(222d) Development of AI Algorithms for Corrosion Prediction in Midstream Industry

Corrosion has been causing the mid-stream industry billions of dollars per year. It is highly desirable to predict pipeline corrosion. For natural gas pipelines, the species cause most concerns on corrosion are water, hydrogen sulfide, carbon dioxide, and oxygen. In addition, the temperature, pressure, and velocity also make impacts. For crude oil pipelines, the species are more complicated.

In this research, the corrosion rate data was generated from sound first-principles based models. Then accurate AI algorithms were developed to predict the corrosion rates in natural gas and crude oil pipelines. This research will help bridge the gap from “post-corrosion” measurement to proactive corrosion management, revolutionize the practice of corrosion management, enhance the reliability of mid-stream assets, improve operation efficiency, and reduce potential environmental risks.