以鬆公差元件組成精密組件之配置方法研究(II)

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，隨著精密機械與3C 產業對於精度的要求提高，公差技術扮演的角色更形重要。然而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，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，計畫將有由單純的雙元件組裝之線性尺度變異，透過理論分析系統地探討如何以較鬆公差的元件來組成精度變異較低的組件，並逐步擴展到多元件之高精度組裝方法，同時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。本計畫於99 年先由組裝結果推論以寬鬆元件公差組成較高精度之組件之可能方法，以求計畫之完整性，同時亦發展以最惡狀況法分析一般分群隨意組配之組裝件尺度變異，目前成果顯示在不重新產製元件避免增加生產成本的條件下，分群組裝可有效降低組裝結果之尺度變異量，對於大量生產之元件組裝匹配有極大效益。100 年則主要發展一般分群隨意組裝之統計公差分析組裝件尺度變異，推導利用此方法之元件公差配置及其限制，其中要探討的統計分佈包含常見之均勻分佈與常態分佈兩種，並將結合均值偏移效果之統計分析。101 年則探討負公差組裝方法理論及其應用技術，推導利用此方法之元件公差配置及其限制，計畫並於100 年起建構公差分析軟體，作為組裝精度工程分析之工具。

Tolerancing technology is one of the foundat ions of indust ry technologies. It is alsoan index of product qual i ty and cost . The role of tolerancing becomes more important asprecision requi rements of a product are get t ing crucial . The variat ion of precision (or therepeatabi l i ty) of an assembly is effected by tolerancing speci ficat ions of i ts components.This impl ies component tolerances of a high precision product must be t ighter, whichmeans higher product ion cost . It is a cri t ical issue to develop methods for a preciseassembly by components wi th loose tolerances. This three-year research project is aimedto develop methods for reducing precision variat ion of an assembly wi thout t ighttolerances for components. Theoret ical analysis on how to achieve higher precisionwi thout t ightening tolerances of components wi l l be first invest igated. Software prototypewi l l be then developed as an engineering appl icat ion.In the past half year, we have invest igated possible methods to achieve t ightrepeatabi l i ty by loosely toleranced components. Worst-case tolerance analysis forgrouping- then-random-assembly method is also developed. The resul t showed thatassembly precision can be improved wi th this method. Product ion cost thus can be highlyreduced wi th lose tolerancing that is very cost -effect ive for mass product ion. Stat ist icalmethods for grouping-then- random-assembly method wi l l be developed for uni form andnormal dist ribut ions in the fol lowing year. Analyt ical method for stat ist ical distr ibut ionwi th mean shift wi l l be also invest igated. Theories and methods for condi t ioned assemblywi th “negat ive tolerancing” wi l l be invest igated in the third year. Tolerance dist ribut ion tocomponents and l imi tat ions of the developed methods wi l l be invest igated as wel l .Software for the developed methods wi l l be developed at the t ime as an engineering toolfor tolerance distribut ion of an assembly by components wi th loose tolerances.