Error Propagation and Uncertainty Analysis: Application to Fault Tree Analysis

Fault tree analysis (FTA) is a widely used method in the evaluation of safety systems. FTA is based on modeling the behavior of a system by a combination of logic gates (AND, OR) in a logic tree that describes the system. Failure data of low level events is used to compute the frequency of the top event of concern. The failure rate data describes the failure of the system components such as pumps, valves, vessels and instruments. All of the failure rate data are uncertain.

This paper presents a methodology for the evaluation of uncertainty in the top event of concern using propagation of error and uncertainty analysis methods. The fundamental theory is presented and a simple fault tree uncertainty analysis is completed to demonstrate the methodology. The uncertainty analysis methodology allows for the identification of the components or assumptions that dominate the risk.