Back to Top
Skip to main content
Industrial Compositional Streamline Simulation for Efficient and Accurate Prediction of Gas Injection and WAG Processes
Project Number

The goal of this project is to show that compositional streamline simulation (CSLS) has very high potential for the simulation of (near) miscible gas injection processes. In this project, researchers are extending and improving a novel and fully adaptive compositional streamline simulator CSLS to three-phase flows. The goal is to increase its computational efficiency so that it will be suitable for use in industry and laboratories for the study of realistic reservoir and production scenarios for a wide variety of gas injection and CO2 sequestration projects. The simulator development is supported by an experimental program designed to improve understanding of three-phase flows, which could lead to improved mathematical models for three-phase relative permeability functions.


Stanford University, Stanford, CA


When gas displaces oil at a sufficiently high pressure, the local displacement efficiency can be high. Due to reservoir heterogeneity and the low viscosity of gas, however, the injected gas may contact only a small portion of a reservoir as it tends to follow the high-permeability flow paths, thus limiting global sweep efficiency. The process performance of gas injection schemes depends upon balancing local displacement efficiency with global sweep efficiency, and both need to be captured accurately by a performance prediction tool.

This situation presents three substantial challenges for reservoir simulators, particularly when a compositional model is used to describe the reservoir fluids:

  • Because the global sweep depends foremost on the underlying heterogeneity and gravity, realistic, high-resolution reservoir models are required for accurate predictions.
  • In modeling the complicated multi-phase, multi-component flows, it is paramount to have a good physical understanding of the interactions between components in the flowing phases.
  • The effects of the physical models used on the performance of the numerical models must be studied carefully, and numerical models must be designed to accurately capture the essential physics.

As a result of these requirements, the computational costs of a reliable compositional reservoir simulation are very high, so computational efficiency becomes critical. In addition, since the physics of multi-phase, multi-component systems is not fully understood, an experimental program that increases this physical understanding is essential.

Achievements and successes include the following:

  • Designed, implemented and tested an improved mapping algorithm to interpolate data from the pressure grid to the streamlines and vice versa without introduction of excessive smoothing.
  • Designed, implemented and tested a high order global time-stepping routine that decreases errors introduced by the sequential solve of pressure and transport equations.
  • Designed, implemented and tested a higher order discretization scheme for the compositional transport solve along streamlines.
  • Designed, implemented and tested a pressure solver on adaptive Cartesian meshes based on variable and compact multi-point flux approximations that lead to a monotone pressure solution. The stencil construction is integrated with a specialized upscaling method to reduce upscaling errors.
  • Developed and implemented a streamline tracing algorithm for adaptive Cartesian meshes. This method is now in the testing phase.
  • Thoroughly investigated parallelization of streamline methods on shared memory machines. Formulated best practices for such parallelization, and tested these practices on three shared memory platforms in use in industry. The outcome of this work has been implemented in the commercial streamline simulator 3DSL (Streamsim Inc.).
  • Designed and implemented a first extension of compositional streamline simulation to non-isothermal processes.

The CSLS has a high potential for use in quick and reliable assessments of reservoir performance and development of management strategies for a wide range of reservoirs involving multi-phase flows. The high-resolution capabilities of CSLS will contribute to a better understanding of the physical processes of multi-phase, multi-component flow in porous media.

CSLS will provide fast and accurate performance predictions of oil and gas reservoirs. It is particularly suited for coupling to optimization algorithms for well placement, surface facilities locations, or reservoir management strategies and will enable optimization studies to be performed in realistic timeframes by industry. This is especially important in remote and/or hostile areas where infill drilling may not be an option and production costs are high. Hence, this proposal addresses the Nation’s need to increase domestic oil production in an environmentally sound manner.

This research also will contribute significantly to the study of CO2 sequestration via injection into subsurface formations. Simulations are needed to predict where CO2 is likely to flow, interpret the volume and spatial distribution of the subsurface reservoir contacted by injectant, and optimize continuous injection operations. CSLS is highly suitable for CO2 sequestration studies because of its ability to handle complex phase behavior and turnaround solutions very efficiently.

