Permeability of fault rocks in siliciclastic reservoirs: Recent advances

It is common practice to create geologically realistic production simulation models of fault compartmentalized reservoirs. Data on fault rock properties are required, to calculate transmissibility multipliers that are incorporated into these models, to take into account the impact of fault rocks on fluid flow. Industry has generated large databases of fault rock permeability, which are commonly used for this purpose. Much of the permeability data were collected using two inappropriate laboratory practices with measurements being made at low confining pressure with distilled water as the permeant. New fault rock permeability measurements have been made at high confining pressures using formation compatible brines as the permeant. Fault permeability decreases by an average of five fold as net confining pressure is increased from that used in previous measurements (i.e. ∼70 psi) to that approaching in situ conditions (i.e. 5000 psi). On the other hand, permeability increases by around the same amount if reservoir brine is used as the permeant instead of distilled water. So overall, these two inappropriate laboratory practices used in previous studies cancel each other out meaning that legacy fault rock property data may still have value for modelling cross-fault flow in petroleum reservoirs. A poor correlation exists between clay content and fault rock permeability, which is easily explained by the application of a simple clay-sand mixing model. This emphasises the need to gather fault permeability data directly from the reservoir of interest. The cost of such studies could be significantly reduced by screening core samples using a CT scanner so that only samples that are likely to impact fluid flow are analyzed in detail. The stress dependence of fault permeability identified in this study is likely to be primarily caused by damage generated during or following coring. So it is probably not necessary to take into account the impact of stress on fault permeability in simulation models unless the faults of interest are likely to reach failure and reactivate.