Failure Pressure Prediction in Oil Transfer Pipe Lines with Incomplete Penetration and Undercut Welding Defects

Abstract

The main target of the current study is the assessment of API 5L X52 pipe line containing lack of penetration and undercut welding defects and to study the influence of the defect length and depth on the failure pressure of the pipeline. The failure pressure is calculated by using Finite element software ANSYS ver19. The results from finite element analysis are used to train an artificial neural network for easily predicting the failure pressure with a wide range of defect depths and lengths. Also the results obtained were also used in the regression analysis to find equations linking the maximum pressure to the defects dimensions. The results revealed a Good agreement between finite element modeling and the experimental tests, theoretical values and available design codes with maximum average error of (13.4%). The failure pressure varied significantly as a function of the defect length and depth.