Model-in-the-Loop Tests Bench as an Efficient Means for Digital Transition: From Descriptive to Simulated Model-Based Systems Engineering and Model-Based Safety Assessment

System architecture models are mainly focused on the static definition of complex systems, particularly in the safety domain, where they enable the design of functional architectures that are mapped to physical architectures, as viewed by the system architect. This functional proposition and its allocation to the physical architecture are typically analyzed separately by safety engineers, who develop their own functional and physical architecture, often corresponding to component groupings in a Physical Breakdown Structure (PBS). Based on this analysis, the safety engineer prepares a System Failure Modes and Effects Analysis (FMEA) to manually determine the failure modes and their effects. Simultaneously, through dysfunction analysis, the safety engineer identifies feared events, usually derived from statistical failure studies of components allocated to the dysfunctional chain, and from in-flight reports, which capture observed failures when the product is in service. The safety engineer then continues with the elaboration of Fault Tree Analysis (FTA), identifying all leaf events (failures of components in the functional chain) and defining logical gate combinations (based on failure propagation paths) to trace back to the top-level feared event. Stochastic simulations estimate the probability of the feared event by simulating many combinations of leaf events. However, this approach often relies heavily on component failure conditions, with less emphasis on the actual fault path. These dysfunctional analyzes are based on the expertise of the safety engineer in mapping the propagation paths of failure, relying on static data and failure reports. This paper proposes a methodology for digitally coupling systems architectures using Model Based System Engineering (MBSE) and safety analysis through Model-Based Safety Assessment (MBSA). This methodology is then validated using a Model-in-the-Loop (MiL) test bench to investigate fault propagation paths, applied to a seat control system featuring a Passenger Control Unit (PCU), an Electronic Control Unit (ECU), and actuators integrated into the seat structure.