Application of bayesian network in safety analysis on civil aircraft systems

Civil aircraft system is one of the typical complicated large systems. Various factors which interplay with each other pose effects on the system safety and form a complicated logical relation. Therefore complicated systems such as civil aircraft system require scientific and effective system safety assessment process to ensure its safety. Currently for the International Civil Aviation industry, the guidance document used for civil airborne system and equipment safety assessment is ARP 4761 (SAE, 1996). Fault Tree Analysis (FTA) has been widely applied in the system safety analysis. But this method cannot perform the dynamic characteristics of system failure logic, and it is very difficult to determine system failure modes and assess failure probability. Dynamic Fault Tree [2] (Gulati et al., 1997) can describe the dynamic performance of system by introducing dynamic logic gate. However, the uncertainty of the system has obviously strengthened as the size of system increases. Then model building and bidirectional inference become very complex. FTA and DFT Analysis have limitation in dealing with these problems.