Value of Information and Hypothesis Testing Approaches for Sample Size Determination in Engineering Component Inspection: A Comparison

Abstract

A sample size determination method is developed for a two-action problem that represents a component maintenance scenario requiring current state estimation. For safety and generation efficiency, each component of a nuclear power plant must be regularly inspected. In terms of safety, the larger the sample size inspected, the less the uncertainty about current and future states of the components; however, such inspections are expensive. Thus, sample size determination becomes an important problem. A key idea for solving this problem is the Value of Information (VoI) and its derivation: the Expected Net Gain of Sampling (ENGS). The ENGS is a function of sample size and represents by how much a decision maker benefits from the observed data. By maximizing the ENGS, the optimal sample size is determined in terms of cost-benefit analysis.