A Next-Generation Reservoir Simulator as an Enabling Technology for a Complex Discrete Fracture Modeling Workflow
- Kok-Thye Lim (Chevron ETC) | Mun-Hong Hui (Chevron ETC) | Bradley T. Mallison (Chevron ETC)
- Document ID
- Society of Petroleum Engineers
- SPE Annual Technical Conference and Exhibition, 4-7 October, New Orleans, Louisiana
- Publication Date
- Document Type
- Conference Paper
- 2009. Society of Petroleum Engineers
- 5.6.4 Drillstem/Well Testing, 5.5 Reservoir Simulation, 5.1.5 Geologic Modeling, 5.4.2 Gas Injection Methods, 5.2.2 Fluid Modeling, Equations of State, 5.3.2 Multiphase Flow, 5.2 Reservoir Fluid Dynamics, 5.7.2 Recovery Factors, 5.4.7 Chemical Flooding Methods (e.g., Polymer, Solvent, Nitrogen, Immiscible CO2, Surfactant, Vapex), 5.6.9 Production Forecasting, 4.3.4 Scale, 5.8.7 Carbonate Reservoir, 5.1 Reservoir Characterisation, 5.5.8 History Matching, 5.2.1 Phase Behavior and PVT Measurements, 5.8.6 Naturally Fractured Reservoir
- 2 in the last 30 days
- 754 since 2007
- Show more detail
- View rights & permissions
|SPE Member Price:||USD 5.00|
|SPE Non-Member Price:||USD 28.00|
We present an innovative workflow for reservoir characterization, gridding, discretization, and simulation of discrete fractures embedded within a single-porosity continuum. This discrete fracture modeling (DFM) workflow provides the capability to realistically model the impact of fractures on recovery without incurring the simplifying assumptions of traditional dual-porosity models. We represent the matrix of the unstructured DFM using 3D polyhedral cells including tetrahedra, pyramids, and prisms. The fractures are represented as polygonal interfaces between matrix cells.
The workflow is enabled by a next-generation reservoir simulator that is designed for robust solution of unstructured grid models. The simulator supports parallel computation on distributed and shared memory machines without manual user intervention. The efficiency of the simulator is due primarily to a parallel linear solver that is designed specifically for unstructured grids. This results in a very efficient and robust reservoir simulator for highly heterogeneous DFMs.
We demonstrate the practicality of this technology by performing a design of experiments (DoE) study that involves a suite of sector models representing multiple discrete fracture realizations of an actual carbonate reservoir. We also show the feasibility of a full-field simulation including over ten thousand discrete fractures, more than a hundred wells, and 3.65 million simulation cells. These simulations are challenging due to the high contrast in matrix and fracture properties, multiphase flow, and the inclusion of complex physics (e.g., rapid gravity segregation in fractures, first- and multi-contact miscibility).
The successful application of the DFM workflow is largely due to the availability of the next-generation technology for reservoir modeling. We could not perform this work using existing simulation technologies designed for structured grids. Our DFM work facilitated the understanding of recovery mechanisms and contributed to business decisions toward optimal development of the carbonate reservoir.
There are many hydrocarbon assets that occur in carbonate reservoirs where the primary storage is in low permeability matrix and the primary fluid production conduit is through fractures. An excellent description of naturally fractured reservoirs (NFRs), ranging from characterization to production case histories, can be found in Narr, Schechter and Thompson (2006).
Application of reservoir simulation to NFRs started when Kazemi et al. (1976) adopted the dual-porosity concept introduced earlier in well test analysis (Warren and Root 1963) for application in multiphase flow simulation. The key assumption is that there exists a locally homogenized geometric shape factor for each grid cell that characterizes the interaction between the matrix and fracture media. Various forms of the shape factor have been derived, including methods to account for multiphase flow. The dual-porosity, and its extension to dual-porosity dual permeability, models requires the assumption of a representative element volume. Shape factors are often used as a history-matching parameter, rather than one that is derived from reservoir characterization.
|File Size||838 KB||Number of Pages||8|