Parallel Computing Techniques for Large-Scale Reservoir Simulation of Multi-Component and Multiphase Fluid Flow
Bibliographic record
Abstract
Parallel Computing Techniques for Large-Scale Reservoir Simulation of Multi-Component and Multiphase Fluid Flow K. Zhang; K. Zhang Lawrence Berkeley National Laboratory Search for other works by this author on: This Site Google Scholar Y.S. Wu; Y.S. Wu Lawrence Berkeley National Laboratory Search for other works by this author on: This Site Google Scholar C. Ding; C. Ding Lawrence Berkeley National Laboratory Search for other works by this author on: This Site Google Scholar K. Pruess; K. Pruess Lawrence Berkeley National Laboratory Search for other works by this author on: This Site Google Scholar E. Elmroth E. Elmroth Lawrence Berkeley National Laboratory Search for other works by this author on: This Site Google Scholar Paper presented at the SPE Reservoir Simulation Symposium, Houston, Texas, February 2001. Paper Number: SPE-66343-MS https://doi.org/10.2118/66343-MS Published: February 11 2001 Cite View This Citation Add to Citation Manager Share Icon Share Facebook Twitter LinkedIn Email Get Permissions Search Site Citation Zhang, K., Wu, Y.S., Ding, C., Pruess, K., and E. Elmroth. "Parallel Computing Techniques for Large-Scale Reservoir Simulation of Multi-Component and Multiphase Fluid Flow." Paper presented at the SPE Reservoir Simulation Symposium, Houston, Texas, February 2001. doi: https://doi.org/10.2118/66343-MS Download citation file: Ris (Zotero) Reference Manager EasyBib Bookends Mendeley Papers EndNote RefWorks BibTex Search nav search search input Search input auto suggest search filter All ContentAll ProceedingsSociety of Petroleum Engineers (SPE)SPE Reservoir Simulation Conference Search Advanced Search AbstractMassively parallel computing techniques can overcome limitations of problem size and space resolution for reservoir simulation on single-processor machine. This paper reports on our work to parallelize a widely used numerical simulator, known as TOUGH2, for nonisothermal flows of multi-component, multiphase fluids in three-dimensional porous and fractured media. We have implemented the TOUGH2 package on a Cray T3E-900, a distributed-memory massively parallel computer with 695 processors. For the simulation of large-scale multicomponent, multiphase fluid flow, the requirements for computer memory and computing time are extensive. Because of the limitation of computer memory in each PE (processing element), we distribute not only computing time but also the memory requirement to different PEs. In this study, the METIS software package for partitioning unstructured graph and meshes is adopted for domain partitioning, and the Aztec linear solver package is used for solving linear equation systems. The efficiency of the code is investigated through the modeling of a three-dimensional variably saturated flow problem, which involves more than one million gridblocks. The execution time and speedup are evaluated through comparing the performance of different numbers of processors. The results indicate that the parallel code can significantly improve capacity and efficiency for large-scale simulations.IntroductionTOUGH21,2 is a general-purpose numerical simulation program for multi-dimensional, multiphase, multicomponent heat and fluid flows in porous and fractured media. The code is written in standard ANSI FORTRAN 77. Since its release in 1991, the program has been used worldwide in geothermal reservoir engineering, nuclear waste isolation, environmental assessment and remediation, and modeling flow and transport in variably saturated media. The numerical scheme of the TOUGH2 code is based on the integral finite difference (IFD) method. The conservation equations involving mass of air, water, chemical components and thermal energy are discretized in space using the IFD method. Time is discretized fully implicitly using a first-order backward finite difference scheme. The discretized nonlinear system of finite difference equations for mass and energy balances are solved simultaneously using the Newton/Raphson iterative scheme. For the basic version (i.e., single CPU), the code is equipped with both direct and iterative solvers.3The development of parallel computers has made it possible to conduct large-scale reservoir simulations. In the past decade, the total number of gridblocks used in a typical reservoir simulation increased from thousands to millions.4 One of the most popular parallel computer architectures is the distributed-memory machine, the massively parallel processor (MPP) computer, which can be made up of hundreds to thousands of processors. Elmroth et al.5 developed a parallel prototype scheme for the TOUGH2 code and implemented the computing time distribution on MPP computer. Their investigation indicates that a parallel code can dramatically enhance computational efficiency. Keywords: modeling & simulation, upstream oil & gas, artificial intelligence, multiphase flow, time step, fluid dynamics, equation system, linear equation system, computer, information Subjects: Reservoir Fluid Dynamics, Reservoir Simulation, Information Management and Systems, Flow in porous media, Multiphase flow This content is only available via PDF. 2001. Society of Petroleum Engineers You can access this article if you purchase or spend a download.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame machine prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. The Gemma side is a direct model label for every work in the frame, read from the title-only record. The Codex side is a classifier learned from the 10,348 direct Codex labels and calibrated to design-weighted sample rates; fields without enough sample support carry no Codex call. Candidate is the union of the two sides; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels.
Distilled classifier scores by category (both heads)
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.001 | 0.003 |
| Meta-epidemiology (narrow) | 0.001 | 0.001 |
| Meta-epidemiology (broad) | 0.001 | 0.001 |
| Bibliometrics | 0.001 | 0.002 |
| Science and technology studies | 0.001 | 0.001 |
| Scholarly communication | 0.001 | 0.002 |
| Open science | 0.001 | 0.002 |
| Research integrity | 0.001 | 0.002 |
| Insufficient payload (model declined to judge) | 0.003 | 0.001 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one source (direct Gemma or distilled Codex), not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".