Abstract
In-service pipelines are often subjected to longitudinal forces and bending moments resulting from, for example, ground movement or formation of free spans in addition to internal pressures. In practice, there are some site-specific cases where corrosion anomalies interact with the external loads. A refined assessment model is required to understand the load carrying capacity of pipe.
In this study, a burst capacity model for corroded pipelines under combined internal pressure and axial compression is developed based on extensive parametric three-dimensional (3D) elasto-plastic finite element analyses (FEA) and artificial neural network (ANN) technique. The parametric FEA employs the ultimate tensile strength (UTS)-based burst criterion and idealizes corrosion defects as semi-ellipsoidal shaped flaws. The FEA model is validated by full-scale burst tests of pipe specimens containing semi-ellipsoidal shaped flaws reported in the literature. Extensive parametric FEA are carried out to evaluate the burst capacity of corroded pipelines under combined internal pressure and axial compression by varying the pipe geometric and material properties, defect depth, length and width, and magnitude of axial compressive stress. Based on the FEA results, an ANN model is developed utilizing the open-source platform PYTHON to predict the burst capacity of corroded pipelines under combined internal pressure and axial compression. The well-trained ANN model is further validated by full-scale burst tests of corroded pipelines under combined internal pressure and axial compression carried out by Det Norske Veritas (DNV).
Predicting the oxidant demand in full-scale drinking water treatment using an artificial neural network: Uncertainty and sensitivity analysis
