基于溃坝事件树的尾矿库水文隶属特征提取方法  

Method of Feature Extraction with the Tailings Safety based on Dam-break Event Tree

作　　者：陈晓方[1] 王刚毅[1] 

机构地区：[1]中南大学信息科学与工程学院,湖南长沙410083

出　　处：《有色冶金设计与研究》2015年第3期15-21,共7页Nonferrous Metals Engineering & Research

基　　金：国家"十二五"科技支撑计划(2012BAK09B04);国家自然科学基金(61374156)

摘　　要：在对尾矿库事故统计分析的基础上,首先选取发生概率最高的漫顶失事作为研究对象,并对不同时期尾矿库的危险源和致灾路径进行了分析,建立了漫顶失事的事件树模型。接着结合事件树模型与可能性理论提出了漫顶失事的水文特征提取流程,从事件树模型的路径分解中提取出高相关的监测参数和相应统计特征,然后采用三角可能性分布函数对参数进行处理,通过仿真确定隶属于漫顶事故的安全特征集。最后对尾矿库状态空间分布进行了实例计算,验证了事件树结合安全特征集进行风险分析的有效性。In this article we select the highest probability of overtopping crash as the research object based on the statistical analysis of tailings on the accident.And the sources of danger and hazard paths at different times of tailings were analyzed to establish the overtopping crash event tree model. Then we combineprobability theorywith the event tree model to propose overtopping crash hydrological characteristics extraction process.Next we extract the high correlation of monitored parameters and corresponding statistical characteristics from the path decomposition event tree model, then to use a probable triangular distribution function to do with the parameters. And to determine the safety feature set to a part of overtopping accident througth simulation. Finally, after the state of instances of the spatial distribution of tailings calculations to verify the safety featuresets combined with event treeis effective to risk analysis.

关 键 词：事件树 致灾路径 特征提取 可能性分布 安全隶属度 

分 类 号：TD76[矿业工程—矿井通风与安全]