长输油气管道是国家的能源命脉，随着新建管道规模增大及在役管道进入老龄化，因管线完整性问题引发的事故频发，严重威胁管道的安全稳定运营。管道环焊缝开裂是油气管道的主要失效形式之一，现有超声和漏磁内检测技术，因受检测环境、环焊缝形貌不规则、检测时间窗口短等诸多因素限制，难以实现环焊缝裂纹缺陷的精确识别和量化检测，导致管道环焊缝裂纹检测成为制约管道安全的世界性难题。本项目针对在役管道环焊缝裂纹非开挖检测这一世界性难题，提出一种基于磁光成像的管道环焊缝裂纹高速内检测方法，构建高速动态下焊缝非均匀复杂结构的漏磁、涡流等多物理效应相互作用机理模型，设计基于物理模型与数据模型融合的多源耦合乘性信号多维度特征分离算法，实现管道环焊缝裂纹的高空间分辨率可视化高速检测和自动识别，为在役长输油气管道安全稳定运行和管线完整性评估提供有力支撑。

Oil and gas pipeline is the lifeblood of China. With the larger scale of new pipelines and the aging of pipelines in service, due to the frequent occurrence of accidents caused by pipeline integrity problems, the safe and stable operation of pipelines is seriously threatened. Crakes in girth weld is one of the main failure forms of oil and gas pipelines. Due to the limited inspection environment, irregular cracks in girth weld, short inspection time window, etc., the existing ultrasonic technique and magnetic flux leakage in-line inspection technique are difficult to accurately identify and quantify the defects of crakes in girth weld, which leads to the correlative inspection has become a worldwide problem restricting pipeline safety. The project is aimed at the worldwide problem of trenchless inspection of crakes in girth weld of pipelines in service, proposes a high-speed magneto-optical imaging in-line inspection method for cracks in girth weld of pipelines, builds a mechanism model of multi-physical effects such as magnetic flux leakage and eddy current in the non-uniform and complex structure of the weld, designs physical model and data model based multi-dimensional feature separation algorithm of multivariate coupling multiplicative signal. In this way, high-spatial-resolution visual high-speed inspection and automatic identification of cracks in girth weld of pipelines are realized, which provides a strong support for the safe and stable operation of long-distance oil and gas pipelines and pipeline integrity assessment.