Smarter Natural Resources - Physically-driven Analytics


Computational Geosciences has emerged as the key technology to help us understand the most challenging and complex of systems in energy, environment, and fundamental research. The promise of Computational Geosciences is to deliver accurate models of complex systems which can be used to develop and advance fundamental understanding of those systems and to manage and optimize resources and production which is dependent on that understanding. Computational Geosciences will play a critical role in the development of many 21st century in Natural Resources industries and, properly deployed, has the potential to seed significant economic growth in high tech and high value industry in the next few years.

IBM is a recognized leader in the field of deep computing and is actively engaged in research and development to deliver Petascale and then Exascale systems, software and solutions in the next decade. IBM’s Blue Gene, Power, and iDataPlex systems are successfully demonstrating the potential of petascale application solutions for several industries. IBM delivers full high performance solutions including hardware, system software, compilers, data management, data center design, middleware, and development tools. IBM Reserch is actively engaged in all aspects of deep computing research including fundamental algorithms, systems research, applications research, and, most strongly, in applications integration and overall solution development.

Physically-based analytics such as modeling, simulation, inversion and optimization offer tremendous opportunities in several key areas: improving our understanding of the earth’s systems, addressing geoscience grand challenges, and providing decision-support tools for energy policy-makers and the industry. With today’s increased demand for energy and constrained natural resources, there is a need to find new resources economically, and efficiently manage known resources. Throughout the world, some resources that are remarkably constrained include: oil, water, copper, and several other minerals. Typical problems in geosciences require advanced numerical algorithms running on high-performance computers. The increasing importance of computation as a powerful tool for prediction and decision-making in the geosciences is driven by advances in three areas: the rapid expansion of our ability to instrument and observe the earth (Smarter Planet); sustained improvements in computational models and solution methods for complex geoscience systems; and the unrelenting growth in computing power. For example, large-scale simulations of the dynamics of surface and subsurface flows are routinely carried out with increasing fidelity. A central challenge in computational geosciences is the systematic assimilation of observational data into large-scale simulations to identify and address model uncertainties. This requires leadership in the development of inverse methods for data assimilation and optimization to match observations. Tackling these problems requires interdisciplinary expertise in geosciences, applied mathematics, computer science, and other areas. We use these disciplines to perform multiple modeling of the earth properties utilizing massive computational resources. This includes basin modeling, geochemistry, gravity, electromagnetics, and time-lapse seismic. We expect that with this approach we should increase the resolution in the earth imaging process and thus increase the rate of finding hydrocarbons, reduce exploration risks, and increase the rate of reserves replacement.

Impact and Benefits:

Brazil is very well placed to take advantage of the opportunity offered by Computational Geosciences. This region is particularly rich in natural resources, and the optimum management of those resources presents an extremely complex computational problem. Computational solutions can benefit all areas of this management problem including the basic design of the engineering solutions to be deployed, understanding of the geological processes which enable the exploration and production of the natural resources, management and optimization of the production operation processes, understanding and management of the environmental impact, and general economic modeling and management

Example project: Parallel Basin Modeling

In this project, we are focused on two challenges in creating simulations of the time evolution of sedimentary basins. The evolution of a basin takes place over millions of years, involves the deposition and erosion of sediments, rock layers undergoing large deformations, and the flow of fluids and gases through porous rock and networks of faults and fractures. The first challenge is the size and complexity of the geometry, and the corresponding demands this places on the computational platform. This requires efficient algorithms for reading, partitioning and manipulating large meshes and solving differential equations for deformation, pressure and temperature distributions on those meshes. The second challenge is to accurately model phenomena taking place on widely different time and spatial scales. This requires conservative techniques like control volume finite elements (CVFE), and methods like discontinuous Galerkin for modeling discontinuities (eg. Faults and fractures).

parallel basin modeling

We collaborated with Eni to develop new large-scale, parallel, numerical simulator using unstructured tetrahedral meshes which provides unprecedented capabilities for the simulation of complex geological structures. New algorithms for dealing with 3D compaction for unstructured mesh were developed using XML technology to integrate related systems. This simulator models the interaction of complex geological processes such as geological basin formation, compaction, and fault effects on pressure and temperature fields. Better numerical models are expected to improve significantly the risk assessment of new prospects in new frontiers as well as in mature basins requiring rejuvenation.

Selected Publications:

Mello, U. T., Rodrigues, J R, Rossa, A, 2009, A Control-Volume Finite-Element Method for Three-Dimensional Multiphase Basin Modeling, Marine and Petroleum Geology, 26:504-518.

Mello, U. T. and Cavalcanti, P. R., 2003, A Topologically Based Framework for 3-D Basin Modeling. In Duppenbecker and Marzi, eds., Multidimensional basin modeling, AAPG/Datapages Discovery Series, No 7, p.255-269.

Anderson, R. N., Guerin, G., Mello, U. T., He, W., 1998, 4-D Seismic Reservoir Simulation in a South Timbalier 295 Turbidite Reservoir. The Leading Edge, 17(10):1416-1418.

Mello, U. T. and Henderson, M. E., 1997, Techniques for including large deformation associated with salt and non-vertical fault motion in basin modeling, Marine and Petroleum Geology, v.14(5), pp. 551-564.

Mello, U. T. and Karner, G. D. -1996- Development of sediment overpressure and its effect on thermal maturation: Application to the Gulf of Mexico basin. American Association of Petroleum Geologists Bulletin, v.80(9):1367-1396.

Mello, U. T.; Karner, G. D. and Anderson, R. N. -1995- The role of salt rock in modifying the maturation history of sedimentary basins. Marine and Petroleum Geology, v.12(7), p. 697-716.

Mello, U. T.; Anderson, R. N. and Karner, G. D. -1994- Salt restrains maturation in subsalt plays. Oil and Gas Journal. Jan. 31, p. 101-107.

Mello, U. T.; Karner, G. D. and Anderson, R. N. -1994- A physical explanation for the positioning of the depth to the top of overpressure in shale-dominated sequences in the Gulf Coast basin, United States. Journal of Geophysical Research, v.99, p. 2775-2789.

Mello, U. T. -1989 - Tectonic controls in the stratigraphy of Potiguar basin: An integration of geodynamical models (In Portuguese). Boletim de Geociências da Petrobrás v.3(4):347-364.

Mello, U. T. and Bender, A. A.-1988- On Isostasy at the Equatorial Margin of Brazil. Revista Brasileira de Geociências, v.18(3): 237-246.