Abstract Ensuring the reliability and integrity of oil and gas pipelines relies heavily on the early detection of internal corrosion threats. This study introduces a machine learning (ML) based predictive model designed to analyze real-time operational data and anticipate the development of internal corrosion. The approach employs a feed-forward multilayer neural network model to predict corrosion rates specific to high-temperature pipeline environments. Results demonstrate that the model can accurately predict internal corrosion threats with an accuracy of 96.7% thereby validating its efficacy and demonstrating its potential for practical implementation in real-world scenarios. Integrating this approach with real-time operational monitoring offers oil and gas operators a practical and data-driven tool to enhance safety, reduce costs, and improve the overall reliability of pipeline operations.
Tharayil et al. (Mon,) studied this question.