Comparison of reservoir simulators and grids.

Deborah Siffert and Guillaume Caumon and Romain Merland. ( 2011 )
in: Proc. 31st Gocad Meeting, Nancy

Abstract

For a given static reservoir model and flow parameters, the forecast of reservoir exploitation may be sensitive to the reservoir simulator used and more significantly, to gridding process on unstructured grids. This project consists in testing the influence of these parameters, and in comparing the results obtained in homogeneous and heterogeneous reservoir. Two simulators are compared, Eclipse and GPRS, using different reservoir models. More particularly, dynamic simulations are run on homogeneous isotropic Cartesian cases in 1D, 2D and 3D, and on the heterogeneous SPE 10th comparative solution project in 2D. The two simulators give very close results, so the following simulations run under GPRS could be considered as non biased. Gridding effects are tested under GPRS on unstructured grids through two geometrical parameters: local refinement and grid melting. These parameters are taken into account in the gridding process using Centroidal Voronoi Tessellations (CVT) algorithms. Static parameter as permeability and dynamic flow parameters as velocity and vorticity, are used to constrain the gridding process.

Download / Links

BibTeX Reference

@INPROCEEDINGS{SiffertGM2011,
    author = { Siffert, Deborah and Caumon, Guillaume and Merland, Romain },
     title = { Comparison of reservoir simulators and grids. },
 booktitle = { Proc. 31st Gocad Meeting, Nancy },
      year = { 2011 },
  abstract = { For a given static reservoir model and flow parameters, the forecast of reservoir exploitation may be sensitive to the reservoir simulator used and more significantly, to gridding process on unstructured grids. This project consists in testing the influence of these parameters, and in comparing the results obtained in homogeneous and heterogeneous reservoir.
Two simulators are compared, Eclipse and GPRS, using different reservoir models. More particularly, dynamic simulations are run on homogeneous isotropic Cartesian cases in 1D, 2D and 3D, and on the heterogeneous SPE 10th comparative solution project in 2D. The two simulators give very close results, so the following simulations run under GPRS could be considered as non biased.
Gridding effects are tested under GPRS on unstructured grids through two geometrical parameters: local refinement and grid melting. These parameters are taken into account in the gridding process using Centroidal Voronoi Tessellations (CVT) algorithms. Static parameter as permeability and dynamic flow parameters as velocity and vorticity, are used to constrain the gridding process. }
}