To quantify the flood risks in cascade dam systems, it is critical to analyze the risk factors and potential breaking failure paths. In this study, Bayesian networks (BNs) were applied to create a flood risk analysis model for a cascade dam system. Expert experiment, historical data, and computational formulas were employed to estimate the prior probability and original conditional probability tables (CPTs) in the BN model; sensitivity analysis was used to ensure the original continuous breaking failure path in the system. To avoid the possible misperceptions of the probability of a certain event, Dam Breach Analysis Model (DB-IWHR) 2014 software and the flood regulation method were used to simulate the dam breaking progress. The posteriori continuous breaking failure paths were obtained, and then the original CPTs were refined based on the new evidence. The proposed method was applied to the Bala-Busigou-Shuangjiangkou (BL-BSG-SJK), which is located upstream of the Dadu River basin in China. The results show that three continuous breaking failure paths could be identified in the researched cascade dam system. A new BN model was created to determine the failure probability of the cascade dam system under the three continuous breaking failure paths. This analytical method may also be useful for other similar cases.
Paleoflood hydrology on the lower Green River, upper Colorado River Basin, USA: An example of a naturalist approach to flood-risk analysis
