| dc.description.abstract | 本研究為探討在影響直通率的各項不良項目，透過六標準差專案分析關鍵不良因子，利用定義 (Define)、量測 (Measure)、分析 (Analyze)、改善 (Improve)、控制 (Control)五個系列階段分析影響直通率的關鍵因子後進行一系列的改善方案，計算直通率改善前後的差異並將流程標準化，以提升總體直通率。
首先定義階段利用柏拉圖的80/20法則將不良項目排序，將影響80%總不良數中20%的不良項目利用魚骨圖分析不良因素，接著利用力場圖將每項不良項目的不良因素做為阻力與改善方案作為助力，助力總分代表該不良項目改善的可行性，最後將不良項目的柏拉圖、魚骨圖、力場圖對應分數彙整成綜合評分表並排列，同樣利用80/20法則綜合分數占80%的項目作為後續階段快速改善的主要不良項目。量測階段首先繪製LCD貼合流程圖分析出可計量的動作，接著利用量測系統分析 （Measurement Systems Analysis, MSA）確保量測人員對於不良分析是否判斷一致，最後使用層級分析法 (Analytic Hierarchy Process, AHP)和因果矩陣圖 (Cause-And-Effect Matrix)計算出可計量的關鍵動作因子。分析階段確認關鍵動作因子對不良項目的影響力，並利用失效模式分析 (Failure Mode And Effects Analysis,FMEA)算出風險順序數 (Risk Priority Number ,RPN)，RPN高於150的製程流程提出對應的對策預防，並依照RPN數值高低決定改善的先後順序，數值越高越先改善。改善階段採用實驗設計 (Design Of Experiment, DOE)將分析階段RPN值大於150分的關鍵動作因子擬定改善策略並決定實驗樣本大小以及最佳依變數，將改善後的RPN值重新評估製程失效模式效應分析 (Process Failure Mode and Effects Analysis ,PEMA)確認RPN值是否有降低趨勢，最後利用常態分佈比較不良項目改善前後達成不良率控制目標機率是否有提高。控制階段經過改善階段取得最佳量測參數，並且得知改善效果為有效的之後，在控制階段會針對分析、改善階段提到的製程流程製作控管計畫並加以實行作業標準化。 | zh_TW |
| dc.description.abstract | This reasurch is to study the all kinds of NG items wich effect the field rate , and by using the six sigma program to anaylsis NG factors .the tool including Define , Measure , Analyze , Improve , control phase . In these five phase will analyze the key factor which affect the field rate and a series of improvement plans are implement.Calculate the difference between before and after of the improvement and standardize the process to improve the overall field rate.
In Define phase using Plato chart 80/20 rule to ranked the fail items and 20% of the fail items that affect 80% of the total fail rate will be analyzed by fishbone diagrams. Then use the Force Field Analysis to use the fail factors of each fail item as restraining force and improve plan as a driving force, and calculate the total score to represent the feasibility of the improvement. Finally, Summarize the Plato, fishbone diagram, and Force Field Analysis turn into corresponding scores and use the same 80/20 rule method with a comprehensive score of 80% as the main fail itmes for rapid improvement in the subsequent stages. In measurement phase , first draws the LCD bonding flowchart to analyze the measurable actions, and then uses Measurement Systems Analysis (MSA) to ensure that the inspector have consistent judgments on the failure analysis, and finally use the Analytic Hierarchy Process (AHP) and Cause-and-Effect Matrix to calculate quantifiable key action factors.In Analyze phase , confirms the influence of key action factors and uses Failure Mode and Effects Analysis (FMEA) to calculates RPN, and proposes corresponding prevention stragety for the manufacture process if RPN higher than 150. determines the order of improvement according to the RPN value. In improvement phase, Design Of Experiment (DOE) is adopted to analyze those improvement strategy which RPN value is greater than 150 points, determine the experimental sample size and the best dependent variable.Then re-evaluate the improved RPN value of the Process Failure Mode And Effects Analysis (PEMA) ,confirms whether the RPN value has a decreasing trend. Finally uses the normal distribution to compare whether the before and after the improvement probability of reaching the failure rate control has increased.Last one, Contol Phase ,after obtaining the best measurement parameters in the improvement phase, and knowing that the improvement effect is effective, in the control phase will made a manufacture control plan and this plan will be implemented and make the Standard Operation Procedure . | en_US |