在離散量測方面,圓形零件的幾何特徵測量是依據一群離散的資料點,取樣的不確定性會導致與真圓度的誤差;樣本數量是決定量測的不確定的關鍵因子。樣本數量的增加能提供完全的資訊和改進評估值的準確性,然而也會導致較高的費用。在最小區域(Minimum zone)標準下只需要四個在關鍵區域內的點即可決定真圓度,但樣本數量的增加亦會增加多餘樣本點數量。 本文以樣本點模擬輪廓零件外形,進而分析兩種策略下真圓度之準確性。在本文裡提供一套數學模型,用以建立樣本與評估圓準確性之間的關係。這個模型提供一個有效率的工具;能降低在量測過程的費用,亦能提供在決定適當的樣本量同時亦給予評估圓較高準確性。
For discrete measurement, circularity is calculated upon a set of discrete data points, which represent a sample of the measured part. The sampling uncertainty causes deviation from the true circularity. The sample size is the critical factor to determine the uncertainty of the measurement. The increment of sample size offers more comprehensive information of the measured part and improves the accuracy of the evaluated value, however, that causes a higher cost in the measurement. The increment of sample size also increases the number of redundant point since circularity under minimum zone criterion is determined by four critical points only. A mathematical model created in this research is to establish the relationship between the sample size and the accuracy of the evaluated circularity. This model offers an effective tool to determine the proper sample size giving the required accuracy of the evaluated circularity with less cost in the measurement. Finally, a simulation of circularity measurement based on two different sampling strategies is given in this research to verify the effectiveness of the proposed model.