On Selection of Importance Measures in Risk and Reliability Analysis

摘要

In risk and reliability analysis a number of importance measures are used, including traditional measures such as the Birnbaum's measure, the improvement potential and the risk achievement worth, as well as uncertainty importance measures reflecting how uncertainties on the component level influence the uncertainties on the system level. Two examples of the latter type of measures are the correlation coefficient and the change in the variance of the reliability of the system when ignoring the uncertainties in the component reliability. In practice it is a challenge to select the appropriate importance measure for different applications of risk and reliability analyses. The purpose of the present paper is to provide a guideline for this selection. A structure for the guidance is established based on three areas of application: design, operation and maintenance (testing). For each of these areas some specific types of importance measures are recommended, covering both traditional measures and uncertainty importance measures. An example is presented to show the applicability of the guideline.