基于贝叶斯网络的沙河集水库大坝风险分析  

Dam risk analysis of Shaheji Reservoir based on Bayesian network

作　　者：周建方[1] 张迅炜[1] 唐椿炎[1] 

机构地区：[1]河海大学机电工程学院,江苏常州213022

出　　处：《河海大学学报（自然科学版）》2012年第3期287-293,共7页Journal of Hohai University(Natural Sciences)

基　　金："十一五"国家科技支撑计划(2006BAC14B07);中央高校基本科研业务费专项(2009B30014)

摘　　要：根据沙河集水库大坝风险分析的内容和破坏模式,分别建立了18个事件树对应的18个贝叶斯网络,再根据引起大坝溃决的原因,合并成由坝体渗漏、坝基渗漏、坝体边坡逐渐破坏、涵洞渗漏及溢洪道渗漏导致溃决的5种状况的贝叶斯网络;根据这5个贝叶斯网络的结构形式和变量情况合并成3个贝叶斯网络;最后进一步凝炼综合成1个贝叶斯网络,该贝叶斯网络至少包含6×18个事件树的信息.运行该贝叶斯网络,可以得到所需要的各种结果,显示出贝叶斯网络在风险分析中的特点和功能.该方法为水库大坝安全风险评价提供了一种新的思路.According to the content and failure mode of darn risk analysis of the Shaheji Reservoir, 18 Bayesian networks （BNs） corresponding to 18 event trees were built, and the 18 BNs were incorporated into five BNs based on the causes of dam failure, which include seepage of the dam body and dam base, the failure of the side slope, and seepage of culverts and spillways. Then, the five BNs were united into three BNs according to their structure forms and variable states. Finally, a synthetic BN was generated, containing information on 6 × 18 event trees. This Bayesian network can be used to obtain various results that are needed, and proves that the BN is powerful in dam risk analysis. This method provides a new way of dam risk assessment for reservoirs.

关 键 词：贝叶斯网络 事件树 大坝风险 沙河集水库 

分 类 号：TV64[水利工程—水利水电工程] N945.1[自然科学总论—系统科学]