【作者】 谷晓鹏；

【导师】 徐国成；

【作者基本信息】 吉林大学 ， 材料加工工程， 2013， 博士

【摘要】 本文对不锈钢薄板激光焊搭接接头的超声波检测方法及理论进行了系统研究。结果表明，对应焊接接头不同部位扫描点的A扫描信号存在显著差异。在母材区域，上层钢板下表面处的反射回波幅值最高；当探头逐渐接近焊缝中心时，由于反射面积的减少，反射回波幅值随之降低；当探头位于焊缝中心时，反射回波幅值达到最低值。同时，A扫描信号的频谱分析结果表明，当探头由母材区域移动至焊缝中心时，信号的主频率由2MHz变为1.25MHz，即反射源由上层钢板下表面变为下层钢板下表面。A扫描信号中两次底面回波之间的散射波分析结果表明，焊缝与母材之间的熔合线界面处不存在明显的反射特征，这也说明在使用15MHz的纵波探头对不锈钢搭接接头进行检测时，不能够直接获得焊缝的熔深信息。原始A扫描信号的小波分析结果表明，信号的主要成分为上层钢板下表面的反射回波，能量主要集中在信号的低频部分；随着频率的升高，反射回波特征逐渐减弱。使用小波包技术对A扫描信号的高频部分进行进一步的分解，确定上层钢板下表面的反射回波集中在0~6.25MHz频带内；当探头由母材向焊缝中心移动时，反射回波幅值随反射面积的减少而降低，这种变化在0~37.5MHz频带内均有不同程度的表现。同时发现，晶粒散射波主要集中37.5~43.75MHz频带内，当探头由母材向焊缝中心移动时，晶粒逐渐粗大，散射波有变强的趋势。超声波检测过程的有限元仿真分析结果表明，上层钢板下表面的反射回波幅值与声束的反射面积呈正比；当探头位于熔合边缘时，回波强度衰减度的仿真计算结果为-6.77dB，与实测信号基本一致。为弥补传统的6dB法的不足，本文提出以等效宽度为目的的计算模型，计算结果表明，检测区域接头内部两层钢板接触面处的熔宽计算精度可以达到0.05mm，完全满足工业条件下焊接质量在线检测的需求。更多还原

