Upscaling Fault Flow Properties for Evaluation of Cross-Fault Flow
An ongoing concern in flow modeling across faults is upscaling the fault-rock flow properties measured in the lab on cm-scale to the meter and tens of meter scale suitable for evaluating within-reservoir flow and exploration-scale sealing. It is proposed that upscaling is best accomplished by stochastic modeling of cross-fault flow guided by the geometries of fault-generated rock bodies (“geobodies”) as observed on outcrop. Preliminary models demonstrate that most resistance to flow transverse to the fault is in the low permeability fault rocks the fault core as opposed to the damage zone of deformation in the fault zone. Published and proprietary outcrop studies have demonstrated that the fault core is heterogeneous normal to the fault. Spatial correlation of lithology transverse to the fault core is on the order of mm to a cm, much too small to improve estimates of upscaled flow properties. Length of constituent bodies in the fault core can be traced along the fault on the scale of cm to meters in different settings. The fault core is best described as a composite of plate-shaped or equant geobodies with different permeabilities and seal capacities. Upscaling by numerical simulations of flow in composite rocks with these geometries can be compared to mathematical upscaling averages. Equant geobodies result in upscaled permeabilities very close to the geometric mean of the constituent geobody permeabilities and their volumes. Sealing depends on the relative size of the geobody relative to the width of the fault core. Plate-shaped geobody constituents lead to upscaled permeability between the harmonic and geometric means of the constituent geobody permeabilities, with upscale permeability approaching the geometric mean as the shape of the constituent geobody approaches equant. The thinner the geobody relative to the fault core width, the greater the seal capacity of the composite medium. Realistic upscaling of lab-measured permeability to real faults therefore depends on both the fractions of geobodies with different interpreted permeability and the likely geometry of the geobodies. Settings with smear and shearing of ductile rocks lead towards platy fault core geometries, whereas settings where lenses or protolith is mixed into the fault core lead towards a more equant geobody shape.
AAPG Datapages/Search and Discovery Article #90189 © 2014 AAPG Annual Convention and Exhibition, Houston, Texas, USA, April 6–9, 2014