This paper proposes a structured framework for dependency modeling based on Generalized Stochastic Petri Nets (GSPN). A reusable library of dynamic gates, developed with Predicate/Assertion–enhanced GSPN (PA-GSPN), is introduced to capture complex dependencies in an interpretable way while supporting repair modeling.
GSPN is extended with predicates and assertions to construct dynamic gates. Validation is performed on two industrial case studies: a Hypothetical Sprinkler System (HSS) including Functional Dependency (FDEP), SPARE and K-out-of-N (KooN) gates, covering both component-level repairs and shared repair crew dependencies; and an Aircraft Fuel Distribution System (AFDS) applying Priority AND (PAND) and Priority OR (POR) gates for priority and sequence logic. All models were simulated in GRIF 2022 using Monte Carlo analysis.
The library of dynamic gates provides models that are interpretable, modular, and easier to update for engineers and researchers. For the HSS under non-repairable conditions, the proposed framework yields a system unreliability of 0.0263 at 1,000 h, compared to 0.0265 in the reference literature, corresponding to an absolute difference of approximately 2.71 × 10−4. Incorporating repair strategies reduces system unreliability from 0.0263 to 0.0061, an improvement of approximately 76.8%, without major structural modification of the model. For the AFDS, the framework produces an unreliability of 0.8803 at 1,000 h, in close agreement with the reference Petri Net model value of 0.8751. Simulation times are 3.31 s for the HSS and 8.69 s for the AFDS. Both case studies confirm that the PA-GSPN approach reduces reliance on gate-specific subnet construction while preserving numerical accuracy and enabling straightforward model updates when new dependencies, repair policies, or environmental conditions are introduced.
To the authors' knowledge, this paper presents the first PA-GSPN framework that systematically formalizes dynamic gates as reusable predicate–assertion modules within a GSPN-based reliability model. It offers a practical and systematic approach that enhances model clarity, interpretability, and reusability. Validation through two industrial case studies demonstrates its applicability and effectiveness, providing engineers and researchers with a reliable tool for modeling complex dependencies and repair strategies in safety-critical systems.
