管道焊缝缺陷多场耦合漏磁行为机理及诊断与量化研究

Magnetic flux leakage testing technology has urgent demand and broad application prospects in the petroleum and petrochemical pipe welds. The basic theory and key technology of the pipeline weld magnetic flux leakage is study from two directions of forward modeling and inversion in this project. The forward direction is the study of leakage behavior mechanism for the pipeline weld of magnetic flux leakage detection, and the inverse direction is the study of pipeline weld defect type diagnosis and defect quantitative. The forward provide conditions and theoretical basis for the inversion, through the forward descript corresponding relation between the defects and generate signals, and then the inversion according to corresponding relation, by detecting signal are deduced with the type of defect and the geometric feature. Mechanical behavior model of the fluid-structure interaction effect between the contact area of the internal defects of the pipe weld and the conveying fluid is set up first time both at home and abroad, and then multi-fields coupling mechanism under the magnetic field is revealed. Compare the theory simulation results through the experiment research by using detection system of this project development. In view of the superposition problem of magnetic flux leakage signal between the weld and the defect, convert signal data into grayscale digital graphics in the form of matrix. Aiming at the problem of weld defect diagnosis recognition, based on gray level gradient - primitive co-occurrence matrix feature extraction method and fusion diagnosis method based on genetic algorithm is put forward. And aiming at the problem of weld defect quantification, based on feature extraction methods of mathematical morphology and chain code algorithm fusion and quantitative defect size based on BP neural network method is put forward. The project will provide theoretical basis and method basis to explain the pipeline weld magnetic flux leakage behavior mechanism and weld defect visual qualitative and quantitative evaluation.

焊缝漏磁法检测技术在石油石化管道中有迫切的需求和广阔的应用前景。本项目从正演和反演两个方向，研究管道焊缝漏磁的基础理论和关键技术，正演研究漏磁法检测管道焊缝的漏磁行为机理，反演研究管道焊缝缺陷类型诊断与焊缝缺陷量化。正演为反演提供条件与理论基础，通过正演表征缺陷与产生信号的对应关系，然后反演根据相对应的关系、由检测出的信号反推出缺陷的类型与几何特征。在国内外首次对管道焊缝内壁缺陷空间接触面与输送流体流固耦合作用的力学行为进行建模描述，揭示在磁力场施加下的多场耦合机理。利用本项目所研发的检测系统对部分理论仿真结果进行实验对比研究。针对焊缝漏磁信号与缺陷漏磁信号叠加的难题，提出将信号数据转换成灰度矩阵形式的数字图形。针对焊缝缺陷诊断识别，提出基于灰度-梯度-基元共生矩阵的特征提取方法及基于遗传算法的融合诊断方法，针对焊缝缺陷量化，提出基于数学形态学和链码算法融合的特征提取方法及基于BP神经网络量化缺陷尺寸方法。项目预期为解释管道焊缝漏磁行为机理及焊缝缺陷可视化定性与定量评价提供理论基础和方法依据。

油气管道泄漏爆炸的事故中，管道焊缝缺陷是影响管道安全工作与否的主要因素之一。当管道焊缝存在裂纹缺陷时，由于输送流体侵入缺陷所在的管道焊缝空间，在流体冲蚀下，加剧裂纹扩展，在进行漏磁检测时，这是管道焊缝内壁缺陷空间接触面与输送流体的流固耦合作用，并伴随磁力场的多场耦合问题。在此背景下，首先，本项目利用虚拟裂纹闭合技术（Virtual crack closure technique, VCCT），研究了随着管道焊缝裂纹动态扩展的漏磁行为的流固磁多场耦合算法。其次，确定了适合管道焊缝结构特点的检测系统模式，采用本项目开发的具有自主知识产权的管道焊缝连续非接触漏磁扫描检测仪进行了相关实验研究。然后，探寻了基于共生矩阵的管道焊缝缺陷漏磁图形特征提取方法，实现了管道焊缝缺陷定性研究。进一步，给出了基于数学形态学和链码算法融合的管道焊缝缺陷定量研究方法，实现了管道焊缝缺陷可视化反演。. 在本项目的资助下，项目负责人以第一作者身份，本项目发表SCI期刊论文7篇、EI期刊论文3篇；项目负责人以第一发明人身份，本项目授权发明专利6项；项目负责人以第一著作权人身份，本项目授权软件著作权6项。项目成果在理论上，丰富了流固磁多场耦合机理的相关内容；在实践上，研制了管道焊缝连续非接触漏磁扫描仪，开发了管道焊缝反演软件，为管道焊缝缺陷定性定量研究提供了一种新途径，对解决漏磁法检测管道焊缝裂纹动态扩展问题及进一步推广图形识别算法在无损检测领域量化评估中的应用具有重要意义，在国家安全生产事故防治领域具有广阔的工程应用前景。