Abstract
Fault tree analysis is an analysis tool that shows how a particular undesired event or accident can occur graphically in a deductive way. It is widely used for probabilistic risk assessment of hazard. However, the fault tree construction is in general very time-consuming . Also, the different fault tree may result from the same system by different engineers or the same engineer at different times. The main purpose of this paper is to study how to synthesize die fault trees in (1) more efficient and rapid (automation), and (2) standardized manner. Many types of unit models have been proposed by different authors. Theoretically, the quality of the fault tree synthesis mainly depends on the quality of the unit models. In this thesis, a new unit model called "Index Table" is proposed. Also, some decomposition strategies to decompose a system to reduce the system size are presented to help us input the system more effiency.