Conditioning channel backward migration modeling to seismic data

in: 78th EAGE Conference & Exhibition 2016 Proceedings, Vienna, Austria

Abstract

Visualizing and interpreting sedimentary geobodies on stratal slices has become a common task in the understanding of reservoir heterogeneity based on reflection seismic data. Large-scale structures such as point bars and abandoned meanders are observable but small-scale structures may require extrapolation interpretation between the noticeable parts. Despite of signifi cant processes over recent years on the visualization of these structures, their quantitative interpretation still essentially relies either on deterministic image analysis or manual interpretation techniques or on geostatistical models constrained by secondary data. This work proposes a hybrid method combining object-based approach and pseudo-genetic concepts in order to facilitate interpretation tasks. The main idea is to generate 3D models of poorly imaged point bar structures around channel objects by retro- migrating an initial channel structure interpreted from the seismic. The main originality of the approach is the inclusion of picked abandoned meanders and point bars in the process. Sedimentary objects are modeled with Non-Uniform Rational B-Splines (NURBS) allowing easy connection and deformation of object models. These models of sedimentary structures are grid free due to their construction with parametric surfaces. We demonstrate this technique on satellite pictures of alluvial deposits presenting oxbow lakes.

Download / Links

BibTeX Reference

@INPROCEEDINGS{Parquer2016,
    author = { Parquer, Marion and Collon, Pauline and Caumon, Guillaume },
     title = { Conditioning channel backward migration modeling to seismic data },
     month = { "jun" },
 booktitle = { 78th EAGE Conference & Exhibition 2016 Proceedings },
      year = { 2016 },
  location = { Vienna, Austria },
       doi = { 10.3997/2214-4609.201601343 },
  abstract = { Visualizing and interpreting sedimentary geobodies on stratal slices has become a common task in the understanding of reservoir heterogeneity based on reflection seismic data. Large-scale structures such as point bars and abandoned meanders are observable but small-scale structures may require extrapolation interpretation between the noticeable parts. Despite of significant processes over recent years on the visualization of these structures, their quantitative interpretation still essentially relies either on deterministic image analysis or manual interpretation techniques or on geostatistical models constrained by secondary data. This work proposes a hybrid method combining object-based approach and pseudo-genetic concepts in order to facilitate interpretation tasks. The main idea is to generate 3D models of poorly imaged point bar structures around channel objects by retro- migrating an initial channel structure interpreted from the seismic. The main originality of the approach is the inclusion of picked abandoned meanders and point bars in the process. Sedimentary objects are modeled with Non-Uniform Rational B-Splines (NURBS) allowing easy connection and deformation of object models. These models of sedimentary structures are grid free due to their construction with parametric surfaces. We demonstrate this technique on satellite pictures of alluvial deposits presenting oxbow lakes. }
}