Estimating corrosion growth rates for underground pipelines is a challenging problem. There are confounding variables with complex interaction effects that may result in unexpected outcomes. For instance, the relationship between soil conditions and AC interference is highly non-linear and challenging to model. This work expands upon prior work using a suite of machine learning tools to estimate corrosion rates. However, instead of estimating a single corrosion growth rate for a single girth weld address (GWA), this work estimates a distribution of potential corrosion growth rates. Modeling distributions provide a more effective risk-measurement framework, especially concerning high volatility or areas of severe tail risk.
This work relies heavily on machine learning and geospatial tools - particularly artificial neural networks and gradient boosted trees to estimate the corrosion rates and non-linear processes. Building upon prior work using data from a North American Operator, the models in this paper use additional variables from recent research in AC interference and microbiologically influenced corrosion to construct a higher accuracy and distribution-based model of pipeline corrosion risk.
One of the pillars of the fourth industrial revolution (4IR) is to let machines make decisions on behalf of humans; this paper describes new technology that allows machines to decide inspection programs and field validation and testing of results. The technology described is a part of integrity management, and uses data, statistics and expert decisions to design inspection programs. These inspection programs are an important part of the safeguarding of equipment to maintain production and safety.
This technology is a data-driven predictive model of material loss from corrosion, based on domain expert input and historical data in the form of non-destructive testing (NDT) tests. The technology trends is based on historical data and SME input, while accounting for uncertainties in NDT measurements, with uncertainties in historical trends and uncertainties in future trends. This produces a more realistic failure prediction to enhance existing RBIs and adds safety by improving on early detection of trends in data. In total, this enables the machine to update inspection plans autonomously, reducing the number of inspections significantly.
The paper also describes how the technology can be developed further to use production data and integrity operating windows to improve predictions, deal with localised corrosion and assess if the test points on a corrosion circuit are sufficient, can be reduced in number or should be manually evaluated by adding more test points.