A censored sequential posterior odd test method in testability demonstration test planning. (August 2020)

• Record Type: Journal Article
• Authors: Zhang, Yong; Wang, Chao; Lin, Xin; Liu, Guanjun; Yang, Peng; Qiu, Jing
• Is Part Of: Proceedings of the Institution of Mechanical Engineers. Volume 234:Number 4(2020:Aug.)
• Page Start: 579
• Page End: 587
• Publication Date: 2020-08
• Subjects: Sequential posterior odd test -- censored sequential posterior odd test -- testability demonstration -- fault detection and isolation; Reliability (Engineering) -- Mathematical models -- Periodiclals; Risk assessment -- Mathematical models -- Periodicals; Engineering design -- Mathematical models -- Periodicals; 620.00452
• DOI: 10.1177/1748006X20918791
• Languages: English
• ISSNs: 1748-006X

Abstract

Testability demonstration tests can effectively verify product capabilities of fault detection and isolation; however, they suffer from insufficient samples, long cycles and high costs due to destructiveness of fault injection tests, which leads to an increasing demand for small sample tests. The sequential posterior odd test can effectively reduce sample sizes, but test results can be random and the sample size may be large. In this article, a censored sequential posterior odd test method is proposed, which can control the incremental risk caused by forced censoring within a contracted range by risk splitting. The development process of the testability demonstration test based on the censored sequential posterior odd test is designed. The number of censored tests and the calculation method of the censored threshold are presented. The case application shows that with the same prior distribution and constraint parameters, the average sample size of the proposed method is smaller than that of the sequential posterior odd test and of the classical method considering risks for both producers and consumers. The presented method can further reduce the risk of misjudgment and the number of test samples, contributing to the reduction of the test cycle and costs.