基于SOM自组织神经网络和K-means方法探究地表水与地下水之间的水力联系  

Exploring Hydraulic Connections between Surface Water and Groundwater Based on SOM and K-means Algorithm

作　　者：张大龙 黄勇[1] ZHANG Dalong;HUANG Yong(School of Earth Sciences and Engineering,Hohai University,Nanjing 211100,Jiangsu,China)

机构地区：[1]河海大学地球科学与工程学院,江苏南京211100

出　　处：《水力发电》2025年第4期6-11,共6页Water Power

基　　金：国家自然科学基金长江水科学研究联合基金项目(U2240217)。

摘　　要：针对地表水与地下水之间的水力联系,引入SOM自组织神经网络和K-means方法,以华北平原某污染河段为研究对象,探讨地表水与地下水之间的水力联系。经分析,发现地表水和1号、2号、6号、7号观测井的地下水水质基本一致,水力联系较强;与3号、8号、9号、10号、12号、13号观测井的地下水水质差异较大,水力联系较弱,研究结果与传统系统聚类方法的结果基本一致。结果表明,此方法能够精确地判别地表水和地下水之间的水力联系,为识别不同含水层的水力联系提供了新的解决思路和技术手段。Aiming at the hydraulic relationship between surface water and groundwater,the SOM self-organizing neural network and the K-means method are introduced to investigate the hydraulic relationship between surface water and groundwater in a polluted river reach of North China Plain.The analyses show that the water quality of surface water is basically the same with the groundwater quality of 1#,2#,6#and 7#well,meaning a stronger hydraulic connection,and the water quality of surface water is different with the groundwater quality of 3#,8#,9#,10#,12#and 13#well,meaning a weaker hydraulic relationship.The analysis results are basically consistent with the results of traditional systematic clustering method.This research shows that the self-organizing neural network and K-means algorithm can accurately identify the hydraulic relationship between surface water and groundwater,which provides a new solution and technical means for identifying the hydraulic relationship between different aquifers.

关 键 词：地表水 地下水 水力联系 水化学分析 SOM自组织神经网络 K-MEANS 聚类分析 

分 类 号：P641.12[天文地球—地质矿产勘探]