Stochastic fault network modeling inside delimited volumetric regions

in: Proc. 32nd Gocad Meeting

Abstract

Due to the sparse or low-resolution nature of the available data on a given reservoir, significant uncertainties about the precise location and shape of faults and about the connection between these faults exist. The seismic surveys are highly valuable sources of structural information. Even if the 3D, or 2D, seismic images are blurred in the vicinity of faults caused by the diffraction of the waves, unfaulted zones can be identified by the continuity of the seismic reflectors. We propose to identify an envelope delimiting a fault occurrence exclusion zone rather than deterministically picking fault sticks from seismic data. This envelope is used to constrain the spatial position of faults during the stochastic simulation of a fault network. To meet this objective, a specific method of data conditioning has been implemented: a post-conditioning process optimizes the geometrical parameters (position of center, dip, strike) of the faults by minimizing a misfit function which quantifies the degree of overlap between the simulated fault and the envelope.

Download / Links

    BibTeX Reference

    @INPROCEEDINGS{JulioGM2012,
        author = { Julio, Charline and Caumon, Guillaume and Cherpeau, Nicolas },
         title = { Stochastic fault network modeling inside delimited volumetric regions },
     booktitle = { Proc. 32nd Gocad Meeting },
          year = { 2012 },
      abstract = { Due to the sparse or low-resolution nature of the available data on a given reservoir, significant uncertainties about the precise location and shape of faults and about the connection between these faults exist. The seismic surveys are highly valuable sources of structural information. Even if the 3D, or 2D, seismic images are blurred in the vicinity of faults caused by the diffraction of the waves, unfaulted zones can be identified by the continuity of the seismic reflectors.
    We propose to identify an envelope delimiting a fault occurrence exclusion zone rather than deterministically picking fault sticks from seismic data. This envelope is used to constrain the spatial position of faults during the stochastic simulation of a fault network. To meet this objective, a specific method of data conditioning has been implemented: a post-conditioning process optimizes the geometrical parameters (position of center, dip, strike) of the faults by minimizing a misfit function which quantifies the degree of overlap between the simulated fault and the envelope. }
    }