Developments on Relative Permeability Computation in 3D Rock Images

Abstract

Abstract Digital Rock Physics (DRP) is commonly perceived as a range of technologies finalized at the calculation of properties of interest to geophysical, geological and reservoir engineering disciplines starting from 3D high resolution x-ray micro-CT images. Provided that verified, physically validated and controllable image acquisition and modeling workflows are available, petrophysical properties computed in this way can in theory be used in association to traditional difficult-to-obtain or often scarce core laboratory measurements to achieve higher insight of the reservoir and reduce the uncertainty in both static and dynamic models. DRP's potential for the industry is also expressed by the possibility of achieving better understanding of recovery mechanisms by probing fluid distributions at the pore scale, as is being nowadays investigated thanks to the utilization of miniaturized flow cells in micro-CT set ups or synchrotron facilities. This could be key for optimization of EOR processes. From the point of view of operating companies, most efforts are still deployed in the R&D Lab with deployment of the technology in operational context still at its infancy and hindered by the issue of representativity of the microscopic imaged or computed scales (or image resolution-scale trade off), by the difficulty one has to unambiguously inform pore scale models with a sufficiently limited but precise set of physico/chemical information and by lack of robust validation procedures. In this dynamic and improving context, we update on our R&D efforts to evaluate and test one particular technology for the simulation of multi-phase flow in digital rocks, the Volume of Fluid method embedded in Paris simulator: the objective is to verify the potential interest in the medium and long term, knowing that other simulation technologies, being simpler to apply, are more mature for utilization in today's context. In this work Paris simulator is utilized to compute single and multi-phase flow properties on TOTAL's supercomputers for both a sandstone outcrop sample from Scotland and two carbonate rocks from UAE. First the simulator is tested for simpler permeability computations and benchmarked against a Lattice-Boltzmann solver; then the code is used for two-phase flow and for relative permeability computations. It is concluded that the particular simulator used in this work, still under development, can be used for different rock types and is particularly efficient in HPC environment: it has therefore strong potential. We conclude describing the future steps of development that will be needed to make the simulator applicable for digital petrophysics.