闭式叶轮荧光渗透检测缺陷尺寸双目测量

发布时间：2018-06-08 07:39

本文选题：闭式叶轮内表面缺陷 + 荧光渗透检测　；参考：《南昌航空大学》2017年硕士论文

【摘要】：某型号航空发动机闭式叶轮,采用精密铸造成型,由于叶轮是高温转动部件,因此其内部缺陷控制和表面缺陷控制关系到航空发动机的可靠性和使用寿命。叶轮由高温合金材料铸造而成,表面缺陷检测最佳方式为高灵敏度荧光渗透后乳化检测,由于叶轮结构复杂,存在许多接近完全闭合的空间,闭合空间表面的荧光渗透检测存在缺陷难以观察、缺陷尺寸测量不易的难题。本文尝试采用双目测量方法对叶轮内表面荧光渗透缺陷进行测量。本文针对叶轮封闭空间的内表面缺陷荧光渗透检测,搭建以紫外光发光二极管、H5内窥镜镜头(直径为5.5mm,镜头长度为31mm,图像分辨率为640*480)为基础的检测环境,自编图像采集软件,可以获得单帧和连续图像,利用该环境采集荧光渗透检测灵敏度测试片图像,该图像在紫外光下显示,光强度分布不均匀且对比度低,在图像光场均匀化、图像降噪、图像分割、边缘提取后,获得灵敏度试片上缺陷的标准图像,并对缺陷尺寸进行测量;搭建了双目测量平台,将渗透的标准试片置于密闭空间中,模拟叶轮闭合环境。对双目测量平台进行张正友标定,在比较归一化互相关算法(NCC算法)与SIFT算法的基础上,采用NCC算法,引入图像匹配的约束条件,提高了匹配效率与匹配精度。实现标准试片上的缺陷尺寸精确测量,随后利用该平台对叶轮内表面实际缺陷进行测量,缺陷尺寸测量相对误差小于5.5%。本文搭建的内窥镜双目测量平台,有效解决叶轮内表面缺陷荧光渗透图像获取难题,所用的光场均匀化、图像降噪、图像分割、边缘提取方法有效、研究获得的标定算法方法简单、精度高,能够很好应用于航空发动机闭式叶轮内表面缺陷检测中。
[Abstract]:The closed impeller of a certain type of aero-engine is formed by precision casting. Because the impeller is a rotating component at high temperature, its internal defect control and surface defect control are related to the reliability and service life of the aero-engine. The impeller is cast from superalloy material. The best way to detect the surface defect is high sensitivity fluorescence permeation emulsification detection. Because of the complex structure of impeller, there are many spaces close to complete closure. It is difficult to observe the defect and measure the size of the defect in the fluorescence permeation detection of closed space surface. In this paper, a binocular method is used to measure the defect of the inner surface of impeller. In this paper, an image acquisition software based on UV light-emitting diode H5 endoscope lens (diameter 5.5mm, lens length 31mm, image resolution 640480) is set up to detect the inner surface defect of impeller. A single frame and a continuous image can be obtained, and the fluorescence osmosis detection sensitivity test image can be collected in this environment. The image is displayed under ultraviolet light, the intensity distribution is uneven and the contrast is low, the image is uniformly distributed in the image light field, and the image is de-noised. After image segmentation and edge extraction, the standard image of the defect on the sensitivity test piece is obtained, and the defect size is measured. A binocular measuring platform is set up to simulate the impeller closure environment by placing the permeable standard test piece in a closed space. The binocular measurement platform was calibrated by Zhang Zhengyou. On the basis of comparing the normalized cross-correlation algorithm (NCC) and sift algorithm, the NCC algorithm was used to introduce the constraint condition of image matching, and the matching efficiency and accuracy were improved. The accurate measurement of the defect size on the standard test sheet is realized, and then the actual defects on the impeller inner surface are measured by the platform, and the relative error of the defect size measurement is less than 5.5. The binocular measurement platform built in this paper can effectively solve the problem of obtaining fluorescence permeable image of impeller inner surface defects. The light field homogenization, image noise reduction, image segmentation and edge extraction are effective. The proposed calibration algorithm is simple and accurate, and can be applied to the detection of inner surface defects of closed impeller of aero-engine.
【学位授予单位】：南昌航空大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：V263.6;TP391.41

【参考文献】