Datapages, Inc.Print this page

Natural Fracture Networks Enhancing Unconventional Reservoirs' Producibility: Mapping & Predicting

Abul Khair, Hani; Cooke, Dennis; Hand, Martin

The success of an unconventional reservoir is strongly reliant on the density and orientation of the pre-existing natural fracture networks. We mapped and predicted natural fractures using 3-D seismic data and geomechanical simulation of tight gas reservoirs in the Cooper Basin/South Australia and compared resulted fractures to image logs, seismic data and well data.

We extracted positive curvature attribute to generate a workflow for modelling subsurface fractures. The workflow included applying structural smoothing on the seismic cube to eliminate acquisition artefacts, then excluding low curvature values that don’t reflect any structural feature. A validation procedure was applied using image logs, well data, and quality seismic, and a high correlation was found between the curvature and fractures as seen on image logs.

Another technique used in this study was to integrate geological and geophysical data extracted from fault and horizon seismic interpretation with geomechanical analyses of stress, strain, and displacements that associated the structural development of the basin. Finite element method (FEM) and Fault elastic dislocation (boundary element method (BEM)) are two ways used to predict fractures generated during the tectonic events of basins. FEM provides a physically-based solution for fractures that takes into account basin evolution, horizon geometry, heterogeneous rock properties and stresses. While the BEM considers the effect of fault displacement on generating stress and strain at every node of the dislocated elastic horizon around the fault. BEM succeeds in simulating strain distribution especially close to main faults, but it doesn't consider rock heterogeneity or the effect of intra-seismic relaxation on fracture generation. Also, with the lack of far field stress data, it is not possible to predict fracture generation away from the major faults.

The validation procedure was applied on fractures predicted from FEM and BEM, and a good correlation was found between the predicted fracture network and the image logs fractures next to the major faults in both methods as fractures in these areas were mostly generated due to stresses exerted during fault displacement. FEM succeeded in predicting fractures close and away from major faults with higher accuracy, while BEM didn't map fractures away from faults. Thus, both FEM and enhanced most positive curvature attributes can be used successfully to model subsurface fracture networks.


AAPG Search and Discovery Article #90163©2013AAPG 2013 Annual Convention and Exhibition, Pittsburgh, Pennsylvania, May 19-22, 2013