在眾多流程改善的工具裡面，六標準差一直是被許多業界廣泛接受與利用的工具之一，由於其「數據導向」的特性，不僅限於製造業的品質改善，服務業、醫療業也均有六標準差的應用案例。 本研究以國內面板業G公司為例，進行六標準差應用在產品稽核抽樣流程改善之研究。G公司為國內專精於生產中小尺寸面板的廠商，主要應用於車用面板、醫療面板及工控面板。為了符合車用產品IATF16949:2016的規範，G公司每年進行兩次的產品稽核，目的希望從客戶的角度出發，評估確認所生產的車用型號品質狀況，達到風險控管的目的，並發掘任何持續改進的可能。儘管每年進行兩次的產品稽核，歷年來所完成稽核的產品種類卻不到全年生產的一半，該如何提升產品稽核的落實度，已為G公司所需要考量的重點之一。 為提升G公司產品稽核的落實程度，本研究利用六標準差的DMAIC手法，逐步利用衡量評估樹一一審視造成落實度不佳的可能變數，並且蒐集相關加以分析以確定根因。最後利用標準化及人員教育訓練等方式，達到控制維護成果的目的。 研究結果顯示，透過DMAIC的步驟手法，能夠有效幫助專案小組成員聚焦於改善的方向，鎖定加強可靠度項目的抽樣方式。經過抽樣方式的調整，成功將車載機種覆蓋率提升至100%，稽核覆蓋率提升5%，並且提升了未完成稽核機種的抽樣數量達30%之多，若能持續進行至下半年，預期能夠顯著改善。在本研究的最後，也闡明了未來可能的研究方向，讓後續持續改善的團隊有執行的依據。 ;Six Sigma, widely accepted and utilized in various industries, is a data-driven tool not limited to quality improvement in manufacturing but also applicable in service sectors and healthcare industry. This study examines the application of Six Sigma in enhancing the product audit sampling process at Company G, a domestic manufacturer specializing in small and medium-sized panels primarily used in automotive, medical, and industrial control applications. To comply with the automotive IATF 16949:2016 standards, Company G conducts biannual product audits to evaluate the quality of automotive panels from the customer′s perspective, aiming for risk management and identifying opportunities for continuous improvement. Despite conducting these audits twice a year, the audited product types over the years have covered less than half of the annual production. Enhancing the implementation of product audits has become a key focus for Company G. To enhance the implementation of product audits at G Company, this study utilized the Six Sigma DMAIC methodology. It systematically evaluated potential variables contributing to poor implementation through measurement and assessment, and collected relevant data for root cause analysis. Finally, standardization and personnel training were employed to achieve control and maintenance of outcomes. The research findings demonstrate that the DMAIC approach effectively assists project team members in focusing on improvement directions, particularly in refining sampling methods to strengthen reliability aspects. Following adjustments to the sampling approach, the coverage rate of audit plan toward automotive models produced in first half of 2024 was successfully increased to 100%, with a 5% improvement in audit coverage. Moreover, the sampling quantity for unfinished audit models increased by up to 30%. Continued implementation of these strategies into the latter half of the year is anticipated to result in significant improvements. In conclusion, the study also elucidated potential future research directions to provide a basis for continuous improvement efforts within the team.
