Safety Analysis for Cyber-Physical Systems Under Cyber Attacks Using Digital Twin

The safety of cyber-physical systems (CPS) in the presence of cyber attacks are crucial for the well-being of personnel operating it and safety of services rendered. For example, CPS may be deployed in industrial automation and critical infrastructure. The interdependence between its components implies that an attack on one, could trigger a cascading failure in the system, due to faults (deviation from conformity behavior) during the component(s) operation. This might impede operations and affect safety. Our goal is to analyse the effects of cyber attacks on components - to better capture the safety event horizon (boundary up to which events cannot affect the safety) of the physical system. To achieve this, we integrate a cascading fault model into the digital twin (DT), using behavioural patterns of system components, based on a probabilistic Mealy automaton. Our DT is designed to enable safety analysis by identifying deviations that may affect functional and physical safety.