臺灣博碩士論文加值系統

事件樹 (ETA) 是一個使用在風險評估上簡單直接的方法。它可提供使用者一個清楚的圖像，使其瞭解何種頂頭事件主控系統的安全。傳統的 ETA 方法使用單一機率值來代表每個頂頭事件的發生率。然而，由於事件狀態本質上的不確定性，因此使用單一明確值來代表其發生率是不合理的。因為模糊理論提供了處理這種不確定性的一個有效工具，所以本論文使用它為數學工具。本論文的主要目的是建立一個評估人為失誤的簡單方法並將之整合進入 ETA。另外，也發展了一套 FETA 的演算法來評估大型系統的風險，並以核能電廠ATWS 事件的一實例來說明此一演算法。

Event tree analysis (ETA) method is a straightforward and simple approach for risk assessment. It can be used to identify various sequences and their causes, and also to give the analyst the clear picture about which top event dominates the safety of the system. The traditional ETA uses a single probability to represent each top event. However, it is unreasonable to evaluate the occurrence of an event by using a crisp value without considering the inherent uncertainty and imprecision a state has. Since fuzzy set theory provides a framework for dealing with this kind of phenomena, this tool is used in this study. The main purpose of this study is to make an effort in constructing an easy methodology to evaluate the human error and integrates it into ETA by using fuzzy concept. In addition, a systematic FETA algorithm is developed to evaluate the risk of a large scale system. A practical example of an ATWS event in a nuclear power plant is used to demonstrate the procedure. The fuzzy outcomes will be defuzzified by using the total integral value in terms of the degree of optimism the decision maker has. At last, more information about the importance and uncertainty of top events will be provided by using the two indices.