(emsl2558)Parameterization and Upscaling for Field-Scale Transport Modeling
EMSL Project ID
2558
Abstract
Field-scale modeling of water flow and solute transport processes in subsurface environments to address DOE's contaminant remediation and waste management problems frequently involves one or more of the following conditions: (1) large simulation domains, (2) the need to assess effects from multiple sites, (3) lack of data characterizing the small-scale heterogeneity of subsurface materials, and (4) a general requirement to assess uncertainties in the modeling results. Accurate parameterization and simulation of very large domains containing highly heterogeneous sediments has proven to be a particularly challenging problem, from both experimental and computational standpoints. The primary objectives of this research are: (1) to develop new methods and algorithms for estimating effective model parameters for simulation of field-scale flow and transport problems, and (2) to develop a framework (and software tools) for systematizing the process of model parameterization, and (3) to test and refine the methods, as needed, by comparing simulation results based on high-resolution and upscaled model parameterizations to data from field-scale flow and transport experiments. The high-resolution modeling will require the use of the parallel computing resources of the MSCF. Computationally efficient and accurate methods will be developed for upscaling the nonlinear constitutive pressure-saturation-permeability relations that are determined at the core scale (cm), to estimate effective parameters for much larger grid blocks (10's of m) that may be required for modeling field-scale problems in 3D. The accuracy of the parameterization and upscaling methods, and the uncertainty in flow and transport predictions, will be assessed by comparing simulation results to data from field-scale flow and transport experiments conducted in heterogeneous sediments at Hanford. The hysteretic nature of the constitutive relations, and the saturation-dependent anisotropy of relative permeability will also be considered in this work. The proposed work will leverage two existing projects. An ongoing EMSP project titled Quantifying Vadose Zone Flow and Transport Uncertainties using a Unified, Hierarchical Approach (EMSP project 70187, PI - Philip Meyer, PNNL) has been developing some of the parameterization and upscaling methods described in the objectives. Preliminary simulations have been conducted of a field-scale flow and transport experiment that was conducted at Hanford ( http://vadose.pnl.gov ), using upscaled model parameters with relatively coarse spatial discretization of the modeled domain (~104 nodes). Further testing of the parameterization and upscaling methods and refinement of the numerical algorithms will require simulation of the field experiment using a large model domain with very fine spatial discretization, owing to the highly heterogeneous nature of the sediments at the site. High-resolution models of the flow transport properties at the field site have been developed for a model domain consisting of 3.4 million grid blocks. Simulation of this problem can only be done by parallel computation. Steve Yabusaki and Mark White of PNNL have been developing a parallel version of the multiphase subsurface transport simulator, STOMP90, under the LDRD project Advanced Simulation Software for Subsurface Science. Our goal is to apply the STOMP90 simulator to the problem described above using the computational resources of the MSCF.
Project Details
Project type
Capability Research
Start Date
2002-07-19
End Date
2002-11-12
Status
Closed
Released Data Link
Team
Principal Investigator
Team Members