基于不确定时间区间TBN的洪水漫坝风险分析

Risk Analysis of Flood Overtopping Based on Uncertain Time Interval TBN

洪水漫坝过程中,自然条件、坝体及应急处置等状态的不确定性所导致的事件持续时间的不确定性,以及应急处置措施的效果是影响漫坝风险的关键因素。首先将传统时间区间改造为不确定时间区间,提出了一种不确定时间区间时间贝叶斯网络(TBN),并推导了该TBN的推理算法。在此基础上构建了考虑应急处置措施的洪水漫坝风险分析模型。仿真结果表明该模型具有信息时间累积能力,能判断各应急处置措施的重要程度,表达了洪水漫坝过程中同时存在的状态不确定性和时间不确定性,提高了漫坝风险分析的精度。

In the process of flood overtopping,the uncertainty of events’ duration time caused by state uncertainty of natural conditions,dam and emergency measures,and the effect of emergency measures are both key factors that affect the flood overtopping risk.The traditional time interval was transformed into an uncertain time interval and an uncertain time interval Temporal Bayesian Network（TBN） was proposed.After that,a new inferring algorithm of the TBN was discussed.On the basis,flood overtopping risk analysis model,in which emergency measures were considered,was established.Finally,the simulation result shows that the model can accumulate information,judge the importance of each emergency measure,express the state uncertainty and time uncertainty which both exist in the process of flood overtopping and accordingly increase the accuracy of flood overtopping risk analysis.