In the absence of in-line inspection (ILI), which of the following is the most useful for predicting the occurrence of external corrosion on buried pipelines?

(a) Design and construction information (e.g. pipe age, coating type, pipe grade)
(b) Operational parameters (e.g. temperature and pressure)
(c) Environmental data (e.g. soil properties, terrain, climate)
(d) Above-ground survey data (e.g. cathodic protection and coating survey results)


This question is a little deceptive, as there really is no correct answer here. Since corrosion occurs due to complex interactions between many different variables, there is rarely one single category of data that can be universally considered as “the most useful.”

But we asked you to pick one answer nonetheless, and it was interesting to see that the majority (60%) of our readers selected above-ground survey data as the most important category.

It’s as good an answer as any.

The efficacy of the cathodic protection (CP) system and the condition of the coating are two extremely important factors that influence external corrosion in buried pipelines – so it stands to reason that these variables should be routinely monitored and used to infer pipeline condition. That’s why they are a core part of External Corrosion Direct Assessment (ECDA) and related techniques.

However, we cannot forget the other types of data. Simple design and construction information – such as age and coating type – are often better predictors of corrosion than anything else, while for a pipeline that has been exposed to its local environment, information about the soil type, chemistry and moisture content can be critical for understanding the occurrence and severity of corrosion.

At ROSEN, we are attempting to capture all of this knowledge within machine-learning models, trained on information from many thousands of pipelines around the world. And although these models offer a robust and elegant solution for external corrosion prediction, they also underline how complex a problem it really is.