地下水流数值模拟中随机因素的灵敏度分析  

Analysis on sensitivity of stochastic factors in numerical simulation of groundwater flow

作　　者：李森[1] 陈家军[1] 叶慧海[2] 孟占利[1] 

机构地区：[1]北京师范大学环境模拟与污染控制国家重点联合实验室,北京100875 [2]交通部公路科学研究院,北京100088

出　　处：《水利学报》2006年第8期977-984,共8页Journal of Hydraulic Engineering

基　　金：教育部重点基金项目(104012);国家重点基础研究发展规划(973)项目(G1999043606)

摘　　要：本文以待定系数摄动随机有限元法为基础,在以渗透系数、给水度、边界、源汇项为随机变量的条件下,建立了二维潜水非稳定流随机模型,编制了二维摄动随机有限元通用程序。通过对太原盆地地下水系统的随机模拟,分析了渗透系数、给水度、边界、开采量4个随机因子对水头模拟值的影响程度。结果表明,对水头期望值影响的灵敏度:渗透系数最大,其次为给水度和边界值,井开采量最小;对水头方差影响的灵敏度:给水度最大,其次为渗透系数和边界值,井开采量最小;对水头摄动量影响的灵敏度,两个随机参数中渗透系数较大,给水度较小。A 2-D stochastic finite element model of unsteady flow in phreatic aquifer based on perturbation coefficient awaiting determination stochastic FEM is proposed. In the model the infiltration coefficient, specified yield, boundary values and withdrawals are regarded as stochastic variables, and the program for computation of computer is developed. The effects of these four stochastic factors on simulation of water table are analyzed respectively through the stochastic simulation of groundwater flow in Taiyuan Basin, Shanxi Province. The results show that the ranking of sensitivities of four stochastic factors to expected water table according to their influence on simulation is as follows： the first one is infiltration, the next are specified yield and boundary values, the last is withdrawal. Whereas, the ranking for water table variances is： specified yield, infiltration coefficient, boundary value and withdrawal. The sensitivity of infiltration coefficient to water table perturbations is higher than that of specific yield.

关 键 词：灵敏度 随机因子 地下水流 数值模拟 太原盆地 

分 类 号：P641.2[天文地球—地质矿产勘探]