渗透系数的条件模拟对污染物运移的不确定性分析

Uncertainty analysis of the contaminant transport fate using conditional simulation of hydraulic conductivity

应用基于顺序条件模拟的蒙特卡罗方法,分析不同的先验渗透系数对污染物运移的不确定性影响.研究结果证明渗透系数的条件模拟可大大降低污染物运移结果的不确定性;随着渗透系数条件点数的变化,通过模拟得到的含水层中污染羽的空间矩(一阶矩和二阶矩)变化并不明显,各点的污染物浓度均值变化也不大.而污染物浓度变化的最大方差(不确定性)则随着渗透系数条件点数的增加而减小,同时运移结果的不确定性还取决于渗透系数条件点在空间上的分布位置.最后,根据渗透系数不同的条件模拟对污染物运移结果的响应,对实际工作中选取条件点的多寡提出相应的建议.

On the assumption that the hydraulic conductivity field follows a lognormal distribution, the sequential Gaussian simulation （SGS）, a commonly used conditional simulation method, is introduced to generate multiple realizations of the hydraulic conductivity field. Then Monte Carlo method, combined with the SGS method, is applied to investigate the effect of the conditional fields of hydraulic conductivity on the uncertainty regarding the contaminant transport fate. For the contaminant plume examined in the synthetic example in this paper the application of conditional simulation to generating hydraulic conductivity field leads to dramatic uncertainty reduction of the result output from the transport model, compared with those results based on unconditional simulations. It is found that the change of the averaged plume in terms of both spatial moments and concentrations output from the transport model is insensitive to the number of the conditioning data of hydraulic conductivity, whereas the peak variance of the plume concentrations at node decreases with the addition of the conditional hydraulic conductivity data. Moreover, the uncertainty of the modeled plume is dominated by the spatial distribution of the certain conditional hydraulic conductivity data. Additional suggestion is made that optimization techniques be used to ascertain the number of the conditional hydraulic conductivity data at the field site.