Abstract: Modelling Flow in Faulted Sandstone Reservoirs

Thomas Manzocchi, Philip S. Ringrose, John R. Underhill

Faults that are too small to be detected seismically are often modelled probabilistically. This work investigates some of the assumptions inherent in these techniques, in particular regarding scaling relationships and fault geometries.

Probabilistic techniques extrapolate the scaling properties of seismically resolved fault systems, usually by assuming that fault populations are self-similar. Self-similarity is assumed based on a power-law relationship between fault frequency and fault size. Such a relationship only proves scale-invariance of a particular geometrical characteristic of the system. For a truly self-similar system, all fault system attributes must have complementary power-law slopes. Field datasets have been studied from high porosity sandstone outcrops in the north of England and Utah. A comparison of the power-law gradients of various fault system attributes shows that scaling is highly lithology controlled, and that these fault systems may be self-affine but are not self-similar.

Small scale faults in porous sandstones are flow baffles, and so flow is controlled not only by fault density, but also by the geometrical architecture of the fault system. A probabilistic fault placement methodology cannot represent the complex fault architecture present at a small scale. Flow simulation models on fault systems with various geometrical characteristics show that effective permeability is highly sensitive to fault geometry, as well as fault density. An improvement to a probabilistic fault placement methodology based on the fault connectivity characteristics allows a better assessment of the effect of small-scale faults on reservoir flow behaviour.

AAPG Search and Discovery Article #90956©1995 AAPG International Convention and Exposition Meeting, Nice, France