Magnetic flux leakage in-line inspection data analysis and integration

油气管道的无损检测是管道完整性管理的基础，也是防止管道泄漏和失效的重要环保措施。漏磁法在线检测是一项关键技术。使用不同配置的检测系统，如轴向MFL和周向MFL，可以提供管道状况的全面评估。然而，由于不同的检测技术固有的异质性，与检测结果相关的不确定性仍然是维护决策的一个挑战，本研究项目是开发机器学习算法来识别ILI数据之间的一致性，探索来自多个检测的共同特征模式。通过对多模态在线检测数据的综合，可以得到不确定性最小的综合评价。因此，本研究将使管道运行的有效管道完整性管理。研究成果将以软件工具的形式提供给工业合作伙伴，以便商业化并更好地服务于加拿大石油和天然气行业的需求。

The non-destructive inspection of oil & gas pipelines is essential for pipeline integrity management and criticalfor environmental protection from pipeline leakage and failure. The in-line inspection (ILI) with magnetic fluxleakage (MFL) method is a key technology. The use of differently configured inspection systems, such as axialMFL and circumferential MFL, can offer a comprehensive assessment of pipeline conditions. However, theuncertainty associated with inspection results remains a challenge for maintenance decision making due to theinherent heterogeneities with varied inspection techniques.This research project is to develop machine learning algorithms to identify the consistency between the ILI databy exploring the common feature patterns from multiple inspections. A comprehensive evaluation with leastuncertainty can be achieved by integrating multi-modal in-line inspection data. Thus, this research will enableefficient pipeline integrity management for pipeline operation. The research outcomes will be in the form ofsoftware tools, which can be delivered to the industrial partner for commercializing and better serving theneeds of Canada's oil & gas industry.