以鬆公差元件組成精密組件之配置方法研究

Abstract

公差技術是工業發展的重要基礎，並且為產品品質與成本的指標，隨著精密機械對於精度的要求提高， 公差技術扮演的角色更形重要， 然而組件的精度變異（ 或誤差範圍） 受其組成元件公差累積所影響， 對於要求高精度的組件其組成元件需要更緊的公差， 也表示該組件的生產成本居高不下， 對於組成元件較多的系統， 則其組成的零件公差要求將需要極高的生產價格， 故亟需有系統地探討能讓元件在不增加生產成本的條件下順利組裝， 甚至於更進一步能將組件的精度變異控制在更小的範圍內， 亦即以較鬆元件的公差來組成精度較高的組件。本計畫延續計畫申請人前期在公差設計分析技術發展之成果， 探討一般要求精度的組件如何在不壓縮元件公差的原則下使用適當的方法來降低組件的精度變異量， 本計畫將有系統地先由探討單純的雙元件組裝之線性尺度變異開始， 透過理論探討如何以較鬆公差的元件來組成精度變異較低的組件， 再逐步擴展到多元件之高精度組裝方法， 並將所發展之方法結合軟體工程技術建立軟體模組， 做為分析模擬與公差配置之軟體工具。計畫將分三年逐步探討並建立以鬆公差的元件來組成精度變異較低組件的方法， 同時建構分析軟體， 計畫第一年重點在於推論以寬鬆元件公差組成較高精度之組件之可能方法， 以及發展以最惡狀況法分析一般分組隨意組配之組裝件尺度變異， 第二年重點則在於發展以統計法分析一般分組隨意組配之組裝件尺度變異， 以及推導此方法之元件公差配置及其限制， 統計方法的對象將包括均勻分佈與常態分佈， 並結合均值偏移效果之統計分析， 同時發展軟體系統使用。第三年則以探討負公差修配之組裝方法理論， 以及利用此方法之元件公差配置及其限制之推導， 同時建構分析軟體， 作為組裝精度工程分析工具。

Tolerancing technology is one of the foundations of industry technologies. Itis also an index of product quality and cost. The role of tolerancing becomes moreimportant as precision requirements of a product are getting crucial. The variationof precision (or the repeatability) of an assembly is effected by the tolerancingspecifications of its components. This implies component tolerances of a highprecision product must be tighter, which means higher production cost. It is acritical issue to develop methods for a precise assembly by components with loosetolerances. This three-year research project extends PI's previous achievements intolerancing engineering to develop methods for reducing precision variation of anassembly without tight tolerances for components. Theoretical analysis on how toachieve higher precision without tightening tolerances of components will be firstinvestigated. Software prototype will be then developed as an engineeringapplication. In the first year, the research will focus on investigating possiblemethods to achieve tight repeatability by loosely toleranced components.Worst-case tolerance analysis for sub-group method will be also developed.Statistical methods for sub-grouped random assembly will be developed foruniform and normal distributions. Analytical method for statistical distributionwith mean shift will be also investigated. Theories and methods for conditionedassembly with “negative tolerancing” will be investigated in the third year.Tolerance distribution to components and limitations of the developed methodswill be investigated as well. Software for the developed methods will be developedat the time as an engineering tool for tolerance distribution of an assembly bycomponents with loose tolerances.