Composite multilinearity, epistemic uncertainty and risk achievement worth

被引:6
|
作者
Borgonovo, E. [1 ]
Smith, C. L. [2 ]
机构
[1] Bocconi Univ, ELEUSI Res Ctr, Milan, Italy
[2] Idaho Natl Lab, Idaho Falls, ID 83415 USA
关键词
Reliability analysis; Risk analysis; Importance measures; Uncertainty analysis; PROBABILISTIC SAFETY ASSESSMENT; COMMON-CAUSE FAILURES; ENGINEERING SYSTEMS; RELIABILITY; FOUNDATIONS; METHODOLOGY; ASSESSMENTS; COMPONENTS; MODELS; ISSUES;
D O I
10.1016/j.ejor.2012.05.008
中图分类号
C93 [管理学];
学科分类号
12 ; 1201 ; 1202 ; 120202 ;
摘要
Risk achievement worth is one of the most widely utilized importance measures. RAW is defined as the ratio of the risk metric value attained when a component has failed over the base case value of the risk metric. Traditionally, both the numerator and denominator are point estimates. Relevant literature has shown that inclusion of epistemic uncertainty (i) induces notable variability in the point estimate ranking and (ii) causes the expected value of the risk metric to differ from its nominal value. We investigate the conditions under which the equality of the nominal and expected values of a reliability risk metric holds. We then study how the presence of epistemic uncertainty affects RAW and the associated ranking. We propose an extension of RAW (called ERAW) which allows one to obtain a ranking robust to epistemic uncertainty. We discuss the properties of ERAW and the conditions under which it coincides with RAW. We apply our findings to a probabilistic risk assessment model developed for the safety analysis of NASA lunar space missions. (c) 2012 Elsevier B.V. All rights reserved.
引用
下载
收藏
页码:301 / 311
页数:11
相关论文
共 50 条