Despite the increasing production and use of renewable power from sources such as biomass, wind, and photovoltaics amongst others, all global scenarios also point to an increasing use of energy from fossil fuels. This fact drives the exploration and production of new deposits of oil and gas.
As is often stated, the "easy oil", that is produced from shallow fields and areas of simple geology has already been discovered and so geoscientists are left with the difficult task of finding accumulations of oil and gas in deepwater horizons and complex geological areas. This is challenging because the subsurface imaging techniques that have been widely used to help exploration in the past, such as Kirchhoff migration, do not take the geological complexities of the new prospects into account in their mathematical formulation.
The solution to this problem is the use of other techniques such as RTM (Reverse Time Migration). To carry work forward in this direction, the tendency is to incorporate into the equations as much of the physical reality as possible; that’s to say, to make a mathematical model that represents the geological environment and the seismic wave propagation in order to take into account characteristics such as viscosity, anisotropy and high velocity and density contrasts to name but a few. Such an approach demands high computing power to handle all these variables.
Fortunately, the availability of new solutions in the computing industry such as clustering, the use of multi- and many-core CPU´s, graphics processing units, reconfigurable logic and a number of other advancements, has provided a significant increase in available computational power. The challenge then becomes to adapt the former algorithms to these new architectures in order to extract maximum performance from them. This is the essence of the HPC4E project’s work regarding the oil and gas industry, which has as its ultimate goal the optimisation of the RTM computer code core and its further adaptation to a full extrapolation scheme, taking into account the newest computer architectures available.
Ricardo de Bragança
Geophysicist - HPC4E researcher - PETROBRAS Research and Development Center
This article also appeared in the LinkedIn Group. Join us!
Other LinkedIn articles: