| 引用本文: | 栾亦琳,孙涛,冯吉才,刚铁.TiAl/40Cr扩散焊未焊合、弱接合和微孔缺陷的智能识别[J].材料科学与工艺,2010,18(6):806-809,814.DOI:10.11951/j.issn.1005-0299.20100614. |
| LUAN Yi-lin,SUN Tao,FENG Ji-cai,GANG Tie.Intelligent recognition of unbonded,kissing bond and micropore in TiAl and 40Cr diffusion bonding[J].Materials Science and Technology,2010,18(6):806-809,814.DOI:10.11951/j.issn.1005-0299.20100614. |
|
| 摘要: |
| 为了解决异种材料扩散焊质量超声波检测时,从回波幅度无法判断界面是否存在微小缺陷的问题,采用支持向量机技术构建了扩散焊界面缺陷识别模型.以TiAl和40Cr扩散焊接头为研究对象,采用超声波水浸聚焦法采集扩散焊界面信号,从信号中提取4个特征值,优化样本数量和核参数后,训练扩散焊界面缺陷识别模型.扩散焊试样界面信号经模型识别后,根据C扫描图像的位置重构识别图像.结果表明,该模型有效地识别出未焊合、弱接合和微孔缺陷,3种缺陷的正确识别率分别为93%、90.5%和91.5%,识别图像直观地显示了界面的缺陷.TiAl/40Cr扩散焊界面缺陷识别模型实现了扩散焊试样界面缺陷的智能识别. |
| 关键词: 扩散焊 缺陷检测 支持向量机 |
| DOI：10.11951/j.issn.1005-0299.20100614 |
| 分类号:TG453.9 |
| 基金项目:国家自然科学基金资助项目(51075097) |
|
| Intelligent recognition of unbonded,kissing bond and micropore in TiAl and 40Cr diffusion bonding |
|
LUAN Yi-lin1, SUN Tao1, FENG Ji-cai2, GANG Tie3
|
|
1.School of Mechatronics Engineering,Harbin Institute of Technology,Harbin 150001,China;2.School of Materials Science & Engineering,Harbin Institute of Technology at Weihai,Weihai 264209,China;3.State Key Laboratory of Advanced Welding Production Technology,Harbin Institute of Technology,Harbin 150001,China
|
| Abstract: |
| To solve the problem that the tiny defects could not be evaluated by the amplitude of the interface signal,when using conventional ultrasonic inspection in diffusion bonding of dissimilar materials,a defects recognition model was established based on support vector machine.By taking the TiAl and 40Cr diffusion bonding joints as the investigated object,the ultrasonic signals reflected from the diffusion bonding interface were collected by immersion focusing testing method.Four characteristics were extracted from ultrasonic signals.The defects recognition model was trained after the optimization of specimen amounts and kernel parameters.The TiAl and 40Cr diffusion bonding interface signals were recognized by the model and the recognition results were reconstructed according to the position of the ultrasonic C-scan images.It is found that the defects of unbonded,kissing bond and micropore are identified by the model well and the recognition ratios of unbonded,kissing bond and micropore are 93%,90.5% and 91.5%,respectively,which shows that the diffusion quality can be revealed by the recognition model. |
| Key words: diffusion bonding defects detection support vector machine |
