Title: Determination of sample size to support diagnostic inspection of components
Authors: Eishiro Higo; Mahesh D. Pandey
Addresses: Department of Civil and Environmental Engineering, University of Waterloo, Waterloo, ON, Canada ' Department of Civil and Environmental Engineering, University of Waterloo, Waterloo, ON, Canada
Abstract: A complex engineering system like a nuclear power reactor consists of a large variety and number of engineering components. As a part of a component aging management program, the diagnostic inspections of various component populations are performed to detect the onset of any unanticipated degradation. A prudent selection of the inspection sample size is necessary to optimise inspection cost. Sample size selection is typically based on the traditional statistical hypothesis test, which tends to result in a fairly large sample size. This paper presents an alternate approach to the sample size determination (SSD) problem based on the concept of the value of information (VoI). The paper provides a comparative analysis of the VoI and hypothesis-testing approaches through illustrative examples. The VoI approach is shown to provide a more meaningful way to minimise the cost of inspection as a function of component-replacement cost and losses arising from a failure. The characteristics and advantages of the VoI approach are analysed.
Keywords: inspection; sample size determination; SSD; Bayesian analysis; value of information; VoI; hypothesis testing; decision analysis.
International Journal of Quality Engineering and Technology, 2018 Vol.7 No.2, pp.128 - 151
Received: 07 Sep 2017
Accepted: 24 Jun 2018
Published online: 15 Jan 2019 *