基于向导点法反演水文地质参数  

Inversion of Hydrogeological Parameters Based on Pilot Point Method

作　　者：何金沙 李春光[1,2] 吕岁菊[2] 杨佩瑶 黄传霁 HE Jin-sha;LI Chun-guang;LÜ Sui-ju;YANG Pei-yao;HUANG Chuan-ji(School of Civil and Hydraulic Engineering,Ningxia University,Yinchuan 750021,China;Research Institute of Numerical Computation and Engineering Applications,North Minzu University,Yinchuan 750021,China)

机构地区：[1]宁夏大学土木与水利工程学院,银川750021 [2]北方民族大学数值计算与工程应用研究所,银川750021

出　　处：《节水灌溉》2021年第10期13-17,共5页Water Saving Irrigation

基　　金：国家自然科学基金项目(11761005);宁夏自然科学基金项目(NO.2021AAC03173)。

摘　　要：为了研究观测井与向导点布置范围以及水文地质参数初值对反演结果的影响,利用二维承压含水层理想模型,分别建立观测井、向导点不同分布范围(占研究区面积16%、36%、64%、81%和100%)以及不同渗透系数场初值的地下水反演模型,讨论其反演规律。其中,初始渗透系数场与实际渗透系数场之间的均方根误差记作R1,表示先验信息精度;初始渗透系数场经过PEST程序反演后的结果称为渗透系数估计场,与实际渗透系数场之间的均方根误差记作R2,表示参数估计精度,R2值越小反演精度越高。结果表明:随观测井、向导点分布范围增加,相应模型的R2值先减小后逐渐保持稳定;同时随向导点分布范围的增加,调用Modflow程序与优化迭代的次数减少;R1值增加,不同渗透系数初值模型的R2值、Modflow程序调用次数与优化迭代也会增加。由此看出,观测井与向导点分布范围越大,初始渗透系数场越接近真实值,反演结果越理想。研究成果为观测井与向导点的科学布置以及初始渗透系数场的取值提供了理论依据,有助于向导点法的推广。In order to study the influence of the layout range of observation wells and pilot points and the initial values of hydrogeological parameters on the inversion results,the groundwater inversion models with different distribution ranges of observation wells and pilot points(accounting for 16%,36%,64%,81%and 100%of the study area)and different initial values of permeability coefficient fields are established by using the ideal model of two-dimensional confined aquifer,and their inversion rules are discussed.Where,the root mean square error between the initial permeability coefficient field and the actual permeability coefficient field is recorded as R1,which indicates the prior information accuracy.While the result of initial permeability coefficient field inversion by PEST program is called permeability coefficient estimation field,and the root mean square error between initial permeability coefficient field and actual permeability coefficient field is recorded as R2,which indicates the parameter estimation accuracy.The smaller the R2 value,the higher the inversion accuracy.The results show that with the increase of the distribution range of observation wells and pilot points,the R2 value of the corresponding model decreases first and then gradually becomes stable.At the same time,with the increase of the distribution range of pilot points,the number of calling Modflow program and optimization iteration decreases;with the increase of R1 value,the R2 value,Modflow program call times and optimization iterations of different initial permeability models will also increase.It can be seen that the larger the distribution range between observation wells and pilot points is,the closer the initial permeability coefficient field is to the real value,and the better the inversion result is.The research results provide a theoretical basis for the scientific arrangement of observation wells and pilot points and the value of initial permeability coefficient field,which is helpful to popularize the pilot point method.

关 键 词：向导点法 向导点分布范围 观测井分布范围 水文地质参数反演 高斯-马夸尔特-列文伯格迭代算法 PEST 

分 类 号：S156.1[农业科学—土壤学]