Corrélation stratigraphique stochastique de puits

Florent Lallier. ( 2012 )
Université de Lorraine

Abstract

Stratigraphic correlation consists in linking boundaries of correlative units between wells or outcrops over a given study area, and is therefore one of the first steps of the characterization of the subsurface geometry, supporting geostatistical modeling of static reservoir properties. However, this early step is subject to uncertainties, since stratigraphic well correlation is constrained only by sparse observations (wells and outcrops), low resolution information coming from geophysics, regional knowledge and geological concepts. The Dynamic Time Warping (DTW) algorithm serves as a base for the development of a generic method stochastically performing stratigraphic correlation between units identified on available wells. The proposed method relies on the definition of correlation rules that are applied using the available data and some regional knowledge, according to the way stratigraphic units are defined. An application to the Cretaceous southern Provence carbonate basin has been performed using correlation rules based on paleo-angles and the theoretical architecture of the depositional environment. The computation of vertical proportions of facies on numerous models generated from the stochastic correlations of sequence stratigraphic units indicates that the uncertainties on the stratigraphic correlation impact the compartmentalization of the modeled reservoirs. The impact of stratigraphic correlation uncertainties on fluid flow behavior is assessed through the example of the Malampaya diagenetic carbonate reservoir. Diagenetic units are correlated on the basis of their wireline log signature and diagenetic types. Different models are generated from the stochastic well correlations, and the corresponding water saturation profiles are computed. They show different displacement patterns, indicating a stratigraphic control of the dynamic property, which contrasts with the synthetic seismic model constructed from the corresponding geomodel. Magnetostratigraphic correlation is another way to study sedimentary basin deposition and deformation history. Adapting the DTW algorithm to magnetostratigraphic data, we generate dating models of Himalayan deposits, for which conflicting interpretations are proposed in the literature. This allows managing the associated accumulation rates uncertainties.

Download / Links

BibTeX Reference

@PHDTHESIS{,
    author = { Lallier, Florent },
     title = { Corrélation stratigraphique stochastique de puits },
      year = { 2012 },
    school = { Université de Lorraine },
       url = { http://ring.gocad.org/ring_dl/public/publications/2012PhdLallier.pdf },
  abstract = { Stratigraphic correlation consists in linking boundaries of correlative units between wells or outcrops over a given study area, and is therefore one of the first steps of the characterization of the subsurface geometry, supporting geostatistical modeling of static reservoir properties. However, this early step is subject to uncertainties, since stratigraphic well correlation is constrained only by sparse observations (wells and outcrops), low resolution information coming from geophysics, regional knowledge and geological concepts. The Dynamic Time Warping (DTW) algorithm serves as a base for the development of a generic method stochastically performing stratigraphic correlation between units identified on available wells. The proposed method relies on the definition of correlation rules that are applied using the available data and some regional knowledge, according to the way stratigraphic units are defined. An application to the Cretaceous southern Provence carbonate basin has been performed using correlation rules based on paleo-angles and the theoretical architecture of the depositional environment. The computation of vertical proportions of facies on numerous models generated from the stochastic correlations of sequence stratigraphic units indicates that the uncertainties on the stratigraphic correlation impact the compartmentalization of the modeled reservoirs. The impact of stratigraphic correlation uncertainties on fluid flow behavior is assessed through the example of the Malampaya diagenetic carbonate reservoir. Diagenetic units are correlated on the basis of their wireline log signature and diagenetic types. Different models are generated from the stochastic well correlations, and the corresponding water saturation profiles are computed. They show different displacement patterns, indicating a stratigraphic control of the dynamic property, which contrasts with the synthetic seismic model constructed from the corresponding geomodel. Magnetostratigraphic correlation is another way to study sedimentary basin deposition and deformation history. Adapting the DTW algorithm to magnetostratigraphic data, we generate dating models of Himalayan deposits, for which conflicting interpretations are proposed in the literature. This allows managing the associated accumulation rates uncertainties. }
}