Uncertainty Quantification of Failure Probability and a Dynamic Risk Analysis of Decision Making for Maintenance of Ageing Infrastructure
Jeffrey T. Fong, James J. Filliben, Nathanael A. Heckert, Dennis D. Leber, Paul A. Berkman, Robert E. Chapman
Risk, as the product of failure probability and failure consequence, has been estimated and applied by engineers and managers to help make critical decisions on (a) maintenance of ageing plants, and (b) planning of new infrastructure. For ageing plants, failure probabilities are more difficult to estimate than consequences, primarily because of a shortage of time-varying data on the condition of the complex systems of hardware and software at varying scales after years of service. A different argument holds for yet-to-be-built infrastructure, since it is also hard to estimate the time-varying nature of future loadings and resource availability. A dynamic, or, time-dependent risk analysis using a time-varying failure probability and a consequence with uncertainty estimation is an appropriate way to manage ageing infrastructure and plan new ones. In this paper, we first introduce the notion of a time-varying failure probability via a numerical example of a multi-scale fatigue model of a steel pipe, and then the concept of a dynamic risk for decision making via an application of the analysis to the inspection strategy for a cooling piping system of a 40-year-old nuclear power plant. Significance and limitations of the multi-scale fatigue life model and the risk analysis methodology are presented and discussed.
Fourth International Conference on Reliability, Safety, and Hazard, Jan. 10-13, 2019. IIT-Madras,