Vol.2 No.2 2009

21/98

Research paper : High accuracy three-dimensional shape measurements for supporting manufacturing industries (S. Osawa et al.)−101−Synthesiology - English edition Vol.2 No.2 (2009) own the CMMs requested official certification for their capability to conduct three-dimensional measurements traceable to the national standard. Therefore, similar to the aforementioned step gauge standards, we established a technological committee for officially accrediting the service providers for three-dimensional measurement in the National Institute of Technology and Evaluation (NITE), and drafted the Guidelines for Technological Requirements[9] for CMM calibration. Currently several companies received the official accreditation for CMMs and are providing calibration services.Since CMM is a multifunctional device, it is impossible to sufficiently evaluate all of its functions just by a few measurements with step gauges. The challenge is to conduct appropriate assessment with as little procedures as possible. The same problem arises in the performance testing for the buyer and the seller of CMM. In ISO, the testing method at the time of delivery of CMM is standardized. AIST has participated in the ISO meetings as an expert to work on standardization. In determining the Guidelines for Technological Requirements for the calibration of CMM, we decided to use the ISO standards[10]. This means that the metrological standard (metrological traceability system) references the industrial standard (standardization and rules for products and services). On the other hand, this ISO standard rules that a standard traceable to national standard must be used, and in reverse, the industrial standard references the metrological standard. We have been conducting R&D under the thinking that effective application of the system is possible for use in industry through unification of metrological and industrial standards. This time, one of our efforts was rewarded.4.2 Calculation of uncertainty in three-dimensional measurementCMM conducts point measurements of each point and gathers the point data. If the measurement is for a circle, least-square fitting is done to calculate the diameter, circularity, and central coordinates. Since there are measurement errors in each measurement point, it is difficult to see at which uncertainty the diameter and circularity were finally calculated. Following items can be listed as factors of uncertainty in three-dimensional measurement (see Fig. 11).1. Uncertainty of probing2. Uncertainty arising from geometrical error3. Uncertainty of data processing (least-square method, etc.)4. Uncertainty arising from measurement procedure (number of measurement points and their arrangement, etc.)5. Uncertainty arising from environmental change (temperature, humidity, etc.)6. Uncertainty arising from positioning of workpiece (holding strength, distortion from own weight, etc.)7. Uncertainty arising from workpiece itself (surface roughness, shape error, etc.)Since various factors of uncertainty affect the measurement, it is fairly complicated to assess the final uncertainty. At AIST, with a grant from the International Joint Research Program, New Energy and Industrial Technology Development Organization (the NEDO grant), research on measurement uncertainty in CMM using Monte Carlo simulation was conducted jointly with the Physikalisch-Technische Bundesanstalt (PTB) of Germany, the National Measurement Institute of Australia (NMIA), the University of Tokyo, and others. The method is called virtual CMM[11], and its basic concept was developed by the PTB. Figure 12 shows the outline of the virtual CMM. Virtual CMM is a computer model of CMM including the errors of factors of uncertainly such as geometrical errors. By conducting about 200 measurements on the virtual model on the computer using the measured position information and other data obtained from actual measurements, the standard deviation and the uncertainty of the measured value obtained from virtual measurements are calculated. Fig. 10 Result of international comparison for ball plate (CCL-K6).The graph shows the difference between AIST values and reference values for 25 spheres of the ball plate. The error bar shows the uncertainty of AIST value, while the blue line shows the uncertainty of reference value.10CCL-K6[mm] [μm] 25 Deviation from reference value (µm)Measured length (mm)-0.5-0.2-0.3-0.4-0.10.00.10.20.30.4501001502002503003504004505002,673,118,12134,169,1714,185,2110,2215,231920,24251Fig. 11 Factors of measurement uncertainty for CMM.Distortion by clampEnvironmental change factors such as temperatureUncertainty of measured valueExecution of measurementData processing such as least-square methodProbe arrangementProbing errorSurface roughness and shape error of measured objectMeasurement procedurezxyTrfrΔ

※このページを正しく表示するにはFlashPlayer9以上が必要です