Abstract: Based on the FTA model inference algorithm, the minimum cut set (path set) and structural importance of the root nodes were determined, and the occurrence and evasion scenario of runway conflict were analyzed accordingly. By the bidirectional reasoning technology of BN, the key risk influencing factors were analyzed, and by the risk factor polymorphism, risk assessment was conducted. Based on the risk analysis conclusion, risk control measures were proposed. The results of the study indicate that runway conflict risk at water aerodrome is most sensitive to the meteorological and hydrological factors, followed by the factors such as the frequency of passing vessels and the cleanliness of the sea runway, therefore, release control should be enforced strictly, and that safety management and personnel factors having a greater impact on runway conflict risk at water aerodrome should be the key point of risk management.

Key words: water aerodrome, runway conflict, risk analysis, Bayesian network(BN), fault tree analysis (FTA)