Online Journal for E&P Geoscientists

Lamine, Sadok ^*1
(1) Shell, Rijswijk, Netherlands.

Naturally fractured reservoirs constitute a large portion of gas and oil reserves. Such reservoirs are highly heterogeneous and complex.

Currently, discrete fracture network (DFN) models provide representative geological models that incorporate the natural fracture features. These models are comprised of fracture sets of different orientations, types and scales that may be entirely of partly connected to form preferential flow paths locally interacting with the rock matrix. On the other hand, the simulation of multiphase flows in naturally fractured reservoirs is generally carried out using dual porosity or dual permeability models. The fracture and matrix media are distinguished and equivalent flow properties for each medium are assigned to each simulation grid cell. The determination of such effective or equivalent parameters is an important task for simulating reservoir flow and predicting future performance. Yet it still is one of the most challenging tasks for fractured reservoir simulation.

In order to handle the inherent heterogeneity of the diverse fracture settings, different fit-for-purpose upscaling techniques that rely on qualitative simplifying assumptions are currently used in reservoir simulation.

The aim of this work is to present a novel hybrid approach that provides a fast and accurate upscaling alternative that preserves the connectivity as well as the scale heterogeneity of the DFN models.