To main content

GeoScale: reservoir simulation on a geological scale

Modern methods for 3D geological modelling and reservoir characterization are leading industry to routinely build very large and detailed reservoir models; grid models of the subsurface geology currently range in size from 10 to 100 million cells and are growing. Such models pose a serious computational challenge to numerical reservoir simulators. Indeed, there is a steadily increasing gap between the level of detail seen in industrial geomodels and the capabilities of current flow simulators. This gap cannot be closed by incremental improvements to legacy simulators.

Contact person

Illustration of the multiscale idea: a fine grid is partitioned into a coarse mesh, over which one computes basis functions that account for the effect of subscale heterogeneity variations. Using these basis functions, one can then formulate a global flow problem that has much fewer degrees of freedom than the original fine-scale problem and thus is faster to solve. However, once the coarse problem is solved, one can use the subresolution of the basis functions to compute approximate fluxes on the original fine-scale grid (and at any intermediate grid resolution).

In 2003, SINTEF set an ambitious goal to try to close this gap by developing a unique and novel suit of multiscale reservoir simulation tools. Through a project portfolio consisting of  two concecutive strategic institute programs as well as a few researcher, expertise-building, and innovation projects, all funded by the Research Council of Norway, we started developing multiscale methods that would be applicable to grids of industry-standard complexity. Our work stood out from most other research on multiscale methods at the time, which focused almost exclusively on regular Cartesian grids. 

The expertise developed in the GeoScale portfolio not only was the start of our development of MRST, but also layed the foundations for our subsequent development of what is state-of-the-art in multiscale reservoir simulation.

Relevant links

Project team

  • Astrid Fossum Gulbransen

    Astrid Fossum Gulbransen

    Head of Risk Modelling DNB
  • Halvor Møll Nilsen

    Halvor Møll Nilsen

    Senior Research Scientist
  • Ingeborg Skjelkvåle Ligaarden

    Ingeborg Skjelkvåle Ligaarden

    Head of Data Science Statnett
  • Jostein Roald Natvig

    Jostein Roald Natvig

    Senior Scientific Computing Engineer SLB
  • Jørg E. Aarnes

    Jørg E. Aarnes

    Global lead - Hydrogen and CCS DnV
  • Stein Krogstad

    Stein Krogstad

    Senior Research Scientist
  • Vegard Kippe

    Vegard Kippe

    Specialist Reservoir Technology, Reservoir Simulation Equinor
  • Vera Louise Hauge

    Vera Louise Hauge

    System Consultant Eikos AS