Probabilistic Evaluation of Running Ductile Fracture Length in Dense-Phase CO2 Pipelines

Abstract A running ductile fracture (RDF) in dense/supercritical-phase CO2 pipelines is a severe failure mode that can release large volumes of CO2 and poses risk to the public. Uncertainties exist in the final RDF length due to varying mechanical and geometric properties along the pipeline. This study uses a probabilistic approach to quantify the final crack length of RDFs in dense/supercritical phase CO2 pipelines by accounting for the inter- and intra-joint variability of key pipeline properties including Charpy-V Notch toughness, wall thickness, and yield strength, as well as these variables’ spatial correlations. A Monte Carlo simulation is developed to evaluate the cumulative distribution function (CDF) of the final RDF length in CO2 pipelines using probability distributions of the previously mentioned variables and a limit state function based on the Battelle two-curve method. 12 cases of CO2 pipelines with X65 grade steel and an outside diameter of 609.6 mm are investigated, including various operating pressures and temperatures, and three fluid compositions consisting of pure CO2 and CO2-rich mixtures with impurities. Two spatial correlation scenarios are considered to investigate the effect of spatial correlation on RDF length. CDFs are developed for each case and compared. The results show that spatial correlation can significantly increase RDF length and should be accounted for in probabilistic analyses. This study is useful for developing risk-based design codes to ensure adequate fracture control in CO2 pipelines.