The goals of this project were to:
This project was selected in response to DOE's Oil Exploration and Production solicitation DE-PS26-01NT41048, focus area Critical Upstream Advanced Diagnostics and Imaging Technology. The goal of the solicitation was to continue critical upstream cross-cutting, interdisciplinary research for the development of advanced and innovative technologies for imaging and quantifying reservoir rock and fluid properties for improved oil recovery
University of Utah
Salt Lake City, UT
Most oil reservoirs are fractured to a certain degree. The models currently used for fractured reservoir simulation do not explicitly consider the spatial fracture characterization. Instead, the fracture presence is smeared out in a dual continuum model. Better reservoir characterization methods are making fracture mappings available. When the spatial characteristics of fractures are considered, the resulting discrete fractur" models are also highly computation intensive. The purpose of this project was to create finite-element models that would be able to incorporate this characterization of faults/fractures directly.
The original reservoir simulators were developed using numerical methods and algorithms that were more suitable for the prevalent computational environments. The more advanced, faster conjugate gradient-like methods for the solution of linear equations, for example, have not been widely implemented. Parallelization also has been beyond the reach of the independent producers because the multiprocessor machines are very expensive.
There has been a great emphasis on the effect of reservoir geologic and petrophysical properties on reservoir performance; however, little research effort has been expended on understanding the effect of control variables on reservoir operation. This is a difficult constrained-optimization problem. When resolved, this will lead to optimum reservoir operation, continuous performance monitoring, and possibly automatic model updating based on reservoir history. The Internet is changing the way business is conducted, and it affords significant opportunities to make the oil business more efficient. Making simulators available on the Internet with online computing modules will open up the technology to independent producers. This project was performed to address some of these issues.
A three-dimensional, three-phase reservoir simulation of complex faulted/fractured reservoirs using the modern linear and nonlinear solvers is now possible. A client-server protocol will make it possible for interested engineers and scientists to create two-dimensional fractured domains and run data files remotely on the University of Utah servers.
If detailed fault/fracture characterization is available, it would now be possible to incorporate this explicitly into three-dimensional, three-phase reservoir simulators. An interactive tool developed as part of this project will help geoscientists experiment with the placement and properties of faults. A new set of tools to deal with complex faulted/fractured systems is now available to the industry.
Among the project highlights:
A three-dimensional, three-phase reservoir simulator based on the Control Volume Finite-Element (CVFE) architecture was developed. Results from this simulator were compared to output from Eclipse. Faults in Eclipse were modeled using a fine-mesh representation.
A new numerical model, based on the mixed finite-element method was developed and verified.
Several two- and three-dimensional simulations of two- and three-phase flow in faulted/fractured porous media were performed to demonstrate the applicability of the methods developed.
It was shown that incorporation of hydraulic fractures was straightforward using the approach described.
Most modern conjugate gradient numerical solvers were used. The simulators were linked to Portable Extensible Scientific Computation toolkit.
Parallel computation schemes using MPI were employed, and scalability was demonstrated on a 18-processor Linux cluster.
The feasibility of being able to embed optimization routines was demonstrated by linking with the Toolkit for Advanced Optimization.
An interactive module based on the client-server protocol was created. This module will let users run the simulators remotely on the University of Utah servers or any other servers where the programs would be installed. This provides affordable parallel computing access to independent producers.
The project is in the no-cost extension period. It was to conclude on August 31, 2005.
First Annual Report, September 2002; Second Annual Report, September 2003; Interim Progress Report.
$170,000 (21% of total)