Graph-based fault network uncertainty assessment (Athabasca basin, Saskatchewan, Canada)

Paul Marchal and Kelsey McKee and Guillaume Caumon. ( 2019 )
in: 2019 Ring Meeting, ASGA

Abstract

Regarding geomodeling, faults are complex geological objects. Structural uncertainty assessment in fault modeling is a challenging issue, due both to interpretation bias and lack of information. It is frequently hard to understand fault systems chronology and behavior, although this information is relevant. For example, in mining, faults can be fluid circulation paths, and thus localize ore concentration. This study is based on an actual data set provided by Orano Canada and aims to afford better understanding of faults in the case of an unconformity associated uranium deposit in the Athabasca basin, Canada. These data are mainly clues of faults and alteration zones, which are collected from drilling analysis and imported as points or segments in SKUA-Gocad. Then, using FaultMod2 module, these observations forms the nodes of a graph in which edges correspond to a probability for two observations to belong to the same fault. This probability is computed based on geological rules which are implemented in the code. Fundamentals rules, such as one based on distance between observations, or even their orientation are already employed. The implementation of new rules is needed to improve geological consistency. In the first place, a Well Zone Rule reduces the probability of building faults in areas where no fault evidence have been found, according to a given well zone. Indeed, no fault evidence in a well might also be considered as a clue too. Then, a Horizon Displacement Rule computes a probability based on the elevation difference, or Dz, between two blocks of a horizon displaced by faults. In our case, it is based on the unconformity elevation data, but it could be generalized on other displaced horizon data. The higher the z is, the greater the probability of fault presence. This work proposes to implement new rules, which are mainly adapted to the data provided, but could be relevant for other cases, due to the fact that they provide more geological reliability to the construction of faults. First results on the Athabasca dataset show that the two new rules tend to generate likelier fault systems.

Download / Links

BibTeX Reference

@INPROCEEDINGS{MarchalRM2019,
    author = { Marchal, Paul and McKee, Kelsey and Caumon, Guillaume },
     title = { Graph-based fault network uncertainty assessment (Athabasca basin, Saskatchewan, Canada) },
 booktitle = { 2019 Ring Meeting },
      year = { 2019 },
 publisher = { ASGA },
  abstract = { Regarding geomodeling, faults are complex geological objects. Structural uncertainty assessment in fault modeling is a challenging issue, due both to interpretation bias and lack of information. It is frequently hard to understand fault systems chronology and behavior, although this information is relevant. For example, in mining, faults can be fluid circulation paths, and thus localize ore concentration. This study is based on an actual data set provided by Orano Canada and aims to afford better understanding of faults in the case of an unconformity associated uranium deposit in the Athabasca basin, Canada. These data are mainly clues of faults and alteration zones, which are collected from drilling analysis and imported as points or segments in SKUA-Gocad. Then, using FaultMod2 module, these observations forms the nodes of a graph in which edges correspond to a probability for two observations to belong to the same fault. This probability is computed based on geological rules which are implemented in the code. Fundamentals rules, such as one based on distance between observations, or even their orientation are already employed. The implementation of new rules is needed to improve geological consistency. In the first place, a Well Zone Rule reduces the probability of building faults in areas where no fault evidence have been found, according to a given well zone. Indeed, no fault evidence in a well might also be considered as a clue too. Then, a Horizon Displacement Rule computes a probability based on the elevation difference, or Dz, between two blocks of a horizon displaced by faults. In our case, it is based on the unconformity elevation data, but it could be generalized on other displaced horizon data. The higher the z is, the greater the probability of fault presence. This work proposes to implement new rules, which are mainly adapted to the data provided, but could be relevant for other cases, due to the fact that they provide more geological reliability to the construction of faults. First results on the Athabasca dataset show that the two new rules tend to generate likelier fault systems. }
}