4.4.1 方法之分類 (categories of methods)

hlperng

章節分配表
4.4 可恃性分析 (Dependability analysis)
4.4.1 方法分類 (Categories of methods)
Dependability analysis methods, which are explained briefly in Annex A, can be classified by the following categories with regard to their main purpose:  
可恃性分析方法，摘要說明如附錄 A，可以依照其主要目的分類如下：

a) methods for fault avoidance, e.g.,
    1) parts derating and selection,
    2) stress-strength analysis;
a) 避免故障方法，例如：
    1) 零件減額定與選擇
    2) 應力-強度分析

b) methods for architectural analysis and dependability assessment (allocation), e.g.,
    1) bottom-up method (mainly dealing with effects of single faults),
        - event tree analysis (ETA),
        - failure mode and effects analysis (FMEA),
        - hazard and operability study (HAZOP);
    2) top-down methods (able to account for effects arising from combination of faults)
        - fault tree analysis (FTA),
        - Markov analysis,
        - Petri net analysis,
        - true table (structure function analysis),
        - reliability block diagrams (RBD),
b) 架構分析與可恃性評鑑（配當）方法，例如：
    1) 由下而上法（主要討論單一故障的效應），
         - 事件樹 (ETA)
         - 失效模式與效應分析 (FMEA)，
         - 危害與操作性研究 (HAZOP)；    2) 由上而下法（可考量故障組合引起的效應）
         - 故障樹分析 (FTA)
         - 馬可夫分析
         - 派翠網分析
         - 真值表（結構功能分析）
         - 可靠度方塊圖


c) methods for estimation of measure for basic events, e.g.,
    - failure rate prediction,
    - human reliability analysis (HRA),
    - statistical reliability methods,
    - software reliability engineering (SRE).
c) 基本事件度量推定法
     - 失效率預估
     - 人員可靠度分析 (HRA)
     - 統計可靠度法
     - 軟體可靠度工程 (SRE)


Another distinction is whether these methods work with sequences of events or time-dependent properties.  If this is taken into account, the following comprehensive categorization results:  
另外一種明顯的分類方式為考量這些方法是否討論事件的後果或時間相關性值。當考量此一因素時，其完整的分類結果如下表所示：

sequence dependent	Event-tree analysis	Markov, Petri, Truth table
sequence independent	FMEA, HAZOP	FTA, RBD
	Bottom-up (single fault)	Top-down (multiple faults)

與次序有關	事件樹分析	馬可夫、派翠網、真值表
與次序無關	FMEA、HAZOP	FTA、RBD
	由下而上（單一故障）	由上而下（多重故障）

These analysis methods allow for the evaluation of qualitative characteristics as well as estimation of quantitative ones in order to predict long-term operating behaviour.  It should be noticed that the validity of any result is clearly dependent on the accuracy and correctness of the input data for the basic events.  

However, no single dependability analysis method is sufficiently comprehensive and flexible to deal with all the possible model complexities required to evaluate the features of practical systems (hardware and software, complex functional structures, various technologies, repairable and maintainable structures, etc.).  It may be necessary to consider several complementary analysis methods to ensure proper treatment of complex or multi-functional systems.  

In practice, a composite approach, with top-down and bottom-up analysis complementing one another, has proven to be very effective, in particular with respect to ensuring the completeness of the analysis.