A review on corrosion modelling for submarine pipeline

Abstract

Undersea pipelines are susceptible to corrosion, leading to resource loss and significant harm to the natural ecosystem. Hence, it is necessary to construct a corrosion model for detection and maintenance. This research primarily examines the existing literature on data-driven models utilising Machine Learning (ML) methods, particularly Artificial Neural Networks (NN’s) and also considers the models based on other theories to provide references for corrosion models. An initial stage involves analysing the main cause of corrosion and identifying the key factors contributing to this structural failure. Then, the review highlights the benefits of ML by listing their composition and current applications. Furthermore, the article analyses corrosion modelling using other methods and examines the potential avenues for optimisation that may provide to ML. Additionally, it considers the cost aspect and provides potential methods and suggestions for reducing costs. This review can serve as a valuable reference for researchers studying corrosive pipeline modelling.