Implicit and Explicit models of fractures and their properties

Colin Daly
Roxar Ltd

Fractures play an important role in the fluid flow behaviour of very many carbonate fields. While significant progress has been made in developing simple to use stochastic fine scale models for clastic heterogeneity, progress has not been as noticeable in stochastic fracture modeling. It is, of course, possible to develop discrete fracture models to provide patterns within a reservoir. These provide a useful tool for conceptual modeling and QC. But the next stage of how to assign flow properties to these models is often difficult as they are highly parametric. Moreover, upscaling of these models in a routine reservoir modeling environment is complex, especially to account for larger fractures which straddle two or more reservoir simulation blocks. The discrete fracture model is very useful as a means of understanding the matrix/fracture interaction terms used in conventional dual porosity/dual permeability models. However the complexity of the modeling process tends to discourage the operational reservoir modellers from tackling the effects of fractures in their reservoir studies.

We propose a model which attempts to model fractures in a simple yet consistent manner. Firstly, fracture density maps (or parameters if calculated in 3d) for 1 or more fracture sets are produced using techniques such as stress/strain modeling. We mention both the classical boundary element method and a stochastic method. These maps are then calibrated to interpreted well test permeability to provide an implicit model for fracture permeability. The maps are also used for a discrete fracture model which may be used for QC as well as for further calculation of the matrix/fracture terms.