厚壁CASS超声检测建模和PCI降噪算法研究

发布时间：2018-03-26 03:14

本文选题：铸造奥氏体不锈钢　切入点：仿真　出处：《大连理工大学》2016年博士论文

【摘要】：铸造奥氏体不锈钢(Cast Austenitic Stainless Steel,CASS)以其良好的断裂韧性、耐腐蚀性和焊接性,被用作压水堆核电站一回路壁厚66～96mm主管道建造材料。核动力装置的检验规范中,对CASS管道加工质量及服役期间状况提出无损检测的强制性要求。目前,现有检测标准要求对厚壁CASS管道进行射线检测和渗透检测。然而,射线检测具有对危害性大的面积型缺陷不敏感、无法进行缺陷定位、具有放射性等缺点。渗透检测无法对亚表面及内部深层缺陷进行检测。相比于射线和渗透检测,超声检测(Ultrasonic Testing,UT)具有对面积型缺陷敏感、能够对深层缺陷定位定量等优点。但是,CASS具有粗大的弹性各向异性等轴、柱状和等轴-柱状混合结构。超声检测过程中会引起严重的散射衰减和结构噪声。当CASS管道中缺陷深度达到50 mm或更深时,缺陷回波声能很弱,检测信噪比甚至小于10 dB,缺陷误检和漏检概率显著增加。正因如此,厚壁CASS主管道的超声检测是无损检测领域内公认的挑战性课题。对此,在开发专用相控阵超声检测设备和探头的同时,研究者们提出利用超声检测仿真建模和信号／图像后处理算法解决低信噪比问题。利用仿真建模描述CASS的超声检测过程,有助于弄清超声波在粗晶结构中的传播机理,为缺陷的有效判别、检测参数优化和检测设备开发提供理论指导。然而,建模过程中晶粒结构、晶体取向和材料弹性参数的量化描述十分困难,使得模拟结果与实际情况存在较大出入,无法有效指导实践。发展用于提高信噪比的后处理算法是解决CASS超声检测问题的另一有效途径。现有超声信号／图像后处理算法主要利用信号滤波和延时叠加(Delay and Sum,DAS)图像重建提高信噪比。其中,DAS后处理算法可通过延时叠加提高成像区域内的信噪比。然而,多数DAS成像的信号采集过程是通过单探头B扫查完成的。此方法会导致厚壁CASS管道深处声能较弱,深层缺陷信噪比增强能力有限。本文以CASS超声检测建模和超声图像后处理算法为研究内容,探索提高厚壁CASS超声仿真模拟的有效性和改善信噪比的方法,主要研究内容和结论如下：利用电子背散射衍射(Electron Back-Scattered Diffraction,EBSD)技术对CASS的晶粒结构和晶体取向进行分析,得到了尺寸为96mm×12mm的轴-径向截面EBSD图谱。通过取向相(差)角选择和晶体取向归一化处理,建立了基于EBSD图谱的CASS超声仿真模型。该模型不仅能够通过扫描微区之间的取向差异描述晶粒结构,而且能够通过取向微区对应的欧拉角描述晶体取向,克服以往模型难以量化描述晶粒结构和晶体取向的问题。利用Bond变换法赋以晶粒在对应晶体取向下的刚度矩阵,确定晶体取向与弹性常数之间的对应关系,实现模型中晶粒弹性特征的量化表征。采用时域有限差分法对模型进行数值计算,结果表明,模拟得到的A扫描信号中重现了粗晶结构引起的散射衰减、主频降低和结构噪声等现象,模拟结果与实际检测结果基本一致。模拟相控阵扇扫描图像的噪声幅值和信噪比与实际检测结果基本吻合。利用数值模拟分析晶粒形态(长径比l/d和晶粒生长方向θ)对超声散射程度的影响,结论如下：当l/d=1时,晶粒为等轴晶结构,晶界对声波的散射程度不随θ(0°≤θ90°)的改变发生变化。当长径比l/d1时,晶粒为柱状晶结构,晶界对声波的散射程度随啪增加而增加。超声散射程度随θ的变化幅度随长径比l/d的增加而增加。柱-等混合结构中等轴晶结构的存在会削弱l/d和θ的超声散射程度的影响。上述工作为揭示超声波在弹性各向异性粗晶结构中的散射机理提供了解决思路。为进一步提高厚壁CASS深层缺陷超声检测信噪比,利用一发一收探头所形成的“伪聚焦”声场提高深层缺陷的回波声能。声场对比结果表明,50～70 mm深度上,探头中心距为100 mm的一发一收探头声能在40～75%,宽度为1.5 mm的单探头声能在25～40%。在此基础上,利用相位相干成像(Phase Coherence Imaging,PCI)算法构建表示DAS图像中各像素点相位分布特征的循环相干因子(Circular Coherence Factor,CCF)和符号相干因子(Sign Coherence Factor,SCF),通过动态加权处理有效降低相位分布散乱的结构噪声幅值。将这种声束传播路径优化设计和图像后处理算法相结合的方法用于壁厚78 mm的CASS焊缝后,深度60 mm的Φ3边钻孔信噪比由20 dB以下提高到30 dB以上。根据相干因子的特点,SCF能够获得较高的信噪比,但SCF图像中的缺陷影像不连续。CCF能够获得缺陷影像连续的图像,但仅在“伪聚焦”声场内具有较高的信噪比。因此,SCF适用于对检测对象进行初步扫查,CCF适用于对“伪聚焦”声场内的缺陷进行精确定量和定位。
[Abstract]:Cast austenitic stainless steel (Cast Austenitic Stainless Steel, CASS) with good toughness, corrosion resistance and weldability, is used as a loop with thickness of 66 ~ 96mm main pipeline construction materials of PWR nuclear power plant. Test specification of nuclear power plant, the processing quality of CASS pipeline and service condition proposed mandatory lossless detection requirements. At present, the existing detection standards for X-ray detection and penetrant testing of thick wall CASS pipe. However, X-ray detection is not sensitive to the type of damage area of large defects, unable to locate the defect, radioactive and other shortcomings. To detect the penetration of sub surface and internal deep defects were detected. Compared to ray and penetration testing, ultrasonic testing (Ultrasonic Testing UT) is sensitive to the defect area, can the advantages of deep defect localization and quantification. However, CASS has a coarse elastic anisotropy Anisotropy axis, columnar and equiaxed columnar structure. Mixing will lead to serious scattering attenuation and noise in ultrasonic detection process. When the defect of CASS pipeline reached a depth of 50 mm or more, the defect echo sound is very weak, the signal-to-noise ratio is even less than 10 dB, defects and false detection probability of missed detection significantly increase. Because of this, ultrasonic testing of thick wall CASS main pipe is a challenging subject in the field of NDT accepted. In this regard, the development of special equipment and phased array ultrasonic detection probe at the same time, researchers have proposed the use of ultrasonic detection of simulation modeling and signal / image postprocessing algorithm to solve the problem of low signal-to-noise ratio by using the simulation. Describe the modeling process of CASS ultrasonic testing, to help clarify the mechanism of ultrasonic propagation in the coarse grain structure, effectively distinguishing defects, provide theoretical guidance for testing parameters optimization and testing equipment development. However, in the process of modeling The crystal structure, crystal orientation and quantitative description of material parameters is very difficult, so the simulation results with the actual situation, can not effectively guide the practice of development. To improve the signal-to-noise ratio of the postprocessing algorithm is another effective way to solve