本研究旨在優化網通產品之 PCBA 螺絲鎖附製程，傳統製程往往面臨結構應力品 質與生產工時兩者互相拉扯之兩難，且常依賴工程師經驗，缺乏系統性依據。為此，本研究提出一套結合田口實驗設計與多目標權重決策之優化架構，透過L9 (34) 直交表進行製程因子（轉速、鎖固順序、鎖固策略）之篩選，並引入「品質優先」、「平衡策略」與「效率優先」三種不同決策權重模型。研究結果顯示，製程參數間存在強烈之非線性交互作用。均值主效應圖雖指出整體大趨勢以「轉速600 rpm、由內向外鎖固、6 KG 直接鎖固」為最優；然而經由交互作用圖深度橫向比對發現，當產線固定採用跨情境最具強健特徵之「由內向外鎖固」以釋放翹曲應力時，中度轉速（400 rpm）在品質與平衡模型中之綜合績效已與高轉速（600 rpm）旗鼓相當、甚至反超。本研究打破傳統單一固定最佳解之盲點，提煉出因應產線UPH與品質KPI之動態調整決策指南，不僅突破多目標限制，更實質提升產線製程管理之便利性與標準化程度，具備靈活之工業應用價值。;This study aims to optimize the PCBA screw fastening process for networking products， addressing the inherent trade-off between structural stress quality and production cycle time. Conventional processes often rely heavily on empirical experience， lacking a systematic approach. To resolve this， we propose an optimization framework integrating the Taguchi method with multi-objective decision-making. Utilizing an L9(34) orthogonal array， we evaluated process factors—rotation speed， fastening sequence， and fastening strategy—under three distinct weighting models: "Quality-First，" "Balanced，" and "Efficiency-First." Experimental results reveal significant non-linear interaction effects among process parameters. While the Main Effects Plot suggests a global optimum of "600 rpm， inside-out sequence， and 6 kg direct fastening，" cross-scenario interaction analysis indicates that when the robust "inside-out sequence" is mandated to mitigate warpage stress， a moderate speed of 400 rpm achieves comprehensive performance comparable to， or even exceeding， that of 600 rpm in both Quality-First and Balanced models. By moving beyond static optimal solutions， this study provides a dynamic decision-making guide aligned with UPH and quality KPIs. This approach overcomes multi-objective constraints， enhances shop-floor standardization， and offers significant industrial flexibility.
