- 學位論文
利用公差分析進行設計及實務檢討
Using tolerance analysis for design and instance review
摘要
產品經由3D圖面設計後，所有零件的組配結果，通常為餘隙配合，不然就是干涉配合的狀況。設計過程中，大多只能靠設計者的經驗及實務來大致上抓出配合間隙及公差範圍，在生產或是製造出成品後，再由組裝結果來反推是否設計得宜，往往沒有科學的方式來進行分析。 本文即針對伺服器在組立零件後，以均平方根RSS (Root-Sum-Square)統計法方式進行公差分析；並且配合製程能力指數Cpk (Process Capability Index) 和6個標準差(Six Sigma)，分析單一零件的公差範圍是否定義得宜並得到一定的生產良率。並利用全尺寸檢查(Full-Article-Inspection)報告，來反證供應商生產的零件也能夠符合設計要求及製程能力。 由結果得知，RSS統計會比Worse Case統計法更適合於公差分析，並且驗證單一零件的設計公差是否定義得宜，並且在生產公差上也得到能否符合Cpk水準的要求。未來可將此方法運用於其他機構設計問題的公差分析，減少在設計過程或製作樣品過程，找出問題點進而爭取開發時間與降低成本成本，後續也期望運用在更寬廣的層面上。
並列摘要
After the product was design by 3D drawings, the assemble result for all of parts were usually clearance or interference. During the design process, most of the situations that only rely on designer’s experience to define the gap and tolerance roughly. After making or manufacture samples, via assembly result to infer the design is proper or not. There is often no scientific way to analyze. The article is focus on server structures that assemble of the parts, using RSS method to tolerance analyze, and cooperate with Cpk and 6-sigma index, analyze the single part of tolerance is suitable or not and get the great produce quality. Finally, it refers Full-Article-Inspection (F.A.I) report to improve the part produced by vendor is ability to meet design requirement and process capabilities.