Probabilistic Risk Assessment

Abstract

Engineering systems have become complex, interconnected, nonlinear, geographically widespread, and expensive to build and operate. Consequences of failures of such systems are often harmful to humans, environment, and the economy. Despite many preventive features, failures of engineering systems remain unavoidable. As the users of these systems become more prosperous, the demands for their safety equally rise, leading to stricter requirements and needs for increased performance and minimum risk. Risk due to a given failure event of the system is formally defined as the expected loss measured by the product of the likelihood of occurrence of that event and its consequences. This article describes the formal processes for estimating the leading risks of complex engineering systems by developing probabilistic and logic‐based risk models. Particular focus is put on identifying low‐frequency failure events involving highly adverse consequences. Risk models used in engineering systems are often probabilistic in nature because one needs to estimate very low frequency events. This article starts with definitions and notions of risk and performance, and presents the elements of a formal probabilistic risk assessment as applied to engineering systems. The article reviews development of the necessary logic models, solutions of which describe risk‐significant scenarios of failure events. Such logic models describe the relationships and performance of the hardware subsystems, structures, components, human actions, and software of a complex engineering system. Finally, methods for solving and probabilistically quantifying these logic models to arrive at the magnitude and leading contributors to risks from complex engineering systems are also presented.