Advances in conformability and sequence order management in stochastic well correlation.

Jonathan Edwards and Florent Lallier and Guillaume Caumon and Cedric Carpentier. ( 2014 )
in: Proc. 34th Gocad Meeting, Nancy

Abstract

The aim of this study is to explain a method to create stratigraphic models from stochastic correlations of sedimentary units identified along wells, according to the sequence stratigraphy paradigm. The stratigraphic model contains two major information: the different units and the conformability. This is the output of our algorithm. The transgressions and regressions being observable at different scales, the sequences can be interpreted at different orders, and this is considered in our method. The incompleteness of the data, their quantity and their varying quality, added to the fact that the processes that control the geometry and the conformability of the sequences are complex and poorly known, lead to uncertainties. The correlation being an early step of the creation of a model, the uncertainties at this step may have a major impact on it. In consequence, a stochastic method is proposed to handle the stratigraphic correlation problem. The algorithm chosen is the Dynamic Time Warping. It allows to find the best correlation between two signals, by computing the cost of each pair of correlations possible, using a set of rules. The new version of the DTW can handle different orders of sequences, different levels of interpretation, by allowing to correlate several sequences of a well to a single sequence of the second well. The algorithm also takes the conformability of the horizons into account. Some secondary rules give costs to the four arrangements possible of the boundaries of a unit (top conformable or not, base conformable or not). This will allow to get the layering of the models, which is important for studies such as fluid flow simulations. Moreover, its output is a set of different possible correlations. The first rule sets that the sequences to be connected have to show corresponding changes of relative sea level (transgression or regression). The second rule, compares the dips of the horizons which are above and below the unit formed by the correlation of the two sequences. A third rule can be added to correlate several wells, by comparing the dips of the horizons of the precedent correlations and the horizons considered.

Download / Links

BibTeX Reference

@INPROCEEDINGS{EdwardsGM2014,
    author = { Edwards, Jonathan and Lallier, Florent and Caumon, Guillaume and Carpentier, Cedric },
     title = { Advances in conformability and sequence order management in stochastic well correlation. },
 booktitle = { Proc. 34th Gocad Meeting, Nancy },
      year = { 2014 },
  abstract = { The aim of this study is to explain a method to create stratigraphic models from stochastic correlations of sedimentary units identified along wells, according to the sequence stratigraphy paradigm. The stratigraphic model contains two major information: the different units and the conformability. This is the output of our algorithm. The transgressions and regressions being observable at different scales, the sequences can be interpreted at different orders, and this is considered in our method.
The incompleteness of the data, their quantity and their varying quality, added to the fact that the processes that control the geometry and the conformability of the sequences are complex and poorly known, lead to uncertainties. The correlation being an early step of the creation of a model, the uncertainties at this step may have a major impact on it. In consequence, a stochastic method is proposed to handle the stratigraphic correlation problem.
The algorithm chosen is the Dynamic Time Warping. It allows to find the best correlation between two signals, by computing the cost of each pair of correlations possible, using a set of rules. The new version of the DTW can handle different orders of sequences, different levels of interpretation, by allowing to correlate several sequences of a well to a single sequence of the second well. The algorithm also takes the conformability of the horizons into account. Some secondary rules give costs to the four arrangements possible of the boundaries of a unit (top conformable or not, base conformable or not). This will allow to get the layering of the models, which is important for studies such as fluid flow simulations. Moreover, its output is a set of different possible correlations.
The first rule sets that the sequences to be connected have to show corresponding changes of relative sea level (transgression or regression). The second rule, compares the dips of the horizons which are above and below the unit formed by the correlation of the two sequences. A third rule can be added to correlate several wells, by comparing the dips of the horizons of the precedent correlations and the horizons considered. }
}