点击次数：

论文类型：会议论文

发表时间：2010-01-01

收录刊物：CPCI-S

页面范围：101-+

关键字：dam overtopping; risk estimation; risk model; Latin hypercube sampling; uncertainty analysis

摘要：Complicated uncertainty exists in dam overtopping risk systems. It has been always one of the hot and difficult topics. In the traditional way, the non normal distribution variables need to be transformed to normal distribution, which would influence the risk analysis. In this study, based on the reservoir forecasted operation practice, the uncertain factors are addressed such as the flood forecasting error, initial water level, the operation delay, wind and wave, and reservoir release error, and so on. A risk analysis model is proposed for risk estimation of dam overtopping. Aiming at the different distribution variables from the uncertain system, Latin hypercube sampling is employed to sample variables' own distributions to solve the model, rather than all the transformed normal distribution. Moreover, it could implement the random combination sampling and could achieve more effective sampling. This presented methodology is a novel progress for dam overtopping risk analysis. It is applied to the risk analysis of dam overtopping for Biliuhe reservoir in Liaoning Province of China, and the risk estimation is evaluated.