Uncertainty management in stratigraphic well correlation and stratigraphic architectures: A training-based method

EdwardsEtAl CAGEO2017

This paper, published in Computers & Geosciences, proposes a new way to sample stratigraphic layering uncertainty by generating possible correlation lines between boreholes.
The method uses a forward stratigraphic model (e.g., DIONISOS, SEDSIM) to infer the likelihood of correlating two stratigraphic units on two different wells and to assess the probaility of a particular unit to pinch out laterally.  Additionally, multi-well stratigraphic ambiguities are managed by a global stratigraphic column, which makes it possible to automatically update reservoir grids conditioned to well data.
This constitutes a new way to correlate stratigraphic series in a reproducible and stochastic way. The proposed method could also be used to help analyze some existing correlations with regard to a particular forward stratigraphic scenario. As a new grid reflecting correlation uncertainties can be generated, this method also clears the path to assess the impact of correlation uncertainties on reservoir flow or other processes (see also Relevance of the stochastic stratigraphic well correlation approach for the study of complex carbonate settings: application to the Malampaya buildup (Offshore Palawan, Philippines)). Nonetheless, the results also illustrate the difficulty to find the optimal stratigraphic architecture event when multiple wells are present, owing to the very large combinatorial complexity of the problem. This motivates even more so the use of stochastic methods.
 
The full paper is freely available from the publisher's web site at https://authors.elsevier.com/c/1V-yCMMTPT6YB until Dec 26, 2017.
 
Results were obtained using the SCube plugin of SKUA-GOCAD. The training model was obtained with DIONISOS.