Modelling complexity in system safety: generalizing the D2T2 methodology

2025-10-20

AI Summary

Traditional fault tree analysis (FTA) and event tree analysis (ETA) struggle to capture dynamic behaviors and multi-level dependencies in complex systems, often leading to overly conservative or insufficiently conservative safety assessments. To address this, we propose a generalized modeling paradigm grounded in Dynamic and Dependency Tree Theory (D²T²), integrating compositional failure analysis with a scalable dependency modeling mechanism. This is the first approach to uniformly formalize arbitrary types and locations of dependency relationships. While preserving the interpretability and computational efficiency of classical FTA/ETA, D²T² significantly enhances the representation of dynamic couplings among components and subsystems. Experimental evaluation demonstrates that the framework achieves a favorable trade-off between modeling accuracy and engineering feasibility, effectively reducing misjudgments in safety risk assessment—and thereby mitigating associated economic costs—thus improving both the reliability and practical applicability of safety analysis for complex systems.

Technology Category

• Reasoning under Uncertainty: Uncertainty Representations
• Constraint Satisfaction and Optimization: Satisfiability Modulo Theories
• Knowledge Representation and Reasoning: Computational Complexity of Reasoning

Application Category

• Graph Algorithms and Modeling for the Web: Efficient manipulation of static and dynamic Web-related graphs
• Security and Privacy: Large-scale security measurements
• Semantics and Knowledge: Data modeling to support human-machine intelligence, including LLMs agents, intelligent system behavior, explanations, and user-friendly interactions

Abstract

Although Fault Tree and Event Tree analysis are still today the standard approach to system safety analysis for many engineering sectors, these techniques lack the capabilities of fully capturing the realistic, dynamic behaviour of complex systems, which results in a dense network of dependencies at any level, i.e. between components, trains of components or subsystems. While these limitations are well recognised across both industry and academia, the shortage of alternative tools able to tackle such challenges while retaining the computational feasibility of the analysis keeps fuelling the long-lived success of Fault Tree and Event Tree modelling. Analysts and regulators often rely on the use of conservative assumptions to mitigate the effect of oversimplifications associated with the use of such techniques. However, this results in the analysis output to be characterised by an unknown level of conservatism, with potential consequences on market competitiveness (i.e., over-conservatism) or safety (i.e., under-conservatism). This study proposes a generalization of the Dynamic and Dependent Tree Theory, which offers theoretical tools for the systematic integration of dependency modelling within the traditional Fault and Event Tree analysis framework. This is achieved by marrying the traditional combinatorial nature of failure analysis, formalised by the Fault and Event Tree language, with more flexible modelling solutions, which provide the flexibility required to capture complex system features. The main advantage of the proposed approach in comparison to existent solutions is the ability to take into account, under the same modelling framework, any type of dependency regardless of its nature and location, while retaining the familiarity and effectiveness of traditional safety modelling.

Problem

• Fault and Event Tree analysis cannot capture complex system dynamics
• Alternative tools lack computational feasibility for dependency modeling
• Unknown conservatism levels affect safety and market competitiveness

Innovation

• Generalizes Dynamic and Dependent Tree Theory methodology
• Integrates dependency modeling into Fault and Event Trees
• Captures any dependency type under unified framework