Hybrid incorporation of physical & social failure mechanisms into probabilistic risk assessment

This paper emphasizes the importance of creating a hybrid Probabilistic Risk Assessment (PRA), capable of covering the interactions of risk contributing factors (e.g. hardware, software, and human aspects of systems, as well as their organizational and regulatory environments). Traditional PRAs use statistical data and expert opinion for estimating the probability of "basic events" in order to calculate system risk. More recently, the focus has been shifting toward modeling the underlying failure mechanisms of the elements of risk scenarios and consequently, PRA requires roots in diverse fields such as reliability engineering, cognitive psychology, organizational behavior, and computer science. This paper provides two building blocks for a hybrid PRA: The first block describes (a) a multi-level framework, called Socio-Technical Risk Analysis (SoTeRiA), as a theoretical foundation for the integration of technical system risk models with both the social features (e.g. safety culture) and the structural aspects (e.g. safety practice) of the organization operating the system and (b) appropriate techniques to operationalize and quantify the SoTeRiA framework. It is argued that using a single modeling technique is insufficient due to the multidisciplinary nature of complex system risk assessments, and a hybrid inclusion of deterministic and probabilistic methods is proposed. The second block relates to the incorporation of physical failure mechanisms into PRAs and its application for modeling Common Cause Failures (CCFs). Last year's disaster at the Japanese Fukushima Daiichi Power Plant is an example that highlights the criticality of modeling CCFs. The purpose is to build a theoretical foundation for the incorporation of Probabilistic Physics-Of-Failure (PPOF) models into PRA in a way that the interactions of physical failure mechanisms (e.g. fatigue and wear) and, ultimately, the dependencies among multiple component failures can be depicted.