【Abstract】 Along with the rapid development of automobile and rail transit manufacturing, themanufacture technology of bodywork based on laser welding has been researched anddeveloped in the main railway car manufacture enterprise at home and abroad to increase thecompetiveness of high quality vehicles. Lap joint is mostly used in the structure of stainlesssteel bodywork. Partial penetrated laser welding is mainly used to acquire the joint. Duringlaser welding, defects such as incomplete fusion and insufficient penetration easily appeardue to the installation ways and the abnormal fluctuation of welding parameters. Therefore,laser welding process control and quality evaluation is a key problem that limits theapplication of laser welding technology.Two kinds of method are mostly used for laser welding evaluation. They are onlineevaluation based on the process control and postweld destructive inspection. The former canprovide quality information partly and can not represent the quality information of jointsaccurately. The latter makes destruction of the joint. Its cost is high and the testing period islong. It can be applied in experimental research and is not suitable for welding production.Thus, efficient and reliable nondestructive testing methods by ultrasonic should bedeveloped to analyze the weld quality of laser welded lap joints of thin stainless steel sheet.Further study on imaging principle, signal de-noising and weld width calculation method hashigh academic and practical value.In this paper, an ultrasonic testing system for laser welding was developed. A-scansignal variation at different positions in the joint was studied using the longitudinal wavereflection method. Experimental results show that there exists relationship between thereflection echo amplitude from the lower surface of the upper steel sheet and the fusion statein the joint. When the probe is at the base metal region, the highest amplitude is received.When the probe moves to the joint centerline, the amplitude decreases due to the decrease ofthe reflection area. When the probe is at the joint centerline, the amplitude is the smallest.Meanwhile, the spectrum analysis of A-scan signal indicates that when the probe movesfrom the base metal region to the joint centerline, the main frequency ratio of the signalchanges from2MHz into1.25MHz. It means that the reflection source changes from thelower surface of the upper steel sheet into the lower surface of the lower steel sheet.When scanning across the joint, there exists relationship between the reflection echo amplitude from the lower surface of the upper steel sheet and the weld width in the joint.C-scan image constructed according to the reflection echo can represent the fusion state inthe joint. However, the analysis on scattered wave of the reflection echo show that there isno reflection characteristic at the fusion line between the base metal and the weld metal. Itmeans that the information of weld penetration can not be received when15MHz longitudeprobe is used to test the laser welded lap joints of stainless steel sheet. Welding parametersand materials characteristic should be in considered to estimate the weld width.In order to further understand the detail characteristic of A-scan signal, wavelettransform technology is used to decompose the original signal. The spectrum composition ofA-scan signal and the jump feature in time domain are analyzed. Three level waveletdecomposition of db4result shows that the signal mainly comprises reflection echo from thelower surface of the upper sheet. The energy mainly exists in the low frequency part. Withthe increase of the frequency, the reflection echo decreases. Through further decomposingthe A-scan signal at high frequency part by means of wavelet transform, it can be conformedthat the reflection echo mainly exists in the frequency range of0~6.25MH. When the probemoves to the joint centerline, the amplitude decreases due to the decrease of the reflectionarea. Such change appears at the frequency range of0~37.5MHz. it is found that thegrain-scattered wave mainly exists in the frequency range of37.5~43.75MHz. When theprobe moves from the base metal to the weld, the grain coarsening causes increase of thescatter wave. Wavelet soft-threshold denoising method is employed to eliminate noise of theA-scan signal based on wavelet decomposition. It can eliminate the effect of scattering noisein the high frequency range so as to strengthen the characteristic of reflection echo. Thus, thequality of C-scan image can be increased.Finite element simulation model of ultrasonic testing on laser welding joints of stainlesssteel sheet is constructed to further analyze the propagation characteristics of ultrasonic inthe joint and the scattering rules in the joint. Experimental result show that when the probemoves to the joint centerline, the amplitude decreases due to the decrease of the reflectionarea. When the probe is at the fusion line, the simulation value of reflection echo attenuationis-6.77dB, which is consistent with the measured value. When the weld width is smallerthan the ultrasonic sound wavelength, the intensity of the reflection echo does not changeobviously. Focusing ultrasonic probe with high frequency and fine ultrasonic beam should bechosen to acquire relatively small lateral resolution. Moreover, testing process with a tiltprobe is simulated. The result shows that the incident angle of the probe has effect on theamplitude of reflection echo from the lower surface of the steel sheet. With the increase ofthe incident angle, the amplitude decreases. There exists linear relationship between the relative attenuation degree and the incident angle.In order to obtain quantitative testing of the weld width, C-scan image of the joint isprocessed and the joint centerline is automatically extracted to construct the model forcalculating the weld width. The analyzing result shows that there exists step edge in C-scanimage constructed according to the reflection echo from the lower surface of the upper sheet.Edge testing of the original image is applied by Roberts operator. Thus the image isdecomposed subsets with obvious weld characteristics. Equivalent centerline of the weld canbe extracted from the decomposed C-scan image by Hough transformation technique. Theend and the discontinuity of the weld can be recognized considering that the weld is round.Thus the endpoint location of the weld can be determined.In the whole scanning region, there exists corresponding relationship between theamplitude of the reflection echo from the lower surface of the upper sheet and the scanningpoint. Therefore, amplitude-position curve is established statistical principle. It can representthe information of the equivalent weld width. The curve is simplified and fitted by leastsquare method. Finally, calculation formula of equivalent weld width is constructed.Calculating result shows that rapid convenient calculation of the weld width can be acquired.It can meet the demand of online testing in the industrial production with high calculationaccuracy and reliability.Further understanding of the ultrasonic testing method and theory for laser welded lapjoints of stainless steel sheet is obtained. It can provide detailed experimental data andtheoretical basis for ultrasonic testing.更多还原