Modelling complexity in system safety: Generalizing the D2T2 methodology

Fault Tree (FT) and Event Tree (ET) analysis remain the most widely adopted methods for system safety assessment, valued for their efficiency and transparency. However, their static structure and assumption of independent basic events limit their ability to represent dynamic behaviours and dependencies, particularly in modern systems where control, automation, and shared resources play a central role. Existing extensions and alternatives improve expressiveness but often at the expense of generality, interpretability, or computational feasibility. The Dynamic and Dependent Tree Theory ( D 2 T 2 ) addresses these shortcomings by preserving the global FT/ET structure while enabling dependencies to be represented at multiple levels of fidelity. In this work, we propose an extension that generalizes the framework to directly capture subsystem and train level dependencies, thereby avoiding decomposition to basic event interactions that would otherwise exacerbate model complexity and state space growth. This capability places virtually no restriction on dependency type or location, and retains FT/ET traceability essential for review and certification purposes. By offering a continuum of modelling choices between classical FT/ET and fully dynamic formalisms, D 2 T 2 provides a distinctive contribution to the state of the art in dependency analysis for safety critical systems.