Streamline Simulation Handles Larger Models, Faster | SLB

Streamline Simulation Handles Larger Models, Faster

Published: 08/08/2007

Blue hero texture

Improve reservoir descriptions

Chevron uses reservoir simulation for waterflood design, infill/horizontal well planning, and coning studies. The accuracy of the reservoir simulation is largely controlled by inherent geological heterogeneities. The need to capture these fine-scale heterogeneities for improved reservoir descriptions prompted Chevron reservoir engineers to create large geological models with millions of cells.

Traditional finite-difference simulation is impractical for assessing the dynamic behavior of very large reservoir models. One Chevron 500,000-cell sandstone reservoir had more than 500 wells and 50 years of production history, with continuous drilling and frequent workover programs followed by a phased waterflood. Chevron reservoir engineers conducted comprehensive field-performance analyses to determine whether—in a number of U.S. reservoirs such as the super-giant sandstone reservoir—there were unrecovered reserves, bypassed oil, and infill drilling possibilities.

Enable fast, accurate simulation

Chevron decided to integrate ECLIPSE FrontSim streamline simulation software into its reservoir simulation workflow because of the simulator's versatile functionality. Streamline simulation offers a practical alternative technology for simulating multimillion-cell models, enabling fast and accurate simulation of large reservoir models and visualization of fluid flow paths throughout the reservoir as described in the following Chevron workflow.

Streamline Simulation Handles Larger Models, Faster
Chevron uses streamline simulation in five of the six reservoir simulation workflow steps (highlighted in blue).
  • Geological model screening and building—in addition to being large, current geological models are also stochastic; multiple, equiprobable realizations or models are obtained from the property distributions in 3D space. Ranking these realizations to identify the optimal realization for history matching is important. There is a range of ranking methods. In the ad hoc method, screening is minimal or even nonexistent; the history-matching process replaces the screening process. Another technique uses a simple statistical measure, such as connected volumes. Streamline simulation provides a more accurate flow-based screening of these large stochastic models. The screening process is prior to scale-up and is included in the geological model construction.
  • Upscaling—Chevron’s scale-up technique uses homogenization with nonuniform coarsening methods. The 3D streamline paths produced by streamline simulation can be used to guide nonuniform coarsening.
  • Assisted history matching—streamline simulation produces flow paths that can be used to guide history matching. This technique improves on ad hoc assumptions by explicitly identifying grid blocks that require modification.
Streamline Simulation Handles Larger Models, Faster
The ECLIPSE FrontSim results reproduce the behavior of the finite-difference model, even allowing for bigger time-steps in the streamline model. These watercut profiles show a very close agreement between the two methods.
Streamline Simulation Handles Larger Models, Faster
These two water saturation streamlines below—from 1966 (top) and 1994 (bottom)—illustrate aquifer and peripheral injection influence.

Reduce costs and time, increase accuracy

Chevron ran several tests proving that for two-phase problems the streamline results were virtually identical to those of finite difference methods. However, speed advantages of streamline simulators and the ability to simulate large reservoir models gives it an edge over finite-difference simulators. In the super-giant sandstone reservoir model with 500,000 cells and more than 50 years of production history, the simulation was completed in only 14 hours. In another example, a Rocky Mountain high-permeability fractured reservoir simulation run was 23 times faster than traditional finite-difference methods.

ECLIPSE FrontSim software—which runs on Linux® or Windows®—enables reservoir engineers to perform very fast and accurate simulation runs on large reser-voir models, reducing costs and history match cycle time. Streamline simulation, an effective solution at all stages of the reservoir characterization workflow, can be used in conjunction with conventional finite-difference methods as an assisted history-matching process and as a validation of upscaling techniques. Used alone, ECLIPSE FrontSim software enables screening of geological model realizations.

 

 

 

“FrontSim lets us quickly and accurately simulate and analyze reservoir models with millions of cells.”

Chevron reservoir engineer

Location
Onshore
Details

Challenge: Capture fine-scale heterogeneities for more accurate multimillion-cell, stochastic model simulations to investigate potential remaining reserves or drilling opportunities in large U.S. reservoirs.

Solution: Chevron integrated ECLIPSE FrontSim software into its reservoir simulation workflow for more accurate, flow-based model screening, with 3D streamline paths to guide nonuniform coarsening and flow paths to guide history matching.

Result: Streamline simulation of the super-giant sandstone reservoir model was completed in only 14 hours, and a Rocky Mountain high-permeability fractured reservoir simulation was run 23 times faster than with traditional finite-difference methods.

Products Used