Oil & Gas - Maximising Recovery Programme (formerly SHARP) IOR Views

Introduction

Streamlines have shown significant promise in the area of miscible gas injection and compositional simulation. These are traditionally difficult problems to solve using cell-based simulation techniques, because large fluid mobility contrasts between injected gas and resident oil in conjunction with strong permeability/porosity contrasts can lead to severe time step restrictions. Streamline simulation largely overcomes these problems by decomposing the three-dimensional (3D) transport problem into a series of independent, one-dimensional (1D) problems along streamlines. The geometry of the streamlines is a direct reflection of preferential flow paths resulting from reservoir heterogeneity, and the time-of-flight (TOF) along the streamlines is a reflection of how quickly composition fronts will travel in the reservoir. Along the streamlines, the 1D transport problem can be solved efficiently and capture the chromatographic separation of components as they travel from the injection well to the production well. Streamlines are periodically updated to honour the changing mobility field and new wells coming on-line or old wells being shut-in.

It is important to underline that streamline simulation is generally considered a "reduced" physics approach, since the theory is framed by the assumption of fluid incompressibility, and dispersive phenomena such as capillary pressure, transverse physical dispersion, and fluid expansion are neglected. Streamline-based flow simulation can be used as a "first-order" approximation of how a reservoir might react to gas injection, particularly for projects that are plagued by early gas breakthrough caused by reservoir connectivity interacting with high injected fluid mobility. For such cases, reduced physics streamline-based flow simulation offers a very powerful alternative to more traditional simulation approaches, particularly for quantifying uncertainty associated with various field development scenarios.

Example Application: Alpine Field, Alaska

The Alpine field is located in the Prudhoe Bay oil field on the North Slope of Alaska, and contains approximately 430 million barrels of recoverable reserves and 1.0billion barrels of OIP. It is a Jurassic marine sandstone with 40-degree API gravity oil and an average thickness of 50ft. A major redevelopment plan for the Alpine Field included drilling horizontal wells and implementing a Miscible Water Alternating Gas (MWAG) process. Key strategy considerations required addressing issues such as: the optimal volume of water injection prior to gas injection; timing of gas injection cycles to delay water breakthrough at the producers; and optimal distribution of gas volumes among injectors.

The traditional reservoir engineering workflow to address such questions is to use a numerical flow simulator to evaluate a matrix of scenarios. While traditional finite-difference simulators can include all the relevant physics, the drawback when used in the context of multiple simulations is the excessive computational cost, particularly if more geological details is desired yet the mobility contrast between the resident and injected fluids is significant. Using a streamline-based approach, on the other hand, allows quick screening and ranking of scenarios as well as a more detailed base geological model. The following approach was used in the case of the Alpine field:

Step1: Calibrate Pseudo Miscible Solution to use Along Streamlines - Because streamline based simulation is a reduced physics approach, it is questionable how much physics the 1D solution for the miscible displacement along each streamline should include. An approach proposed by Thiele et al. [1] is to calibrate a simpler, pseudo miscible model using a full-physics finite difference model. This can be done using a 1D problem as shown in Figure 1, where a number of 1D profiles are created using the pseudo miscible model and the best match to the full-physics solution is considered an acceptable solution.

Figure 1: Using a Simple 1D Numerical Problem, Parameters of the Pseudo Miscible Model can be Calibrated Using a Full-Physics Simulator

Step 2: Compare Streamline Simulator and Full Physics Simulator on 3D Base Case - Once the 1D problem has been calibrated, the next step is to compare a 3D solution between the streamline simulator and a full physics finite difference simulator. Reasonable agreement between the two is important to lend credibility to the simulation of multiple scenarios using the streamline simulation alone. It is important to remember that the streamline simulation is essentially a way of decomposing a difficult 3D-transport problem into a number of simpler 1D problems. While there are a number of assumptions involved in this step, it is important to understand what the "first-order" effects are in the displacement and ensure these are not eliminated in a streamline simulation (i.e.: heterogeneity, well rates & locations, major fluid PVT behaviour, etc). Figure 2 shows the streamlines (left) for a 100 x 160 x 6 Alpine model (42,110 active cells), the corresponding Flux-Pattern map (middle) and the comparison with a full-physics simulator (right). The good agreement between the full-physics compositional simulator comes despite the simplified assumptions introduced by the pseudo-model along each streamline and a much smaller number of time steps to simulate the 36 year forecast (2352 time steps for FD compared with 79 for SL).

Figure 2: (Left) Streamlines at a Particular Time Step for the Alpine Field and (Middle) the Corresponding Flux-Pattern Map. (Right) Comparison of Field Production Rate Between a Full-Physics Finite Difference Simulator (2352 Time Steps) and 3DSL (79 Time Steps)

Step 3: Simulate Scenarios Using Streamline Simulator - At this point the streamline simulator can be used to test multiple field development scenarios. The speed and efficiency can be used to simulate a much larger number of cases than is typically possible with a finite difference simulator. In addition, data unique to streamlines, such as well-allocation factors, can be used to extend the reservoir engineering analysis, particularly in the context of improving sweep efficiencies between efficient injector-producer pairs and minimising breakthrough for inefficient injector-producer pairs. Figure 3 shows six simulations of MWAG scenarios using different injection frequencies (0.25, 0.50, 0.75, 1.0, and 2 years) for water and gas plus the waterflood only case for a 30 year forecast. Each run took approximately 15min on a P4/XP/3.2GHz PC. Notice how the simulations reveal that oil production is insensitive to the WAG-frequency (solid lines), but that there is a significant amount of incremental oil recovery compared to the base waterflood case (dashed lines).

Figure 3: Forecast For Several MWAG-Cycle Frequencies (0.25, 0.50, 0.75, 1, 2 Years) Compared to the Base Case of a Pure Waterflood. Each Simulation Took Approximately 15min on o P4/XP/3.2GHz PC for a Total Run Time of Approximately 90min

Conclusions

Streamline simulation is a powerful alternative to finite difference simulation when the primary focus is on understanding how geology and high, injected fluid mobility might interact in the context of miscible gas injection projects. Multiple development scenarios can quickly be evaluated and first order parameters affecting recovery and project NPV identified.

References

• Thiele, M.R., Batycky, R.P., and Blunt, M.J.: "A Streamline Based 3D Field-Scale Compositional Reservoir Simulator," SPE paper 38889 in proceedings of the 1997 ATCE, San Antonio, TX (October).
• 3DSL, User Manual v2.20, Streamsim Technologies, 2004.
• McKishnie, R.A., Malik, S., Chugh, S., Lavoie, R.G., and Griffith, P.J. "Streamline Technology for the Evaluation of Full Field Compositional Processes - Midale, A Case Study," SPE paper 89363 presented at the 14th SPE/DOE Symposium on Improved Oil Recovery, Tulsa, OK, 17-21 April, 2004.
• Redman, R.S.: "Horizontal Miscible Water Alternating Gas Development of the Alpine Field, Alaska," SPE paper 76819 presented at the SPE Western Regional Meeting, Anchorage, AK, 20th-22nd May, 2002.
• Seto, C.J., Jessen, K., Orr, F.M. Jr.: "Compositional Streamline Simulation of Field Scale Condensate Vaporization by Gas Injection," paper SPE 79690 in proceedings of the 2003 SPE Reservoir Simulation Symposium, Houston, TX (Feb. 2003).
• Thiele, M.R., Batycky R.P., and Kent, L.T.: "Miscible WAG Simulations Using Streamlines," in proceedings of the 8th European Conference on the Mathematics of Oil Recovery (ECMOR), Freiberg, Germany, 3-6 Sept., 2002.