CMG Collaborative Research: Stochastic Multiscale Modeling of Subsurface Flow and Reactive Transport

该CMG项目的总体目标是开发数学模型和高效准确的数值算法，以解决高度非均质多孔介质中流动和反应输运的大规模问题。目前的项目涉及三个跨学科的地球系统建模研究小组：俄克拉何马大学（OU）、匹兹堡大学（UPitt）和德克萨斯大学奥斯汀分校（UT-Austin）。组建的研究团队建议通过提高对控制多个空间和时间尺度下地下现象的物理过程的理解，推进必要的数学和地球科学基础，以增强模拟器的预测能力。目标应用包括复杂地球系统的可靠和高效建模，不确定性评估以及随机和确定性模型的有效耦合。本项目旨在取得以下成果：(1)开发和分析用于估计随机系统物理特性和统计量的新型离散化方法；(2)建立多尺度随机问题模型，量化非均质性的大尺度不确定性和子域系统参数的小尺度不确定性；(3)提高随机反应输运精度和效率的迭代耦合和时间步进研究；(4)涉及蒙特卡罗模拟和传统矩展开的有效替代方法的随机问题大规模求解方法的数值实现。人类与广泛的自然和工程地球系统相互作用，包括垃圾填埋场、污染场地、含水层和化石燃料储存库。因此，理解和模拟复杂的地球系统对于管理和优化环境清理和能源生产活动至关重要。这种CMG的努力提供了在大规模地理系统不确定性评估领域的学术、政府和工业合作的可能性，具有推动软件商业化的潜力。此外，在该项目下开发的预测和计算工具将影响当前对各种系统的科学理解，包括生物组织、大气、多孔复合材料和智能材料。拟议的研究将涉及在开放大学，UPitt和UT-Austin的跨学科环境中培训本科生，研究生和博士后研究员。

The overarching objective of this CMG project is to develop mathematical models and efficient and accurate numerical algorithms for solving large-scale problems of flow and reactive transport in highly heterogeneous porous media. The present project involves three interdisciplinary geosystems modeling research groups: The University of Oklahoma (OU), University of Pittsburgh (UPitt), and The University of Texas at Austin (UT-Austin). The assembled research team proposes to advance the mathematical and geoscience foundations necessary to enhance the predictive capabilities of simulators through an improved understanding of the physical processes that govern subsurface phenomena on multiple spatial and temporal scales. Target applications include reliable and efficient modeling of complex geosystems, uncertainty assessment and the effective coupling of stochastic and deterministic models. The project aims to achieve the following results: (1) development and analysis of novel discretization methods for estimating physical characteristics and statistics of stochastic systems; (2) modeling of multiscale stochastic problems for quantifying large-scale uncertainty in heterogeneity and small-scale uncertainty in subdomain system parameters; (3) investigation of iterative coupling and time stepping for improving accuracy and efficiency of stochastic reactive transport; and (4) numerical implementation of large-scale solution methods for stochastic problems involving efficient alternative methods to both Monte Carlo simulations and traditional moment expansions.Humankind interacts with a broad range of natural and engineered geosystems, including landfills, contaminated sites, aquifers and fossil fuel reservoirs. Therefore, understanding and simulation of complex geosystems are essential in managing and optimizing environmental cleanup and energy production activities. This CMG effort offers the possibility of academic, governmental, and industrial collaboration in the area of large-scale geosystems uncertainty assessment with the potential of driving software commercialization. Moreover, the predictive and computational tools developed under this project will impact current scientific understanding of a diverse array of systems, including biological tissues, the atmosphere, porous composite materials, and smart materials. The proposed research will involve the training of undergraduates, graduate students and post-doctoral fellows at OU, UPitt and UT-Austin, in an interdisciplinary environment.