目前台灣產業已由OEM走入ODM的階段,製造仍是今天市場上的競爭武器;然而,當工程師設計出來的「樣品」沒有考慮易製性,則將會導致樣品無法變成「產品」,且若製造成本太高、品質太差或交期過長,則更無法成為「商品」。因此,為了強化企業的製造競爭優勢,本研究欲提出品質相關問題之解決手法,透過資訊化工具、量化分析和團隊努力的概念,期望在早期產品設計階段解決製造問題。 本研究主要針對製造業提出適當的品質診斷系統,且以印刷電路板業為例進行探討。運用自組織映射網路,對設計/製程之「品質異常楚理單」做製程品質缺點的失效原因屬性分類,歸納出製程品質問題中相關屬性之異常原因建構失效模式與效應分析,提出有效的矯正預防對策,進而提升製造品質。研究結果資料挖掘技術較一般品質特性要因手法,所歸納出失效模式異常真因準確度高出34.69%。驗證本研究所提出之架構,對設計工程人員在設計產品製造流程時,可作為有效的運用工具。
Nowadays the type of industry in Taiwan has transferred from ODM to OEM, but manufacturing is still an efficient instrument in the market. However, if the product engineers design the samples without considering the “Reducibility in,” cost, or quality, the products can’t be turned into the commodities in the market. Therefore, the purpose of this research is to establish the effective device of solving the problems of quality by the information technology, quantification analysis and teamwork. We explored the various quality diagnostic systems of manufacturing, and focused on the industry of printed circuit board. Using data mining technology generalized the abnormal causes of process quality issues, and established the quality diagnostic system of failure modes and effects analysis (FMEA) to improve the quality. Furthermore, we used the self-organizing maps (SOM) to categorize the process quality drawbacks of design/process quality abnormal process order, and analyzed the process failure modes. The research shows that using data mining technology to generalize to abnormal causes is accurate than using general cause-and-effect diagram by 34.7%. As a result, the system being verified in this research is an effective tool for the R&D engineers to design the proofread process flows.