創新可靠度配當方法

Abstract

可靠度的配當結果直接影響到產品的可靠度與競爭力，目前目標可行性法 (Feasibility-of-objectives) 常被使用於軍用機械-電子系統可靠度設計上，評點可靠度配當法 (Average weighting allocation) 也被廣泛的使用於商用產品可靠度設計上；然而現行的可靠度配當方法有六個主要缺點：(1) 未考慮權重，(2) 衡量尺度範圍違反統計假設，(3) 未考慮從系統安全性及合理的風險評估角度進行可靠度配當，(4) 配當時未考慮因子間之間接影響關係，(5) 配當過程中未同時考慮系統故障率預估值。(6) 在設計或修改一個複雜的系統時，未同時考慮技術能力或經費限制條件常會影響最大可靠度目標值之達成。因此，為改進傳統可靠度配當方法的缺點，並考量在可靠度設計初期，依據各項資訊可取得之情況下，本研究提出三種可靠度配當方法：如在設計初期缺乏量化數據情況下，建議使用最大熵順序加權平均法 (Maximal entropy ordered weighted averaging)；如可獲得部分系統參考量化數據且考量成本因素情況下，建議運用情境化最大熵順序加權平均運算法(Situational maximal entropy ordered weighted averaging，SME-OWA)，進行以成本為基礎的可靠度配當；如已具備相當之使用參考數據且以安全性及風險性為設計上之優先考量，建議以FMEA為基礎結合決策實驗室分析法進行可靠度配當；本研究並以先進戰機之雷達系統、起落架系統及引擎資料獲得單元三個例子，進一步將所提出之三種方法與傳統可靠度配當方法進行比較。結果證明使用本研究所提之方法可獲得正確的可靠度配當數值，可提供設計者與管理階層製訂彈性之決策參考資訊。

Reliability allocation is one of the most important factors to consider when determining the reliability and competitiveness of a product. The feasibility-of-objectives (FOO) technique has become the current standard for assessing reliability designs for military mechanical-electrical systems, whereas the average weighting allocation method is widely used for commercial applications. However, the conventional reliability allocation methods have six fundamental problems. The first problem is the measurement scale, while the second problem is that the system allocation factors are unequal weighted to one another, the third problem is most reliability allocations methods often neglect many important features, such as maintainability and risk issues. The forth problem is that they do not consider indirect relations between subsystems or components; the fifth problem is they do not consider predicted failure rate in the apportionment process and finally the sixth problem is in designing or modifying a complex system, where the technological or financial constraints might dictate the maximum achievable reliability. To address these issues, this study proposes three reliability allocation methods. Depending upon the operational environment and data required on design process, if data not available, the maximal entropy ordered weighted averaging (ME-OWA) method is suggested, which efficiently resolves the shortcomings of the FOO technique and the average weighting allocation method. If cost is highest concern and data is available, such as predicted failure rate, predicted unit cost etc., the cost-based situational maximal entropy ordered weighted averaging (SME-OWA) approach is suggested that associates the required cost, to a particular technology level, of a predicted failure rate of subsystems. If the risk issue is highest concern and we can get FMEA data before the design process, the combined failure mode and effects analysis (FMEA) and decision-making trial and evaluation laboratory (DEMATEL) technique is suggested, which considers severity of failure, occurrence of a failure, and detection of a cause of failure in reliability design. This study evaluates reliability allocation in the context of a fighter aircraft airborne radar system, landing gear system and engine data acquisition unit (EDAU). The results from these three comparisons show that the proposed methods are both accurate and flexible for actual commercial application.