The development of failure injection heuristics to enhance minimum cutsets detection of complex systems

Abstract Safety analysis became an exhausting process after introducing complex critical systems in the past few decades. Classical safety analysis methods can be insufficient to efficiently study a system’s behavior. That is because classical methods are performed independently from the system design process and depend on the engineers’ skills. Model-Based Safety Analysis (MBSA) is an approach in which design and safety engineers share a common abstracted model. The common model is created by selecting relevant portions of the design model, with the addition of failure behaviors and safety-related information. One of the main bottlenecks of MBSA is the exponential expansion of system states along with the computation power required to search for the Minimum Cutsets. This manuscript proposes an object-oriented tool that applies safety analysis to design models. The tool encloses components’ failure behavior models in terms of object-oriented classes. The developed tool maps Simulink design files into a set of interconnected object-oriented components. A developed failure injection heuristics is applied in the abstracted design model to produce Minimum Cutsets automatically. The heuristics are based upon a mix of depth-first search and breadth-first search along with component sensitivity in the system. The heuristics are developed to reduce the amount of computation required to capture all Minimum Cutsets.