1, two new methods are presented in the next two sections. The first new method uses the median rank of unreliability as the estimate for each ordered time-to-failure of the failed sample in the test, and then the specific time-to-failure is derived with this median rank with the reliability target for the service life. In the second new method, it uses the derived expectation of the total time-of-testing in z-space with the transformation of z = tβwhere t is the test time and β is the shape parameter in Weibull time-to-failure distribution. In both methods, the governing equation for the test design is derived, and the test sample size and the test duration can then be determined. The new methods along with the existing methods are compared for the test design result accuracy with the numerical calculations, and it shows that the two new methods and the existing method with Binomial distribution are acceptable for the test design. On the other hand, the new methods have advantages of the close-form solutions in determining the sample size and the test duration and the relaxed condition for the application of the method.

Fields of Study & Subjects

Reliability and Maintenance Optimization Software Reliability and Analysis Research Statistical Distribution Estimation and Applications

Publication details

Journal / source
Annual Reliability and Maintainability Symposium (RAMS)
Pages
1–6
Publication date
26 January 2026
Language
EN