中文摘要
機械器具在設計時，往往知其總和尺寸和裝配組合總公差，對於個別零件尺寸公差的訂立，憑藉的往往是設計者的個人經驗與直覺。而傳統的最差狀況法與統計學方法，也都沒考慮到各個零件的製造方式與主要零件尺寸的限制，在實際應用上並不十分適當。本研究以實驗規劃法中的反應曲面法檢討影響尺寸公差分配的因素，考慮加工方法及使用上功能的限定，規劃實驗中的各個公差尺寸水準，而求得零件組合的公差迴歸函數，並經由不同的實驗數據來修正此函數的準確性。此模式可以加入多重拘束條件，如成本要求及功能需求等等，並以量測誤差值最小化作目標方程式，求出裝配公差配置的最佳化組合，提供設計者一種簡單而有效的公差分配模式。
ABSTRACT
It is not unusual in the process of product development,that one knows only the assembly dimension and its tolerance instead of the individual parts.Assignments of the tolerance for individual parts are very often depend on the experience and intuition of the design engineer. Many design engineer sdopt,Worst Method and Statistic Method in specifying tolerances for the parts. However,these two methods do not take into account the special dimensional requirement sometime has to be applied to the individual part. And manufacture requirements and cost are not an issue for these methods. This is not practicle for most application. In this study,we adopt the Reaction Surface Method to plan the experiment,factor affect the distribution of individual parts tolerances are considered, especially way of processing parts and its functional constraint are integrated in the prrocess of regression. Data used in the regression process are obtained through actual measurement,which in turn will increase the accuracy of the regression function. The Method adopted in this research can be modified by the actual constrain,like cost cost or performance requirements. Minimum measure