Geo-models production forecast with a non linear n-dimensional interpolator

in: XI European Conference on the Mathematics of Oil Proceedings, European Association of Geoscientist and Engineers

Abstract

One of the most challenging problem in reservoir modelling is to handle the uncertainty on the reservoir flow performance. Common uncertainty analysis approaches use a large number of equiprobable model,however among them only a limited number are considered for detailed flow simulation. This paper proposes an approach based on a n-dimensional response surface, to forecast the reservoir flow performance on non-simulated model. This approach is based on the Discrete Smooth Interpolation Algorithm. This algorithm designed to work in an n-dimensional space is particularly convenient because uncertainty on the data and contradictory data can be taken into account. The approach has been tested on realistic model and results are consistent and some time even better than classical techniques.

Download / Links

    BibTeX Reference

    @INPROCEEDINGS{Fetel2004ECMOR,
        author = { Fetel, Emmanuel and Mallet, Jean-Laurent and Voillemont, Jean-Charles },
         title = { Geo-models production forecast with a non linear n-dimensional interpolator },
         month = { "sep" },
     booktitle = { XI European Conference on the Mathematics of Oil Proceedings },
          year = { 2004 },
    organization = { European Association of Geoscientist and Engineers },
       address = { Cannes, France },
      abstract = { One of the most challenging problem in reservoir modelling is to handle the uncertainty on the reservoir flow performance. Common uncertainty analysis approaches use a large number of equiprobable model,however among them only a limited number are considered for detailed flow simulation. This paper proposes an approach based on a n-dimensional response surface, to forecast the reservoir flow performance on non-simulated model. This approach is based on the Discrete Smooth Interpolation Algorithm. This algorithm designed to work in an n-dimensional space is particularly convenient because uncertainty on the data and contradictory data can be taken into account. The approach has been tested on realistic model and results are consistent and some time even better than classical techniques. }
    }