Assessment of the harmfulness of corrosion defects in a pressurized pipeline with a single corrosion defect using FEA and ANN

External corrosion is a significant issue that can lead to ruptures in long-distance liquid transportation facilities such as pipelines. Highlighting the need for regular monitoring of their condition, which can compromise their integrity and safety. Three types of external corrosion defects were studied since it is necessary in some to prioritize interventions when multiple defects appear simultaneously on different sections considering burst pressure of the structure. These pressures were determined using numerical simulations in SolidWorks, based on models validated by previous experimental results. An artificial neural network was developed to quickly predict burst pressures, optimizing the network’s topology through detailed analysis and thus avoiding the need for complex finite element analyses in the future. The results indicate that the defect having parabolic shape is generally the most detrimental to the structure, although the circumferential defect may be more damaging in certain cases. Additionally, the depth of the defect is a crucial parameter compared to geometric characteristics, emphasizing the importance of continuous monitoring to prevent corrosion risks.