Time-variant reliability analysis of corroded gas pipelines considering dynamic defect evolution and interaction

This study presents a comprehensive methodology for estimating the time-variant reliability of corroded gas pipelines, considering the dynamic progression and interactions of external corrosion defects. The growth of corrosion depth is modeled using a gamma process, while linear growth models are used to describe the evolution of defect length and width. Graph theory is employed to capture both direct and indirect interactions between defects, providing an efficient and robust approach to track and identify interaction paths among all defects. For reliability analysis, subset simulation (SS) is adopted, which offers significant computational advantages, particularly when assessing low failure probabilities. Several case studies with varying defect arrangements are conducted to demonstrate the effectiveness, efficiency, and robustness of the proposed method. The results show that neglecting potential interactions may lead to underestimating failure probabilities by up to three orders of magnitude compared to the accurate results in certain scenarios. The extent of this underestimation is influenced by the number of interacting defects as well as their geometric characteristics. The case studies also reveal interesting transitions between different failure modes, ranging from single-defect-induced failures to those driven by defect clusters. These findings provide deeper insights into the underlying failure mechanisms of pipelines with multiple corrosion defects, which is critical for enhancing the integrity management of gas pipelines.