This short paper presents an overview of some critical issues related to fracture modelling and upscaling, using as an example, borehole image data from a carbonate field in the Middle East.

Relationship between fracture density and mechanical layer characteristics is common knowledge. A fundamental question is whether fracture-prone units have a larger number of fracture corridors and mega-fractures. Fracture corridor spacing may be large controlled by the fault size and is independent of mechanical layers. Alternatively, the spacing of fracture corridors and mega-fractures may be controlled largely by mechanical layer characteristics to a large extent. In this Field, the fracture prone units seem to have higher number of fracture corridors and mega-fractures than less fracture- prone units.

One important step in fractured reservoir characterization is estimating the percentage of fracture fairways (fault zones) detectible from seismic data. Without such an estimate, fracture models can only yield a global history match and can not be used for well-planning. The importance of mapping fracture fairways is high when the fracture corridors are much clustered in fracture fairways and fault zones. For a low degree of clustering, it may not be so important to locate all fracture fairways, since fracture corridors will be widely spread with a low degree of connectivity. Fracture analysis showed that seismic fault maps show only half of fracture fairways and small faults that are present in this field. The remaining half of the faults must be generated stochastically and used in stochastic fracture corridor discrete and upscale models along with actual seismic faults. Introduction of stochastic faults poses a serious challenge to reservoir simulation because classical history match procedures can no longer be used to evaluate a fracture model.

You can access this article if you purchase or spend a download.