基于概率熵判据的大坝位移预警概型比较  

Comparison of early warning model for dam displacement with probability based on maximum entropy

作　　者：张炯 叶琦 

机构地区：[1]江西省赣抚平原水利工程管理局,江西南昌330096

出　　处：《江西水利科技》2015年第6期405-409,共5页Jiangxi Hydraulic Science & Technology

摘　　要：基于数理统计原理,依据概率分布采用小概率估计方法拟定大坝位移预警指标是实现大坝安全监控的常用手段.借助系统可靠度分析方法,选取能够良好刻画系统服役性态的威布尔(Weibull)分布(三参数及二参数)、伽玛(Gamma)分布和对数正态(Log-normal)分布,针对大坝典型坝段位移监测序列,分别建立其概率分布模型,实现对分布参数的求解,再以最大熵原理为判据,选取能最客观描述大坝服役性态的分布,结合大坝失事概率,完成概率分布模型比较研究和位移预警值估计工作.实例表明:相较于其他分布,三参数Weibull分布能最客观描述典型坝段服役性态,故利用其对坝段位移预警值进行估计更加可靠.With the help of system reliability analysis method, the distribution of Weibull, Gamma and Log-normal that can be used to describe the system service behavior characteristics was selected to establish models of probability distributions according to the sequences of typical dam sections which has great significance to achieve the dam safety monitoring. The distribution that can describe the dam service behavior characteristics most objectively was chosen based on the criterion of the theory of the maximum entropy. The work that comparison between probability distributions and estimation of the value of displacement early warning was finished combined with the dam accident probability. The results showed that： the best model that can describe the typical dam sections service behavior characteristics most objectively is three-parameter Weibull distribution, it can be a more reliable way to estimate the value of dam displacement early warning.

关 键 词：概型比较 最大熵原理 位移预警 大坝 

分 类 号：TV698.11[水利工程—水利水电工程]