Low-coherence interferometry based roughness measurement on turbine blade surfaces using wavelet analysis
Low-coherence interferometry based roughness measurement on turbine blade surfaces using wavelet analysis
复制标题
DOI：
10.1016/j.optlaseng.2016.02.011
复制
发表时间：
2016-07-01
影响因子：
4.6
通讯作者：
Reithmeier, Eduard
中科院分区：
文献类型：
--
作者：
Zou, Yibo;Li, Yinan;Reithmeier, Eduard
介绍了一种用于测量涡轮机叶片表面粗糙度的非接触光学系统--低相干干涉仪（LCI）。该系统不仅具有较高的垂直分辨率，能够获得粗糙形貌，而且与其他常用光学系统相比，具有较大的垂直扫描范围。后一个特性使我们能够测量具有大曲率的涡轮机叶片表面，而不会在透镜和测量对象之间发生碰撞。在获得表面形貌后，应用小波分析将原始表面分解成多个带宽进行多尺度分析。结果表明，开发的LCI系统证明了良好的性能，不仅在获得表面形貌的粗糙度尺度，但也能够测量具有复杂的几何形状的物体在一个大的垂直范围内的表面。此外，本研究中所应用的双正交小波在提取整个表面的粗糙度微结构方面表现出良好的振幅和相位特性。最后，传统的粗糙度参数，如平均表面粗糙度S-a和均方根（RMS）粗糙度S-q，在每个分解子带的评估和它们与每个子带的尺度的相关性进行了分析。(C)2016爱思唯尔有限公司版权所有
In this paper, a non-contact optical system, a low-coherence interferometer (LCI), is introduced for the purpose of measuring the surface roughness of turbine blades. The designed system not only possesses a high vertical resolution and is able to acquire the roughness topography, but also it has a large vertical scanning range compared to other commonly used optical systems. The latter characteristic allows us to measure turbine blades surfaces with large curvature without collisions between the lens and the measurement object. After obtaining the surface topography, wavelet analysis is applied to decompose the original surface into multiple bandwidths to conduct a multiscale analysis. The results show that the developed LCI system proofs a good performance not only in obtaining the surface topography in the roughness scale but also in being able to measure surfaces of objects that possess a complex geometry in a large vertical range. Furthermore, the applied biorthogonal wavelet in this study has performed good amplitude and phase properties in extracting the roughness microstructures from the whole surface. Finally, the traditional roughness parameters, such as the mean surface roughness S-a and the Root Mean Square (RMS) roughness S-q, are evaluated in each decomposed subband and their correlations with the scale of each subband are analyzed. (C) 2016 Elsevier Ltd. All rights reserved.