Statistical analysis and stochastic simulation of {Fracture} {Networks}.

Francois Bonneau and Radu S Stoica. ( 2022 )
in: 2022 {RING} {Meeting}, pages 11, ASGA

Abstract

Fracture networks (FN) are systems of complex mechanical discontinuities in rocks. Such networks dramatically impact fluid flow acting as a drain or a barrier. This work proposes to build a stochastic mathematical model and to use it for fracture characterization. For a few decades, the mathematical framework of marked-point process has been successfully used to study fracture networks. In this work, we will also use this approach to approximate FN with a collection of marked-point standing for straight-line segment. The “point” locates the segment barycenter and holds two real valued “marks”: (1) the length and (2) strike azimuth of the segment. Geologists usually use a characterization workflow integrating first order metrics of the marked point process theory to describe the number and the geometry of fractures, i.e. their density and their mark distributions. Recently, second-order characteristics have been used to characterize fracture network inner correlation and spatial variability. Investigating stochastic models that are known to produce realizations present similar first and second order characteristics may be the key of a final understanding of FN. In this work, we built a mathematical model to produce stochastic realizations of marked point process that considers and reproduces such observations. Also, we investigated the optimization of model parameters in the light of a natural FN pattern. This work may open the path to a thinner classification of FN and to predictive stochastic simulations.

Download / Links

BibTeX Reference

@inproceedings{bonneau_statistical_2022,
 abstract = { Fracture networks (FN) are systems of complex mechanical discontinuities in rocks. Such networks dramatically impact fluid flow acting as a drain or a barrier. This work proposes to build a stochastic mathematical model and to use it for fracture characterization. For a few decades, the mathematical framework of marked-point process has been successfully used to study fracture networks. In this work, we will also use this approach to approximate FN with a collection of marked-point standing for straight-line segment. The “point” locates the segment barycenter and holds two real valued “marks”: (1) the length and (2) strike azimuth of the segment. Geologists usually use a characterization workflow integrating first order metrics of the marked point process theory to describe the number and the geometry of fractures, i.e. their density and their mark distributions. Recently, second-order characteristics have been used to characterize fracture network inner correlation and spatial variability. Investigating stochastic models that are known to produce realizations present similar first and second order characteristics may be the key of a final understanding of FN. In this work, we built a mathematical model to produce stochastic realizations of marked point process that considers and reproduces such observations. Also, we investigated the optimization of model parameters in the light of a natural FN pattern. This work may open the path to a thinner classification of FN and to predictive stochastic simulations. },
 author = { Bonneau, Francois AND Stoica, Radu S },
 booktitle = { 2022 {RING} {Meeting} },
 pages = { 11 },
 publisher = { ASGA },
 title = { Statistical analysis and stochastic simulation of {Fracture} {Networks}. },
 year = { 2022 }
}