金属板超声层析成像无损检测的研究

发布时间：2018-05-27 03:11

本文选题：金属铝板 + 表面缺陷　；参考：《西华大学》2017年硕士论文

【摘要】：随着社会发展,金属薄板被广泛地应用于各个工程领域当中。金属薄板在使用过程中,因材料本身的缺陷,或一些由金属薄板制造而成的构件和设备在使用过程中产生的缺陷,都会造成一定的安全隐患,甚至引发灾难性后果。因此对金属板进行缺陷探伤是非常重要的环节,它对于安全生产的意义重大。兰姆波在无损检测技术领域中具有快速高效的特点,非常适合用来对大面积薄板状结构进行无损检测。而计算机层析成像技术更是早已被广泛的应用于医学领域当中。本文的主要研究工作是将超声波无损检测技术与计算机层析成像技术相结合,用于有限元仿真中铝板表面缺陷的检测识别,最后通过数字图像处理的相关技术对得到的铝板超声层析成像进行分割和平滑处理,并运用信息熵和区域一致性参数来对图像处理结果进行定量评价。本文的主要研究内容如下:通过广泛查阅国内外文献资料,了解了超声兰姆波在各向同性和各向异性材料中的传播特性,由于本文的研究对象铝板属于各向同性材料,因此在文中只给出了超声兰姆波在各向同性材料中的频散方程和频散曲线。详细地介绍了计算机层析成像技术和文中图像分割所涉及的相关基础理论,包括投影与反投影基本原理、中心切片定理、扇形束卷积滤波反投影算法、脉冲耦合神经网络、遗传算法等,为有限元仿真中超声激励信号走时的提取和铝板超声层析成像及其图像处理提供了理论依据。利用ABAQUS有限元软件建立缺陷大小和位置不同的铝板模型,通过加载五周期汉宁窗口正弦信号来模拟兰姆波在铝板中的传播,提取走时数据。用MATLAB编写扇形束卷积滤波反投影算法程序,将走时数据代入算法中对铝板模型进行重建成像,取得了较好的成像效果,板中的缺陷信息清晰可见。最后对铝板成像运用全局阈值分割、PCNN分割、GA改进PCNN分割这三种不同的图像分割方法进行图像处理,并使用信息熵和区域一致性参数这两个定量评价标准对图像处理结果进行分析和评价。研究结果表明本文使用的方法能够有效的检测出金属铝板模型中缺陷的位置、形状等,也通过对信息熵和区域一致性参数这两个定量评价标准进行数值计算证实了全局阈值分割、PCNN分割、GA改进PCNN分割三种方法的分割效果依次增强,为金属铝板超声层析成像无损检测在实验环境下的进一步研究奠定了良好的基础。
[Abstract]:With the development of the society, metal sheet is widely used in various engineering fields. In the process of using metal sheet, the defects of the material itself, or the defects produced in the use of the components and equipment made by the metal sheet, will cause certain hidden dangers and even disastrous consequences. Plate flaw detection is a very important link. It is of great significance for safety production. Lamb wave has the characteristics of rapid and efficient in the field of nondestructive testing technology. It is very suitable for nondestructive testing of large area thin plate structure, and computerized tomography has long been widely used in medical field. The main research work of this paper is to combine the ultrasonic nondestructive testing technology with the computer tomography technology for the detection and recognition of the surface defects of the aluminum plate in the finite element simulation. Finally, the ultrasonic tomography of the aluminum plate is segmented and smoothed by the related technology of digital image processing, and the information entropy and the region are used in the same area. The main contents of this paper are as follows: through extensive reference to domestic and foreign literature, the propagation characteristics of ultrasonic Lamb waves in isotropic and anisotropic materials are understood. Because aluminum plates in this paper belong to the same homosexual materials, only the ultrasonic orchids are given in this paper. The frequency dispersion equation and dispersion curve of MBO in isotropic materials are described in detail. The basic theories related to computer tomography and image segmentation are described in detail, including the basic principle of projection and back projection, central slice theorem, fan beam convolution filter back projection, pulse coupling neural network, genetic algorithm and so on. In the finite element simulation, the extraction of ultrasonic excitation signal and the ultrasonic tomography of aluminum plate and the image processing provide the theoretical basis. Using the ABAQUS finite element software to establish the aluminum plate model with different defect size and position, by loading the sinusoidal signal of the five period Hanning window to simulate the propagation of lamb wave in the aluminum plate and extracting the travel time data. M ATLAB writes the backprojection algorithm program of fan beam convolution filter, and reconstructs the aluminum plate model in the travel time data algorithm, and achieves better imaging results. The defect information in the plate is clearly visible. Finally, three different image segmentation methods are applied to the aluminum plate imaging using global threshold segmentation, PCNN segmentation, and GA to PCNN segmentation. The image processing results are analyzed and evaluated using the two quantitative evaluation criteria of information entropy and regional consistency parameters. The results show that the method used in this paper can effectively detect the position and shape of defects in the metal aluminum plate model, and also through the two quantifications of information entropy and regional consistency parameters. The value calculation of the evaluation standard confirms that the three methods of global threshold segmentation, PCNN segmentation and GA improved PCNN segmentation are enhanced in turn. It lays a good foundation for further research on the nondestructive detection of ultrasonic tomography of metal aluminum plates in the experimental environment.
【学位授予单位】：西华大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TG115.28;TP391.41

【参考文献】