高级检索
代欣欣, 郑俏俏, 季玉坤, 高向东. 焊接缺陷磁光成像磁场分布数值模拟与试验分析[J]. 焊接学报, 2019, 40(12): 102-108. DOI: 10.12073/j.hjxb.2019400321
引用本文: 代欣欣, 郑俏俏, 季玉坤, 高向东. 焊接缺陷磁光成像磁场分布数值模拟与试验分析[J]. 焊接学报, 2019, 40(12): 102-108. DOI: 10.12073/j.hjxb.2019400321
DAI Xinxin, ZHENG Qiaoqiao, JI Yukun, GAO Xiangdong. Numerical simulation and experimental analysis of magnetic field distribution of magneto-optical imaging in weld defects[J]. TRANSACTIONS OF THE CHINA WELDING INSTITUTION, 2019, 40(12): 102-108. DOI: 10.12073/j.hjxb.2019400321
Citation: DAI Xinxin, ZHENG Qiaoqiao, JI Yukun, GAO Xiangdong. Numerical simulation and experimental analysis of magnetic field distribution of magneto-optical imaging in weld defects[J]. TRANSACTIONS OF THE CHINA WELDING INSTITUTION, 2019, 40(12): 102-108. DOI: 10.12073/j.hjxb.2019400321

焊接缺陷磁光成像磁场分布数值模拟与试验分析

Numerical simulation and experimental analysis of magnetic field distribution of magneto-optical imaging in weld defects

  • 摘要: 以激光对接焊的焊接缺陷为对象,研究基于数值模拟的焊接缺陷漏磁场的分析方法. 建立对接焊焊接缺陷检测的三维模型,利用漏磁场理论对比分析不同几何缺陷与漏磁场信号之间的关系规律,并用试验进行验证. 结果表明,裂纹的深度越深,磁感应强度越大,未熔合、凹坑分别随着角度、宽度的增大而磁感应强度减小,并且验证漏磁场信号可以作为焊接缺陷检测的依据. 采用RGB分割法对磁光图像进行分割并提取几何特征,用模糊C-均值聚类(FCM)对不同焊接缺陷进行识别,有良好的识别率.

     

    Abstract: Taking the weld defects of laser butt welding as the object, the analysis method of magnetic leakage field of weld defects was studied based on numerical simulation. The three dimension finite models of the weld defects detection of laser butt welding was established by Magnet. The relationship between different geometrical defects and their leakage magnetic field signals was compared and analyzed by using the theory of leakage magnetic field, and verified by experiments. The numerical simulation results show that the deeper the crack depth is, the larger the magnetic induction intensity is. The magnetic induction intensity decreases with increase of the angle and width of unfused and concave pits. Magnetic leakage signal can be used as the basis of weld defect detection. Besides, the magneto-optical image was segmented by RGB segmentation method and the features were extracted. Fuzzy c-mean clustering (FCM) was used to identify different weld defects. The results show that the recognition rate is good.

     

/

返回文章
返回