Abstract:

Tensile, opening mode fractures created in a variety of low matrix permeability rocks have initial, unstressed apertures in the µm to mm range as determined from image analyses of X-ray CT scans. Subsequent hydrostatic compression of these fractured samples with synchronous radial strain and flow measurement indicates that both mechanical and hydraulic aperture reduction varies linearly with the natural logarithm of effective normal stress. These stress-sensitive single-fracture laboratory observations are upscaled to models of fracture populations displaying frequency-length and length-aperture scaling laws commonly exhibited by natural fracture arrays. Functional relationships between reservoir pressure reduction and fracture network porosity, compressibility and directional permeabilities are generated that can ultimately be exported to the reservoir simulator for improved naturally fractured reservoir performance prediction.

Introduction

Many fine-grained reservoirs (clastics, carbonates and mudrocks) require the additional permeability associated with partially open natural fractures (NF's) to achieve economic flow rates. Depletion-driven matrix compaction is routinely accounted for when simulating the performance of conventional reservoirs such as deep-water sands (Guenther et al, 2005; Pourciau et al, 2005) however predicting in a similar fashion the impact of declining fluid pressure on naturally fractured reservoir (NFR) productivity is less well-established.

Uniaxial strain testing of recovered core provides direct measurement of formation compressibility and matrix permeability reduction (Crawford et al, 2011; Ewy et al, 2012) that can subsequently be incorporated in conventional reservoir simulation studies to account for compaction. Challenges in capturing analogous geomechanical effects in NFR performance prediction can in part be attributed to a paucity of hydromechanical measurements of NF response to effective stress changes appropriate to the hydrocarbon reservoir environment and difficulties in upscaling the laboratory stress-sensitivity of single fractures to in situ NF populations exhibiting frequency-size distributions in geometric attributes.

We report new measurements of stress-dependent mechanical and hydraulic aperture reduction in fractured specimens and describe a workflow for upscaling this dynamic response to discrete fracture network (DFN) models which we use to rank the influence of in situ stress orientation and magnitude, pressure depletion and fracture surface roughness on network compressibility and permeability response.

This content is only available via PDF.
You can access this article if you purchase or spend a download.