Empirical Failure Pressure Prediction Equations for Pipelines with Longitudinal Interacting Corrosion Defects Based on Artificial Neural Network

Conventional pipeline failure pressure assessment codes do not allow for prediction of interacting defects subjected to combined loadings. Alternatively, numerical approaches may be used; however, they are computationally expensive. In this work, an analytical equation based on finite element analysis the API 5L X52, X65, and X80 corroded pipes with a longitudinal corrosion defect loadings is proposed. An artificial neural network (ANN) trained obtained from (FEA) varied spacings, depths lengths, axial compressive stress were used develop equation. Subsequently, parametric study effects spacing, length, depth, pipe was performed demonstrate correlation between geometries pipes, using The new predicted pressures these grades coefficient determination (R2) value 0.9930 error range ?10.00% 1.22% normalized spacings 0.00 3.00, effective lengths 2.95, 0.80, 0.80.