摘要
本文从土石坝安全等级划分的有关规定出发,给出了动态的土石坝安全等级概率描述,利用Bayes方法,以有效挖掘和利用新信息,实现对土石坝安全性状更为合理的评估和预测。运用离散随机变量的Bayes表达式,将土石坝安全鉴定专家评级赋值的先验概率和实时检测信息的似然概率相综合,更新土石坝安全等级评定的概率信息,并将这一信息运用到土石坝除险加固排序等决策中。这一方法有利于减少大坝安全性定量评估的难度和不确定性,有利于实现与现行安全决策准则的衔接,从而使目前的大坝安全风险分析方法更趋实用。
Based on the regulation for division of security level of dams in China, the dynamic probability characterization of the security level of dams is suggested. By using the Bayesian approach and effective information mining as well as the utilization of renewed information, the more rational evaluation and prediction of the security level of dams can be realized. By adopting the Bayesian expression of discrete stochastic variables, the priority probability of the security level of dams given by experts is combined with the likelihood probability of the real-time check information, so that the probability information for the evaluation of security level of dams is renewed. Such probability index is then applied to the refusal decision-making of dam rehabilitation. The proposed method reduces the difficulty and uncertainty of the evaluation of security level of dams and realizes the smooth connection with the current decision-making rules in China. It is helpful to the applicability of the current analysis methods of security level of dams.
出处
《水利学报》
EI
CSCD
北大核心
2008年第8期922-926,共5页
Journal of Hydraulic Engineering
基金
国家自然科学基金项目(50579038
50679043)
国家"十一五"科技支撑计划项目(2006BAC14B03)
关键词
BAYES方法
大坝安全信息挖掘
安全等级动态概率评定
排序决策
大坝除险加固
Bayesian approach
dam safety information mining
dynamic probability evaluation of dam security level
refusal decision-making
dam rehabilitation
