Data-driven Methods to Predict the Burst Strength of Corroded Line Pipelines Subjected to Internal Pressure

A corrosion defect is recognized as one of the most severe phenomena for high-pressure pipelines, especially those served a long time. Finite-element method and empirical formulas are thereby used strength prediction such pipes with corrosion. However, it time-consuming finite-element there limited application range by using formulas. In order to improve strength, this paper investigates burst pressure line pipelines single subjected internal based on data-driven methods. Three supervised ML (machine learning) algorithms, including ANN (artificial neural network), SVM (support vector machine) LR (linear regression), deployed train models experimental data. Data analysis first conducted determine proper pipe features training. Hyperparameter tuning control learning process then performed fit best corroded pipelines. Among all proposed models, model three layers has highest training accuracy, but also presents largest variance. The provides both high accuracy validation accuracy. performance in terms generalization ability. These can be surrogate transfer new coming data future research, facilitating sustainable intelligent decision-making