| 摘要: | 目的 – 本研究的目的提出一種全新的評估方法以強化『失效模式與效應分析 (FMEA)』的能力；藉由實務例子的再運用，本研究所提出的評估方法能夠求得各模式相對失效的排序以及求出各失效模式的改善配當量。同時針對各失效模式提出具有效益性的管理意涵。 方法 – 雖然『失效模式與效應分析』能提供快速與簡易方法以鑑別產品或系統的所有失效模式的機率排序，但在『失效模式與效應分析』排序方法中係以風險優先指數(RPN)為準則。此風險優先指數為模式的嚴重度、發生率與偵測度的乘積。此嚴重度為第一個輸入因子亦即模式失效後所衍生的後果程度；第二個輸入因子為發生頻率亦即該模式會發生失效的概率；第三個輸入因子為偵測度亦即模式在失效發生前能夠被偵測出來的程度。『失效模式與效應分析』的主要缺點之一為各失效模式從不同嚴重度、發生率與偵測度的組合乘績，可能會有相同的風險優先指數因而造成決策的困擾。另一個缺點為『失效模式與效應分析』不探討各輸入因子的貢獻參數以至於缺乏因子量化的改善。藉由資料包絡分析法(DEA)與其延伸的模式，本研究從既有的嚴重度、發生率與偵測度即可求得風險排序，而不再由RPN做排序。再者DEA提供各因子的改善尺度。本研究方法藉由既有的實務FMEA的原始資料以DEA的模式展開。第一例為電腦軟體程式開發FMEA，當某一執行指令未完全時，軟體會失效。本例在展現DEA工具手法如何運用在FMEA的完全表達。第二例為醫療體系系統FMEA，當監控系統未展開與執行時，醫療糾紛會發生。本例藉由結合FMEA和DEA的方法提供醫療系統的風險管理。 主要發現 - 經由軟體開發以及醫療系統DEA的再運用，本研究發現所提出的研究方法論，證明DEA不但可以如傳統FMEA的風險排序，而且能強化評估能力，較特別的是能提供FMEA所缺乏的因子量化改善方向。這些量化指標對管理者提供更深層的管理意涵，尤其提供資源再分配與風險管理的參考。 實務應用–本研究方法論讓管理者或設計者在開發一新系統或新產品階段即能夠掌握失效模式。檢討各貢獻因子而不是只看RPN的大小。找出新的因子度量與原始因子度量的差異以決定努力的方向。最後群集分析提供最具經濟效益的投資組合。 價值 – 本研究提出具獨特性、穩健、具體的新方法論以做為失效分析的工具。本方法論克服傳統FMEA的缺失，不但考慮到多重條件與因子權重，並且權衡各因子的貢獻度。之外，本論文提供一些未來值得研究的方向。 Purpose - The purpose of present study is to propose a state-of-the-art new approach to enhance assessment capabilities of failure mode and effects analysis (FMEA); and demonstrates it through practical examples to show how relative failure rankings can be determined and to identify improving scales for failure modes. Subsequently provides helpful managerial implications for management. Methodology - Failure mode and effects analysis (FMEA) offers a quick and easy way for identifying ranking-order for all failure modes in a system or a product. In FMEA the ranking methods is so called risk priority number (RPN), which is a mathematical product of severity (S), occurrence (O), and detection (D). The SODs are input criteria where S is the seriousness of the effect of the failure. The second input is O, which is the probability or frequency of the failure. The third input is D, which is the probability of failure being detected before the impact of the effect is realized. One of major disadvantages of this ranking-order is that the failure mode with different combination of SODs may generate same RPN resulting in difficult decision-making. Another shortfall of FMEA is lacking of discerning contribution factors, which lead to insufficient information about scaling of improving effort. Through data envelopment analysis (DEA) technique and its extension, the proposed approach evolves the current rankings for failure modes by exclusively investigating SOD in lieu of RPN and to furnish with improving scales for SOD. To demonstrate how DEA is implemented to enhance FMEA, the paper examines two real case examples origin in FMEA setting. One is computer software coding where failure could happen during execution if process of coding shall incomplete. The case demonstrates how DEA is applied on existing FMEA. The other case is healthcare discipline where failure could arise if surveillance measures did not fully implement and execute. This case is to provide risk management in healthcare through integration of FMEA and DEA. Findings - By demonstrating an illustrative example and extended case study, the proposed approach supports that DEA can not only complement traditional FMEA for improving assessment capability but also, especially, to provide corrective information regarding the failure factors – severity, occurrence and detection. Further application of DEA extensions also reveals that the utilization of this methodology is useful to managing resource allocation and risk management. Practical implications – It is shown that the proposed approach enables manager/designers to prevent system or product failures at early stage of design. Moreover the approach is able to provide managerial insight of SOD more effectively rather than justifying the efforts on RPN alone. Projection of each SOD is determined to help manager examine scale of efforts. Finally the stratification analysis offers the economical allocations of failure modes with respect to the incurred costs and the efficiency. Originality/value – The paper proposes a state-of-the-art new approach, robust, structured and useful in practice, for failure analysis. The methodology, within a firmed methodology, overcomes some of the largely known shortfalls of traditional FMEA: it takes into account multiple criteria and restricted weighted; and it analyses the failure modes’ ranking considering not only the direct impacts of failure indexes, but also the contribution of these indexes. The paper also provides worthwhile future research directions. |
