2019 AAPG Annual Convention and Exhibition:

Datapages, Inc.Print this page

Modeling Natural Fractures in Reservoirs by Incorporating Structural History

Abstract

We present a workflow, adaptable to a broad range of geological and tectonic settings, which uses the geological and deformation history to predict fracture orientations and intensities. The method incorporates structural balancing and validation to ensure data-model consistency and uses geological proxies derived from deformation modelling to calculate fracture properties. The paucity of measured subsurface fracture data leads to difficulty in extrapolation of fracture models to undrilled areas of geological models. Modelling the mechanisms that result in natural fracture formation to predict fracture orientations and intensities is shown here to accurately reproduce measured fractures allowing confident calculation of orientations and intensities of natural fractures in areas with no data, regardless of structural position in a model. Application of a stress field to predicted fractures, and calculation of shear and normal stresses allows derivation of dilation and slip tendency, among other relationships. This can aid in exploration and production, including informing the fractures’ effect on fluid flow in the subsurface and delineating areas of likely failure based on an increase in pore pressure. Two case studies, from conventional and unconventional production settings, are presented to demonstrate the applicability and capability of the workflow. In both cases a geologically meaningful fracture prediction was achieved while providing a good match with measured fracture data. The first case study involves a fractured sandstone reservoir in the Wind River Basin, Wyoming, where CO2 from an ongoing flood is being lost downhole, associated with a natural fracture network. Fractures predicted from structural history modelling showed that fluids were likely to be lost due to E-W orientated fractures, which agreed with records of sub-surface fluid movement in the area. Another case study is presented from an unconventional field in the Midland Basin, Texas. Fluid production rates and hydraulic stimulation diagnostics suggest that natural fractures may influence outcomes. The project aim was to predict fractures in under-sampled areas, and inform future drilling campaigns. Structural analysis indicated that fracture formation was a result of fault reactivation. This reactivation was modelled and fractures predicted to accurately reproduce measured data, and enable prediction of fracture orientations and intensities in areas not yet drilled.