不可见焊接缺陷旋转场能堆积磁光成像检测关键问题研究

Targeting the invisible defects inside weld seam and on its sub-surface of ferromagnetic substance, this project proposes an innovative approach for the nondestructive inspection of weld defects during magneto-optical imaging with controllable magnetic field energy. The rotating magnetic field is controlled and the weld seam is then magnified, which cause magnetic vector nearby the weld seam to change periodically and result in magnetic field energy accumulation. By utilizing Faraday magneto-optical effect, magnetic field distribution nearby the weld defects can be viewed in the form of images. A spectrum measurement model of weld seam defects is established based on amplitude parameters and phrase parameters of frequency domain spectrum, which helps to realize automatic identification of such welded defects as crack, porosity, incomplete fusion and surface micro crack inside weld seam. This project investigates how weld seam defects affect the physical process of rotating magnetic field and the representation mechanism of magnetic field energy accumulation. It also specifies the internal pattern and mapping correlation between the magnetic field nearby weld seam defects and magneto-optical imaging, and constructs the feature descriptions of amplitude parameters and phrase parameters based on the power spectrum density of magneto-optical images. The inter features of invisible weld defects are presented from the perspective of energy representation mapping of weld seam magneto-optical images, and a quantitative recognition algorithm of weld defects with certain identification competence is also established. The project is expected to provide theoretical basis and experimental reference for the automatic identification of weld defects. It has been proved to be of significant scientific and practical values in solving problems arising from the non-destructive testing of weld defects, ensuring weld quality and preventing major accidents.

针对肉眼无法识别的焊接结构件内部和亚表面不可见焊接缺陷，研究一种可控场能状态下磁光成像的焊接缺陷无损检测新方法。通过控制旋转磁场能量，使焊接缺陷处磁场矢量周期性变化并形成场能堆积，同时应用法拉第磁光效应使缺陷处磁场畸变式分布以图像形式表现，再以频域功率谱幅值参数和相位参数构建焊接缺陷谱图，建立中心切片磁光成像傅里叶变换和缺陷几何形态重构算法，实现焊缝内部裂纹、气孔、未熔合等缺陷的识别。项目探索焊接缺陷对旋转磁场物理过程的扰动规律及场能堆积表征机理，揭示焊接缺陷处磁场畸变及场能堆积与磁光成像间的内在规律，从磁化区域场能变化及磁光图像能量表征映射的角度探索不可见焊接缺陷的内在特征，建立具有显像和定位能力的焊缝内部缺陷识别算法。项目为不可见焊接缺陷的识别提供理论基础和科学依据，对解决焊接缺陷无损检测问题、保证焊接质量和避免重大事故具有重要科学和现实意义。

项目针对肉眼无法识别的焊接结构件内部和亚表面不可见焊接缺陷，研究可控多场能状态下磁光成像的无损检测新方法。通过试验和分析，探索了多场能（交变磁场、旋转磁场、变频磁场及组合磁场）激励下不可见焊接缺陷磁光成像机理、缺陷场能堆积与磁光图像特征参数之间的内在规律，同时应用法拉第磁光效应使焊接缺陷处磁场畸变式分布以图像形式表现。以频域功率谱幅值参数和相位参数构建焊接缺陷谱图，建立磁光成像傅里叶变换和缺陷几何形态重构算法，实现焊缝内部裂纹、气孔、未熔合等缺陷的识别。针对噪声统计特性的不确定性和模型误差，探索了适用于不可见焊接缺陷可视化无损检测的磁光成像系统，研究出基于磁光成像特征的焊缝缺陷识别算法。项目揭示焊接缺陷磁场畸变及场能堆积与磁光成像间的内在规律，建立磁光成像的不可见焊接缺陷可视化无损检测识别分类模型，为焊接缺陷的无损检测、保证焊接质量和避免重大事故提供了理论和试验依据，具有重要科学和现实意义。项目研究成果共发表学术论文89篇，其中国际期刊SCI论文35篇，EI期刊论文24篇。