Reliability analysis for complex systems based on fuzzy generalized stochastic Petri nets considering competing failures

This paper presents a novel methodology for assessing the reliability of complex systems, particularly those with a balanced subsystem, while accounting for competing failure modes and uncertainties. The proposed approach combines generalized stochastic Petri nets (GSPN) with dynamic fault trees (DFT) to establish a comprehensive framework for reliability analysis. GSPN is utilized to model the balanced subsystem due to its capability to effectively represent multi-state components and competing failures modes. Meanwhile, DFT is initially employed to describe other system modules, which are subsequently converted into GSPN for a unified analysis. To address uncertainties in component parameters, triangular fuzzy numbers are used to enhance the robustness of the reliability assessment. Additionally, an extended Monte Carlo simulation algorithm is proposed to solve the fuzzy GSPN model through a two-layer sampling technique. This enables the computation of key reliability metrics, including the fuzzy failure probability and component importance measures. Finally, the applicability and effectiveness of the proposed methodology are demonstrated through a practical engineering case study.