Abstract
Submarine pipelines in the sea are applied for oil, gas, water and mixed transportation. Among them, 91% of the pipes contain CO2. Here, based on the existing pipeline internal inspection data of submarine pipeline, the APRIORI algorithm and least-square-support-vector-machine (LSSVM) are applied to analyze the distribution rules and defect characteristics of internal defects along the pipeline. The corrosion defects are divided into 7 types and the pipeline section is divided into 12 intervals. Also, the pipe segment has been defined as J (general pipe), W (weld) and C (close to weld). The contents include the analysis of the characteristics and types of defects, the distribution of defects along the pipe, the severity of the corrosion defects, the size characteristics of defects, and the comparison of the data detected in multiple rounds. The defect depth of four kinds of pipelines is mostly 10%–20% of the wall thickness, hereby the severity of defects is studied via the percentage distribution of corrosion depth. The data of multi-round inspection shows that the corrosions in the mixed pipeline are active and the defects are increasing. The methods and results in this paper can be employed to predict the most likely defect type, mileage location, clock orientation, and shape size of submarine pipeline corrosion. This is helpful for the integrity management of submarine pipelines.
