ANN- and FEA-Based Assessment Equation for a Corroded Pipeline with a Single Corrosion Defect

Most of the standards available for assessment failure pressure corroded pipelines are limited in their ability to assess complex loadings, and estimations conservative. To overcome this research gap, study employed an artificial neural network (ANN) model trained with data obtained using finite element method (FEM) develop equation predict a pipeline single corrosion defect. A analysis (FEA) medium-toughness (API 5L X65) subjected combined loads internal longitudinal compressive stress was carried out. The results from FEA various geometric parameters were used as training dataset ANN. After ANN trained, its performance evaluated, weights biases development equation. prediction newly developed has good correlation value, R2 0.9998, percentage errors ranging −1.16% 1.78%, when compared results. When estimates based on Det Norske Veritas (DNV-RP-F101) guidelines, standard more conservative than study.