Stochastic structural modeling in sparse data situations.
in: Proc. 31st Gocad Meeting, Nancy
Abstract
Resource estimation is a key step for decision-making in natural resource management because it determines the economic viability of a deposit with regard to exploration and production costs. Uncertainties on fault network layout may lead to risky reserve estimation in compartmentalized resevoirs, especially when few data is available. In such context, this paper suggests to generate stochastic structural models using prior information such as fault orientation, size-displacement relations and interpretations of both faults and stratigraphic horizons on particular cross-sections.
Interpreted fault data on 2D seismic lines provide significant information to constrain fault simulation and reduce uncertainty space. Indeed, the fault slip is generally assumed to be related to the fault length. Consequently a global fault size range can be inferred from observed displacements and is an input parameter for the fault simulation algorithm. During the simulation of a fault object, an initial fault trace is assigned to the fault, then the algorithm computes a clustering probability for unassigned fault traces that depends on their distance to the fault center location and to the initial fault trace, and on the size-displacement law.
For each simulated fault network, stratigraphic modeling is performed to honor interpreted horizons using an implicit approach. Geometrical uncertainty on stratigraphic horizons can then be simulated by adding a correlated random noise to the stratigraphic scalar field. Implicit modeling provides a convenient framework for such perturbations as horizon-to-fault contacts are not explicitly represented, hence are consistently maintained.
The method is applied to the Nancy field case where fault networks and stratigraphy are generated from interpreted 2D seismic lines. Generated models show the impact of the fault network and the stratigraphy on the reserve estimation uncertainties.
Download / Links
BibTeX Reference
@inproceedings{Cherpeau1GM2011,
abstract = { Resource estimation is a key step for decision-making in natural resource management because it determines the economic viability of a deposit with regard to exploration and production costs. Uncertainties on fault network layout may lead to risky reserve estimation in compartmentalized resevoirs, especially when few data is available. In such context, this paper suggests to generate stochastic structural models using prior information such as fault orientation, size-displacement relations and interpretations of both faults and stratigraphic horizons on particular cross-sections.
Interpreted fault data on 2D seismic lines provide significant information to constrain fault simulation and reduce uncertainty space. Indeed, the fault slip is generally assumed to be related to the fault length. Consequently a global fault size range can be inferred from observed displacements and is an input parameter for the fault simulation algorithm. During the simulation of a fault object, an initial fault trace is assigned to the fault, then the algorithm computes a clustering probability for unassigned fault traces that depends on their distance to the fault center location and to the initial fault trace, and on the size-displacement law.
For each simulated fault network, stratigraphic modeling is performed to honor interpreted horizons using an implicit approach. Geometrical uncertainty on stratigraphic horizons can then be simulated by adding a correlated random noise to the stratigraphic scalar field. Implicit modeling provides a convenient framework for such perturbations as horizon-to-fault contacts are not explicitly represented, hence are consistently maintained.
The method is applied to the Nancy field case where fault networks and stratigraphy are generated from interpreted 2D seismic lines. Generated models show the impact of the fault network and the stratigraphy on the reserve estimation uncertainties. },
author = { Cherpeau, Nicolas AND Caumon, Guillaume AND Levy, Bruno },
booktitle = { Proc. 31st Gocad Meeting, Nancy },
title = { Stochastic structural modeling in sparse data situations. },
year = { 2011 }
}