基于“S”形试件的切削痕迹检测技术的研究

发布时间：2018-07-05 12:29

本文选题：复杂曲面 + 异常痕迹　；参考：《电子科技大学》2017年硕士论文

【摘要】：随着中国航空航天、车辆船舶制造业的快速发展,复杂自由曲面应用越来越广泛,如机翼、叶轮叶片等。但复杂自由曲面其复杂的几何特征,如变曲率、变扭曲角、开闭角变换等,常常导致加工后的表面结构不均匀,甚至出现异常加工痕迹,难以表征和评价。目前,对复杂自由曲面的表面质量检测评价仍主要集中在轮廓度、粗糙度等指标评价。然而,机加工表面常存在的异常加工痕迹没有确定的可参照标准用以检测评定,常常被划入粗糙度或波纹度进行评价。因此,需要探索一套检测复杂曲面轮廓并对切削痕迹进行提取评价的方法。本文基于“S”形检验试件,搭建适合的测量平台,并重点探讨一种新的提取表征试件型面异常切削痕迹的方法,作为传统粗糙度表征的补充。采用数学“显微镜”——小波分析下的小波多尺度分析方法提取加工表面异常切削痕迹特征;针对小波分析Mallat算法所产生的边界效应进行边界延拓研究对比,分析得出合适于机加工表面的延拓方式;并基于M估计稳健权函数提出小波分析稳健迭代方法以增强滤波稳健性;在评定表征痕迹上,类比支撑长度率提出表面损伤率和损伤面积比概念。最后推导了线性小波分析的算法不确定度传递关系,并基于Matlab GUI完善了异常加工痕迹的评定分析系统。本文具体在以下几方面做出大量的理论研究和对比分析,并得出相关结论:(1)根据“S”形检验试件结构特点,设计搭建了型面轮廓测量平台;通过小波分析提取表面轮廓的异常痕迹特征,并研究了小波基函数的选取以及分解层数的影响因素及正确计算。(2)分析了小波分析边界效应产生原因,研究了各边界延拓方法,并对比得出最适合于机加工表面的延拓方法。引入M估计稳健权函数到小波分析,建立权函数可调参数与表面轮廓数据关系,实现小波稳健滤波分析。(3)提出表面损伤率和损伤面积比概念作为评定指标,有助于痕迹表面直观的定量评价。(4)推导所采用的线性相位小波分析算法的不确定度传递关系,基于Matlab GUI集成开发了表面异常痕迹的检测评定软件系统。
[Abstract]:With the rapid development of aeronautics and spaceflight in China, the application of complex free-form surfaces, such as wings and impeller blades, is becoming more and more extensive. However, the complicated geometric features of complex free-form surface, such as variable curvature, variable twist angle, open and close angle transformation, often lead to uneven surface structure, even abnormal machining trace, which is difficult to characterize and evaluate. At present, the evaluation of surface quality of complex free-form surface is mainly focused on the evaluation of contour and roughness. However, there is no definite reference standard for measuring and evaluating abnormal machining traces on machined surfaces, which are often evaluated by roughness or corrugation. Therefore, it is necessary to explore a method to detect the contour of complex curved surface and to extract and evaluate the cutting trace. Based on the "S" shape test, a suitable measuring platform is set up, and a new method to extract abnormal cutting marks of the surface of the specimen is discussed, which can be used as a supplement to the traditional roughness characterization. The feature of abnormal cutting trace of machined surface is extracted by means of wavelet multi-scale analysis under "microscope" -wavelet analysis, and the boundary effect caused by Mallat algorithm is studied and compared. Based on M estimation robust weight function, wavelet analysis robust iterative method is proposed to enhance filtering robustness. The concepts of surface damage rate and damage area ratio are put forward by analogy of bracing length ratio. Finally, the uncertainty transfer relation of linear wavelet analysis algorithm is deduced, and the evaluation and analysis system of abnormal machining trace is improved based on Matlab GUI. In this paper, a large number of theoretical research and comparative analysis are made in the following aspects, and relevant conclusions are drawn: (1) according to the structural characteristics of the "S" shape test specimen, a profile measuring platform is designed and built; The abnormal trace features of the surface contour are extracted by wavelet analysis, and the selection of wavelet basis function, the factors affecting the number of decomposition layers and the correct calculation are studied. (2) the cause of the boundary effect in wavelet analysis is analyzed, and the methods of boundary continuation are studied. The extension method which is most suitable for machining surface is obtained by comparison. The robust weight function of M estimation is introduced to wavelet analysis, and the relation between the adjustable parameters of weight function and the surface contour data is established, and the wavelet robust filtering analysis is realized. (3) the concepts of surface damage rate and damage area ratio are proposed as evaluation indexes. It is helpful for the quantitative evaluation of trace surface. (4) the uncertainty transfer relation of the linear phase wavelet analysis algorithm is deduced. A software system for detecting and evaluating the surface anomaly trace is developed based on Matlab GUI integration.
【学位授予单位】：电子科技大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TG501

【参考文献】