The purpose of this challenging numerical and computational project is to develop a reliable and very efficient simulator for assessment and optimal design of gas injection processes. An advanced, and relatively costly recovery process is more likely to be implemented if accompanied with a reliable simulator. Current simulation of gas injection processes is computationally very expensive, requiring either the use of very coarse grids, which leads to incorrect predictions of the manner in which the gas sweeps the oil reservoir, or simplified physics, which makes it very difficult to correctly predict the behavior of the injection process. Streamline simulation is suitable for gas injection processes that are strongly advection dominated, provided the errors introduced by decoupling the pressure and transport solvers, which are solved on two different grid systems, and the mappings between the two grid systems are reduced and controlled. Also, for industrial applications, it is paramount to implement the streamline simulator on parallel computers. Although it has often been claimed that streamline solvers are trivially parallelizable, no parallel streamline simulator exists to date.

The specific objectives for this research include:

  • Identify and understand the main sources of errors introduced in streamline simulation of the very strongly nonlinear gas injection processes.
  • Suggest and test novel numerical methods to control and reduce these errors.
  • Investigate efficient parallelization techniques for streamline solvers on shared memory systems, which are the computer systems of the future.
  • Design a parallel compositional streamline simulator.
Current Status

(January 2009) 
This project has been completed and the final report is available below under "Additional Information".

This project was selected under DOE’s Oil & Gas Program master solicitation DE-PS26-04NT15450 (focus area: Gas Processing).

Project Start
Project End
DOE Contribution


Performer Contribution

$210,026 (21 percent of total)

Contact Information

NETL - Chandra Nautiyal ( 918-699-2021)
Stanford U. - Margot Gerritsen ( or 650-725-2727)

Additional Information

Final Project Report [PDF-6.12MB]

Semi-annual reports for October 2004 through June 2007 available from DOE.

Younis, R., and Gerritsen, M.G., “Multiscale Process Coupling by Adaptive Fractional-Stepping: An In-Situ Combustion Model,” Proceedings of SPE/DOE Symposium on Improved Oil Recovery, 2006, SPE 93458.

Mallison, B.T., Gerritsen, M.G., and Matringe, S.F., “Improved Mappings for Streamline-Based Simulation,” SPE Journal, Volume 11, Number 3, September 2006, pp. 294-302, SPE 89352.

Jessen, K., and Gerritsen, M.G., “High-Resolution Prediction of Enhanced-Condensate-Recovery Processes,” Proceedings of SPE/DOE Symposium on Improved Oil Recovery, 2006, SPE 99619.

Nilsson, J., Gerritsen, M.G., and Younis, R., A Novel Adaptive Anisotropic Grid Framework for Efficient Reservoir Simulation, Proceedings of the SPE Reservoir Simulation Symposium, 2005, SPE 93243.

Nilsson, J., Gerritsen, M.G., and Younis, R., An Adaptive, High-Resolution Simulation for Steam-Injection Processes, Proceedings of the SPE Western Regional Meeting, 2005, SPE 93881.

Lambers, J.V., and Gerritsen, M.G., An Integration of Multilevel Local-Global Upscaling with Adaptivity, Proceedings of SPE's Annual Technical Conference and Exhibition, 2005, SPE 97250.

Jessen, K., Gerritsen, M.G., Lambers, J.V., and Mallison, B.T., A Fully Adaptive Streamline Framework for the Challenging Simulation of Gas-Injection Processes, Proceedings of SPE's Annual Technical Conference and Exhibition, 2005, SPE 97270.

AMR is used to construct a grid that is dynamically refined near solution fronts as they advance, as shown in this gas injection problem with an injector in the lower left corner.
AMR is used to construct a grid that is dynamically refined near solution fronts as they advance, as shown in this gas injection problem with an injector in the lower left corner.