摘要
针对大坝安全分析预警,本文采用k-mean及局部离群系数方法,通过分析大坝渗流监测数据LOF值的时空分布及演变规律,揭示了渗流时序具有标度不变性,并有效检验局部离群及区域异常值,定量判断监测数据的异常情况。结合工程案例,采用matlab程序计算了LOF值,基于统计学方法确定了大坝渗流发生破坏的预警阈值,为开展基于监测预警提供技术参考。
For the dam seepage safety analysis and early warning,this paper adopts the K-mean and local outlier coefficient method.By analyzing the temporal and spatial distribution and evolution law of LOF value of dam seepage monitoring data,it reveals that the seepage time series perform scale invariance,which can effectively test local and regional outliers,and quantitatively judge the abnormality of monitoring data.Combined with the engineering case,the LOF value is calculated by MATLAB program,and the early warning threshold of dam seepage damage is determined based on statistical method,which provides a technical reference for monitoring and early warning.
作者
陶源
周志维
华剑峰
TAO Yuan;ZHOU Zhiwei;HUA JianFeng(Jiangxi Jianhong Water Conservancy Consulting Co.,Ltd.,Nanchang Jiangxi,330029,China;Jiangxi Academy of Water Science and Engineering,Nanchang Jiangxi,330029,China)
出处
《江西水利科技》
2023年第1期18-22,共5页
Jiangxi Hydraulic Science & Technology
基金
江西省水利厅科技项目(202223YBKT14).
关键词
渗流
预警
局部离群系统
数据挖掘
Seepage
Early warning
Local anomaly system
Data mining
