文章摘要
于润桥,卢超.ART-2网络在超声检测信号分析中的应用[J].,2007,26(1):60-63
ART-2网络在超声检测信号分析中的应用
Application of ART-2 network in the signal analysis of ultrasonic testing
  
中文摘要:
      本文以焊缝超声检测常见的缺陷(夹渣、气孔、未焊透、未熔合)为对象,选取缺陷时域波形的上升时间、波宽、上升角度和频域特征的波动次数、底宽等特征值组成特征矢量作为ART-2的输入,来实现缺陷性质的识别,并与射线检测结果对比。实验结果表明:采用该方法得到的正确识别率达到81.5%,具有良好的应用前景。
英文摘要:
      Based on the four types of weld flaws in Ultrasonic testing,this paper presents the ART-2 method to recognize the defect pattern.Five characteristics in the time do- main and frequency domain are chosen as the input vector for the ART-2 network.The results were compared with those of X-Ray testing.Experimental results show that the recognition rate can reach 81.5% due to the method and it has a future for applications.
DOI:10.11684/j.issn.1000-310X.2007.01.013
中文关键词: 超声波检测  焊缝缺陷  ART-2网络  模式识别
英文关键词: Ultrasonic testing  Welding defects  ART-2 network  pattern recognition
基金项目:
作者单位
于润桥 南昌航空工业学院无损检测技术教育部重点实验室,南昌,330034 
卢超 南昌航空工业学院无损检测技术教育部重点实验室,南昌,330034 
摘要点击次数: 2534
全文下载次数: 136
查看全文   查看/发表评论  下载PDF阅读器
关闭