| dc.description.abstract | Fractures in oil and gas reservoir rocks play an important role in the storage and transport
of these fluids. The fluid flow is controlled by these structures, which are presented in different scales, ranging from networks of connected pores in the rock matrix to kilometric fracture networks. Thus, to understand the behavior of the reservoir and estimate parameters for the extraction of such resources, the fracture networks are studied by geologists and engineers through numerical representations (deterministic and stochastic) based on seismic acquisitions and profiles of wells. However, studies of reservoirs located at great depths suffer from the limitation of the scale of the former and the sparsity of data from the latter, and thus, analogous outcrops are used to fill this gap. Fracture information on outcrops can be acquired in the field or from interpretations in digital representations obtained through Light Detection and Ranging or photogrammetric surveys with the aid of unmanned aerial vehicles. The most common method consists in delineating fracture traces in orthophotos obtained from photogrammetric processing, but these products do not allow the acquisition of three-dimensional information such as fracture dip, for example. Thus, the hypothesis is that three-dimensional models enable a more detailed description of fracture networks, with measurements that can be made on a
computer screen and in immersive virtual reality, resulting in more appropriate input data sets for mathematical models. That said, this work aimed to propose a stochastic approach for the development of fracture networks based on immersive digital models of outcrops. To this end, 2D and 3D discrete fracture network models were generated based on interpretations in digital and virtual immersive models of outcrops. The models were validated through statistical analyzes and comparisons to works with studies on the same outcrops investigated in this research. In particular, in terms of fracture intensity per volume (P32) the results of the 3D stochastic model had an average value of 0.15m−1 and a standard deviation of 0.08m−1, which confirmed the established hypothesis, since the values of P32 are close to values obtained from 2D interpretations, however the dip information in these models are estimated. Finally, this research highlights the potential of using immersive reality to support geological interpretations in rocky outcrops analogous to fractured reservoirs, to provide geologists and engineers with more adequate parameters for flow simulations that aim to define reservoir production estimates and
exploration criteria. | en |
