A dynamic failure propagation methodology supporting the risk assessment of multidisciplinary systems

Modern critical infrastructure systems have grown to be increasingly complex. Among the many reliability and system safety (RSS) characteristics of the system, failure propagation is critical to understand. Understanding failure propagations can significantly reduce the system's risk since corrective design actions can be taken early on. Beyond traditional RSS methods, some are centered on failure propagation including fault tree analysis (FTA), the BowTie method, fishbone diagrams, etc. The BowTie analysis is a method for assessing the prevention and recovery attributes of a complex safety-critical system. The proposed methodology in this paper addresses the prevention aspect of the BowTie analysis. Specifically, we proposed a method based on physics-based multidisciplinary model to accurately simulate the failure propagation of the system. The failure propagation paths are developed naturally by the simulation model and are therefore more complete. The novelty of such an approach is that practitioners do not need to predict the paths. The methodology is demonstrated using a case study of a three tank system with one critical function. The case study results show that the proposed method can successfully identify failure propagation from "causes" to "hazards" and its multidisciplinary nature helps capturing paths that cross system disciplines (such as propagation through the environment).