Stochastic modeling of solute transport in 3-D heterogeneous porous media with random source condition
Stochastic modeling of solute transport in 3-D heterogeneous porous media with random source condition
复制标题
具有随机源条件的 3-D 异质多孔介质中溶质输运的随机建模
DOI：
--
复制
发表时间：
2006
期刊：
影响因子：
--
通讯作者：
M. Sekhar
中科院分区：
文献类型：
--
作者：
A. Chaudhuri;M. Sekhar
在多孔介质流动输运的概率分析中，往往要考虑控制参数的空间非均质性所引起的不确定性。源条件的随机性对因变量分布的随机行为也起着重要作用。本文主要研究了控制系统参数不确定性和输入源条件的影响。针对这种情况，提出了一种与随机有限元法相结合的方法，并举例说明了随机源条件下三维非均质多孔介质中浓度分布的概率分析。在第一步中，由于控制参数的空间异质性，采用SFEM对单位源脉冲进行概率求解。此外，将单位源脉冲情况的结果应用于随机源条件下的多脉冲情况的数值卷积分析。源条件被建模为随机质量在固定时间间隔内的离散释放。比较了确定性系统和随机系统以及不同系统参数值下浓度的均值和标准差。源条件不确定度的影响还可以通过域内不同位置浓度的均值和标准差来证明。
During probabilistic analysis of flow and transport in porous media, the uncertainty due to spatial heterogeneity of governing parameters are often taken into account. The randomness in the source conditions also play a major role on the stochastic behavior in distribution of the dependent variable. The present paper is focused on studying the effect of both uncertainty in the governing system parameters as well as the input source conditions. Under such circumstances, a method is proposed which combines with stochastic finite element method (SFEM) and is illustrated for probabilistic analysis of concentration distribution in a 3-D heterogeneous porous media under the influence of random source condition. In the first step SFEM used for probabilistic solution due to spatial heterogeneity of governing parameters for a unit source pulse. Further, the results from the unit source pulse case have been used for the analysis of multiple pulse case using the numerical convolution when the source condition is a random process. The source condition is modeled as a discrete release of random amount of masses at fixed intervals of time. The mean and standard deviation of concentration is compared for the deterministic and the stochastic system scenarios as well as for different values of system parameters. The effect of uncertainty of source condition is also demonstrated in terms of mean and standard deviation of concentration at various locations in the domain.