A Distance-based approach to automatically identify similarities and differences between faults association models.

@inproceedings{oumar_distance-based_2022,
 abstract = { The interpretation of faults from geological or geophysical data is often prone to structural uncertainties due to the limited quality of seismic data, limited outcrop observations and subsurface data and subjectivity of interpretations.This paper focuses on uncertainties resulting from the association of sparse faults observations. For this, Godefroy et al. (2021) proposed a graphbased simulation method to assess these uncertainties by generating several realizations of fault associations from incomplete local fault observations. Each realization is a way to regroup fault data into a variable number of faults, which can then be used to generate geometric scenarios in geomodeling software. In this paper, we build on this approach to explore and to organize the space of realizations to allow for orienting the simulation algorithm towards either local neighborhood search of global exploration. For this, we store association scenarios in a tree structure as previously proposed by Cherpeau, Caumon, and Le´vy (2010); Taty Moukati et al. (2021), and we exploit this structure to define two distances between the fault association realizations. The unordered distance depends only on how the fault data are associated together, whereas the ordered distance also depends on the simulation order, as each order may correspond to a different geological history and fault network connectivity. To study the link between these two static distances and the hydrodynamic behavior of subsurface reservoirs, we perform a 2D two-phase flow experiment on geometric realizations of the faults. In our tests, we observe a moderate but clear trend between static and hydrodynamic distances, both for the unordered distance and the ordered distance. This opens interesting perspectives for the resolution of inverse flow problems in structurally controlled or fractured aquifers and reservoirs. },
 author = { Oumar, M’kemeltou AND Caumon, Guillaume AND Taty-Moukati, Fabrice },
 booktitle = { 2022 {RING} {Meeting} },
 pages = { 13 },
 publisher = { ASGA },
 title = { A Distance-based approach to automatically identify similarities and differences between faults association models. },
 year = { 2022 }
}