Trial manufacturing of isotropic random surfaces for calibration and uncertainty analysis of surface profiling instruments

湿喷砂法和电火花线切割磨削法适用于制造市场上仿形仪校准用的各向同性随机曲面。应用质量工程技术，以获得最佳的加工条件，均匀的表面纹理，可以成功地产生。另一方面，根据自回归模型，编制了一个计算机程序，生成各种离散地形数据集。介绍了一种通用的数据文件格式（部分修改的SDF），用于测量表面地形数据和计算机计算的模拟数据。建立了将测量不确定度分为与测量原理有关的部分（包括仪器硬件）和计算某些地质统计参数的软件引起的不确定度的方法。在代表性商业仪器所附的软件中，参数定义存在一些差异。本研究项目中的测量不确定度分析仅限于传感装置的动态行为。这意味着没有考虑主要由机械结构引起的不确定性分量。基于光学非接触方法的商业实用表面仿形仪大致可分为点式位移传感器的传统扫描系统和面干涉仪系统。在前一种情况下，在2D线性系统中进行异常值检测和数据校正。对于后一种情况，在三维面系统中采用了类似的数据预处理。利用随机曲面，通过具有代表性的地质统计参数，对这些预处理程序的性能进行了研究。最后，开发了通用的表面形貌分析软件Summit。

Wet-blasting and Wire Electric Discharge Grinding methods are applicable to manufacturing the isotropic random surfaces for calibration of profiling instruments on the market. Quality engineering techniques were applied to obtain the optimized machining conditions by which uniform surface texture can be successfully produced. On the other hand a computer program to generate various sets of discrete topographic data by following the auto-regressive model was accomplished. The common data file format (partly modified SDF) was introduced for both measured surface topographic data and simulation data calculated by a computer. Established was the methodology by which we can divide the measurement uncertainty into the part related to measuring principle including instrument hardware and that caused by software to calculate certain geo-statistical parameters. Some differences in the parameter definitions were disclosed among the software attached to representative commercial instruments. The measurement uncertainty analysis in this research project was limited to the dynamic behavior of sensing devices. It means that the uncertainty component caused mainly by mechanical structure was not taken into account. Commercial and practical surface profiling instruments being based on the optical non-contact method were roughly classified into conventional scanning system by a point-type displacement sensor and areal interferometer system. In the former case, outlier detection and data correction was conducted in the 2D linear system. For the latter case, the similar data pre-processing was utilized in the 3D areal system. The performances of those pre-processing programs were studied through the representative geo-statistical parameters using the manufactured random surfaces. Finally the all-purpose surface topography analyzing software named by "Summit" was completed.