Dynamic Bayesian Network–Based Approach for Probabilistic Fracture Failure Assessment of Supercritical CO2 Transportation Pipelines

Fracture failure in supercritical CO2 transportation pipelines can arise from contaminants, third-party damage, or welding defects, primarily due to the unique phase properties of the conveying medium. This study presents a probabilistic fracture failure assessment of supercritical CO2 pipelines using a Dynamic Bayesian Network (DBN) approach. The DBN model captures the causal evolution of pipeline failure from a microscopic damage mechanism perspective, incorporating both ductile and brittle fracture behaviors. It explicitly accounts for the temporal dependencies among contributing factors of pipeline fracture failure. Based on the model, the dynamic failure probability of pipeline fracture is quantified. Additionally, a diagnostic analysis is performed to identify the most probable factors leading to CO2 pipeline fracture failure. The methodology is demonstrated through a case study, which can support risk management strategies for supercritical CO2 pipeline fracture failure.