the CASS problem. The existing ultrasonic ultrasonic signal / image postprocessing algorithm mainly uses signal filtering and the time delay (Delay and Sum, DAS superposition) image reconstruction SNR. The DAS postprocessing algorithm can improve the imaging area through the superimposed signal to noise ratio. However, most of the DAS imaging signal acquisition process is through a single probe B scan completed. This method will lead to the depths of thick walled pipes CASS energy is weak, deep defects of SNR enhancement ability is limited. This paper takes CASS ultrasonic modeling and ultrasonic image postprocessing algorithm as the research content, exploring and improving the simulation of thick wall CASS ultrasound simulation The validity of the method and improve the signal-to-noise ratio, the main research contents and conclusions are as follows: using electron backscatter diffraction (Electron Back-Scattered, Diffraction, EBSD) of CASS crystal structure and crystal orientation are analyzed, the size of 96mm * 12mm - axis radial section of the EBSD map. The phase angle orientation (difference) and the crystal orientation normalization, established the CASS simulation model of EBSD based on ultrasonic map. This model can not only describe the difference between the orientation of grain structure by scanning micro area, but also the orientation of micro area corresponding to the Euler angles to describe the crystal orientation, overcomes the model to quantitative description of grain structure and crystal orientation. By using Bond transform method to Fu grain in the crystal orientation of the corresponding stiffness matrix, to determine the relationship between crystal orientation and elastic constants, elastic model in grain Quantitative characterization. Using FDTD method to model numerical calculation, the results show that the simulated A scan signal obtained to reproduce the scattering caused by coarse grain structure and reduce the structure noise attenuation, frequency and other phenomena, simulation results and actual test results are basically the same. The basic agreement with the measured analog phased array fan scanning image noise amplitude and signal-to-noise. By numerical simulation analysis of grain shape (length / diameter ratio l/d and grain growth direction) conclusion effect of ultrasonic scattering degree is as follows: when l/d=1, the grain equiaxed grain structure, the degree of scattering of acoustic boundary with angle (0 DEG to 90 DEG theta) change. When the ratio of length to diameter l/d1, the grains are columnar crystal structure, degree of acoustic scattering increases with the increase of grain boundary. With ultrasonic scattering degree of variation of theta with slenderness ratio l/d increases. The column - mixed structure The equiaxed grain structure can weaken the effect of ultrasonic scattering degree of l/d and theta. The above work to reveal the mechanism of ultrasonic scattering in the coarse grain structure of elastic anisotropy in providing a solution. In order to further improve the thick wall CASS deep defect ultrasonic signal to noise ratio, using a charge probe formed by the pseudo focus on the defect echo sound field to improve the deep sound field. The comparison results show that the depth of 50~70 mm, the probe center distance of 100 mm sending and receiving probe energy in the 40 ~ 75%, the width of 1.5 mm single probe energy in the 25 ~ 40%. based on coherent imaging using phase (Phase Coherence Imaging, PCI) algorithm for constructing DAS pixels in each phase distribution of the cyclic coherence factor (Circular Coherence Factor CCF (Sign) and Coherence symbol coherence factor Factor, SCF), through dynamic weighted effective reduction The structure of low phase noise amplitude distribution. The scattered beam propagation path optimization design method and image postprocessing algorithm combining CASS mm 78 for weld wall thickness, depth of 60 mm with 3 side drilling signal-to-noise ratio by 20 dB increased to more than 30 dB. According to the characteristics of the coherence factor. SCF can obtain higher signal-to-noise ratio, but the defect image SCF image discontinuities in the.CCF can obtain image defects of continuous image, but only in the "pseudo high focus acoustic signal-to-noise ratio. Therefore, SCF is suitable for detection of initial object scanning, CCF for accurate quantitative and positioning in the the defect of" pseudo focus "in a sound field.

【学位授予单位】：大连理工大学
【学位级别】：博士
【学位授予年份】：2016
【分类号】：TG115.285

【相似文献】