The Application of a Non-Linear Model for Evaluating the External Corrosion Protection of a Gas Pipeline

Abstract

Since the majority of pipelines in China are required to be installed underground, external corrosion becomes a major threat to their useful lifetime. The factors affecting the likelihood and severity of external corrosion of buried pipelines includes soil corrosivity, coatings, presence of moisture (e.g., river crossing), electrical fields, cathodic protection system, coatings, etc. In order to help understand the potential impact of this problem, a survey of a 70-km gas pipeline was conducted, including measurements of soil corrosivity, soil resistivity, stray electrical currents, IR drop, depth of cover, coating integrity, and the effectiveness of cathodic protection. Additional information was collected regarding interference potential by Pipeline Current Mapper (PCM), Closed Interval Potential Survey (CIPS), and Direct Current Voltage Gradient (DCVG) techniques. The collected information was integrated into a database which was used to build a non-linear model using the Support Vector Machine (SVM) method. By investigating the non-linear relationship among these factors with respect to their influence on the external corrosion of pipelines, it is possible to predict the most likely locations of external corrosion along other pipelines. The model was applied to another (50 km) pipeline and the results indicate the method is accurate and efficient.