【作者】 黄江中；

【导师】 刚铁；

【作者基本信息】 哈尔滨工业大学 ， 材料加工工程， 2009， 硕士

【摘要】 随着现代工业技术的高速发展,厚壁钢结构越来越广泛地应用于机械工业、核工业及航天航空等领域。由于低合金结构钢具有较高的强度和较好的塑性与韧性,其成为厚壁焊接结构中常用的材料。超声TOFD检测技术以其原理简单,操作方便,检测速度快,对焊缝中裂纹类缺陷检出率高,特别适合于大厚件焊缝检测等特点越来越受到人们的青睐。因此,对中厚板低合金结构钢焊缝的TOFD检测就具有重要的理论和实际意义。本文研制了TOFD检测系统的手动扫查装置,开发了基于USB接口的便携式超声TOFD采集系统,该系统具有结构简单,操作方便等特点,适合于工程现场检测与分析。由于工件表面不平整等原因使得探头移动不够平稳,从而造成检测图像发生畸变现象。本文提出了采用互相关分析法和灰度差法对检测图像进行矫正。研究结果表明:该方法能够很好地矫正图像,使侧向波保持在相同的水平位置,为缺陷的准确定量定性和目标提取做准备。为分析TOFD法检测图像特征,本文选用24mm厚度的低合金钢板,采用手工电弧焊制备出大量的缺陷样本,共计129个,包括气孔、夹渣、裂纹及未焊透等。利用开发的系统采集了上述缺陷的超声信号,分析总结了典型缺陷检测图像特征。与射线照相底片相对比,TOFD检测对缺陷性质的判断大致与射线检测相符合。在图像矫正基础上,提取了检测图像中焊缝区域,并自动将检测图像进行线性化处理。这种处理有利于识别缺陷的深度及高度,简化了检测过程,减少了人为因素的干扰,提高了缺陷定位的精度。为了提高工作效率、减低误判率,利于后续缺陷特征提取与分类识别,本文采用基于背景差法进行缺陷目标提取。结果表明:该方法能够很好地模拟并去除TOFD法检测图像的背景并提取出缺陷。更多还原

【Abstract】 Recently thick-walled steel structures are widely used in machinery industry, nuclear industry, aerospace and other fields. Low-alloy structural steel with high strength as well as good ductility and toughness, are commonly used in thick-walled structures. As an accurate and rapid testing method, TOFD has advantages on weld defects testing for thick plate, especially on crack defects. It has been widely concerned. Therefore, the TOFD testing for thick plate weld of low-alloy structural steel is significant of theory and practice.A USB-based portable type ultrasonic TOFD acquisition system was developed in this paper, and the manual scanning device of the system was also designed. The developed system is easy to operate, simple for structure, and suitable for work-site detection and analysis.Because of the unsmooth sample surface, the probes motion is not easily stable, which result in testing image distortion. Therefore, the cross-correlation analysis and the gray-level difference method were used to calibrate these images. The results show that the methods are able to correct the images, which maintain the lateral wave in the same level that prepares for accurately quantitative, qualitative analysis and object extraction for defects.For TOFD testing image feature analysis, samples of a large number of defects were prepared using low-alloy steel of 24mm thickness plate by arc welding, which haves a total of 129, including the pores, slag, cracks, lack of penetration and so on. The ultrasonic signals of the above-mentioned defects were collected by the developed system, and the characteristics of typical defects in the testing images were analyzed and summarized. The identification of defects properties by TOFD technology approximately agree with radiographic film.In this paper, weld zone was extracted from the calibration image, and linear algorithm was used for automatic processing the image, so as to help for identification the defect depth and height, simplifying the testing process, and increasing the accuracy of defect location and quantity. For increasing of efficiency, decreasing the misjudge rate and defect features extracting and classification, differential analysis between background and defect target was applied to extract the target of defects and image segmentation. The result shows that this method not only can well simulate and remove the background of TOFD D scan image, but also extract the defect effectively.更多还原