The key deliverable for phase 2 will be a refined RRM workflow that is able to incorporate additional data sets, and model and simulate more complex reservoir and production scenarios. This workflow may have been refined after discussion and review with researchers from the industry partners, in order to incorporate scenarios that are of key interest to all industry partners. We currently envision the following deliverables, which will be reviewed and amended jointly with the industry partners after completion of phase 1.

Input Data (1)

Data input capability will be extended to include (a) depth-converted seismic data, (b) data from wells (well logs, well tops), (c) models from GereSim, in collaboration with Dr. Waldemar Celles (TecGraf/PUC-Rio) and (d) outcrop data, such as spatially located photo-panoramas. The new data sources will be combined with the data inputs developed in year 1, to allow creation of hybrid models from a variety of sources. In addition, year two will see further expansion of the library of template surfaces for SBIM, to include a broader range of geological shapes. These shapes could be identified in collaboration with the industry partners.

Surface Extraction and Manipulation (2)

New surfaces can be extrapolated and/or created from seismic and well log data, and merged with surfaces generated from different data sources to create new reservoir geometries. A key deliverable will be functionality to modify existing surfaces such that they are consistent with well data. Tools from phase 1 and phase 2 will be integrated with interactive multi-touch display surfaces, such as the Microsoft Surface (https://www.microsoft.com/surface/en/us/default.aspx).

Gridding (3)

Unstructured gridding capabilities will be extended based on the well-established Computational Geometry Algorithm Library (CGAL – http://www.cgal.org/). This will allow for the automatic translation of surfaces and wells generated in (2) and/or imported in (1) into high quality triangular and tetrahedral finite element meshes and complementary finite volumes, including mesh repair functionality. Using fully unstructured tetrahedral elements and adaptively refined grids will allow us to discretize highly complex geometries, surfaces and volumes compared to standard pillar gridding and basic tetrahedral grids implemented in Phase 1. The project will hire a dedicated CGAL employee part-time during this phase to ensure that CGAL is fully integrated into the RRM workflow as needed, and meshing is fully automated with little user interference. Both finite element and finite volume techniques will be provided to aid mechanical and flow simulations building on the on-going research activities of HWU, ICL and UFPE in this field.

Characterisation (4)

The library of default rock types and rock properties will be expanded to include a broader variety of rock types and corresponding rock properties. These rock types and properties could be identified in collaboration with researchers at the industry partners. A library for basic fluid properties (e.g. viscosity, density, formation volume factor) will also be developed, again in collaboration with researchers at the industry partners.

Computation (5)

The key incremental development in year 2 is to add the capability to perform dynamic simulations to the static calculations developed in year one, including

1. Transient well-test response for (extended) injection and withdrawal tests (single-phase flow only) can be simulated; this is an important extension to the calculations developed during Phase 1, which only allow static reservoir conditions to be computed. The work may include implementing a more realistic well-model (e.g., using analytical solutions to account for pressure drop in wells).
2. The stress state in the reservoir can be computed in response to the injection and/or withdrawal of reservoir fluids.
3. The RRM framework can be coupled with external optimization tools (e.g. DAKOTA) to analyse different production scenarios, building on the on-going joint research between UFPE and Petrobras. In addition to these dynamic calculations, the static calculations will be extended to include real-time calculation of connected volumes as the well trajectory is changed, for example to guide well planning.

Exploratory Visual Analysis (6)

Existing operators will be expanded to provide new visualisation options for simulation and modeling results on desktop computers and interactive display surfaces.

Scenarios (7)

The RRM prototype will be tested jointly with researchers from the industry partners to review its suitability and user friendliness and develop future modelling scenarios and visualisation techniques that should be incorporated in the RRM workflow. Deliverables may be amended for Phase 3 and future work accordingly.

Output (8)

The RRM framework will provide output interfaces of new reservoir geometries to GereSim and the Open Source platform DAKOTA for optimisation calculations.