台灣的半導體產業發展至今已有近四十年的歷史，製程與產業結構鏈已經相當的完整與成熟。由於半導體產業鏈非常的複雜，並且環環相扣，若在過程中任何一個步驟出現錯誤，將對整體造成生產週期增加與產線停滯等嚴重後果。隨著預診斷及運行狀況管理(Prognostics and Health Management)的應用逐漸普及，本研究也沿用此方法的概念來進行改善。 本研究針對半導體產業鏈的關鍵問題找出預防及改善的方法，應用製程失效模式與影響分析(Process Failure Mode and Effect Analysis, PFMEA)分析潛在失效模式，透過風險優先數值(Risk Priority Number, RPN)找出高風險的問題，並結合故障樹分析(Fault tree analysis, FTA)，區分出造成影響的嚴重性，在還沒發生故障前，擬定適當的製程改善對策及措施，降低可能造成的風險並提升製程、設備之可靠性。 ;Taiwan′s semiconductor industry has developed for about 40 years. The process and industrial structure chain has been quite complete and mature. Because the semiconductor process is very complicated and interlocking, if some mistakes occur in any step of the process, the overall production cycle will increase and the production line will be stagnant. As the application of Prognostics and Health Management has become popular. This study has also used the concept of this technology to improve the problem. This study identifies prevention and improvement methods for key issues in semiconductor processes. Applies process failure mode and effect analysis (PFMEA) to analyze potential failure modes and find out the high-risk problem through Risk Priority Number (RPN), and combined with the Fault Tree Analysis (FTA). Distinguish the severity of the impact, and formulate appropriate process improvement strategy and measures before the failure has occurred. Reduce the possible risks and improve the reliability of the process and equipment.
