Mimetic Finite Differences and Algebraic Multilevel Methods for Reservoir Performance Predictions: CRADA Final Report
- Los Alamos National Lab. (LANL), Los Alamos, NM (United States). Richard P. Feynman Center for Innovation
- Los Alamos National Lab. (LANL), Los Alamos, NM (United States)
The need to better predict and control reservoir performance is driving the development of increasingly sophisticated reservoir models that use unstructured general polyhedral grids to capture highly detailed stratigraphy. However, both discretization methods and scalable iterative solution algorithms for these situations are an active area of research. In this work we compared the performance of a new Mimetic Finite Difference (MFD) method with two new Mixed Finite Element methods, and LANL and the CRADA Participant investigated the performance of Los Alamos National Laboratory’s algebraic multigrid code on realistic flow problems. Furthermore, uncertainty quantification requires rapid evaluations of reservoir models. To do this, fine-scale reservoir models need to be coarsened and yet important fine-scale features and processes that have a significant impact on flows need to be preserved. To this end, we investigated the use of multi-resolution methods, such as the Multilevel Multiscale Mimetic (M3) method, to coarsen reservoir models.
- Research Organization:
- Los Alamos National Laboratory (LANL), Los Alamos, NM (United States); ExxonMobil Upstream Research Company, Houston, TX (United States)
- Sponsoring Organization:
- USDOE
- DOE Contract Number:
- AC52-06NA25396
- OSTI ID:
- 1330827
- Report Number(s):
- LA-CP-16-20433; LA-UR-24-21930; CRADA: LA08C10594-PTS-001
- Country of Publication:
- United States
- Language:
- English
Similar Records
Numerical Multilevel Upscaling for Incompressible Flow in Reservoir Simulation: An Element-Based Algebraic Multigrid (AMGe) Approach
Performance of fully-coupled algebraic multilevel domain decomposition preconditioners for incompressible flow and transport.