The objective of this study is to prove the concept of using a seismically derived discrete fracture network (DFN) calibrated with borehole measurements, for complex hydraulic fracture modeling. This study was successfully applied on a two-well horizontal pad in the Avalon shale located in the Delaware basin, New Mexico.

The interaction of propagating hydraulic fractures with natural fractures plays a crucial role in defining the complexity and extent of the hydraulic fracture geometry. In a multi-well pad and infill-well scenarios, a good understanding of the effective drainage area of the hydraulic fracture created is essential for optimum well spacing to reduce interference between wells. In multi-stacked laterals, an understanding of the fracture height is necessary for placing the next well to avoid hydraulic fracture interference vertically and to avoid water wet zones. A DFN model built from borehole image logs is limited in its ability to account for the lateral variability of the natural fracture network properties away from the wellbore; these variations have a significant impact on the created hydraulic fracture geometry, and thereby drainage, away from the wellbore.

In this study, the DFN model was created from depth-converted 3D narrow azimuth seismic data. The seismic discontinuity plane (SDP) extraction and fracture characterization method was used to extract the seismic-scale fractures out of the best matching seismic attribute cube. Using orientation, P32 density (fracture area/unit volume), and fracture length distribution per individual seismic-scale fracture sets, a corresponding subseismic population of discrete fractures was stochastically modeled. Then the extracted seismic-scale fractures and the modeled subseismic fractures were combined to form a comprehensive multi-scale DFN model. The result was then incorporated into a state-of-the-art unconventional fracture model (UFM) that models the explicit interaction of the hydraulic fractures with natural fractures, to determine a good representation of the created complex fracture geometry observed from microseismic data.

Available microseismic data was used to validate the DFN and UFM models. The DFN model showed natural fracture intensity variation along and away from the wellbore and for the most part agreed with the footprint of the microseismic events. This variation was also captured by the UFM simulator, while modeling the hydraulic fracture geometry with the observed treating pressure profiles.

The workflow presented in this paper shows the successful application of seismic data in creating discrete fracture networks required to model hydraulic fractures in unconventional reservoirs. This workflow is critical for understanding the variation of the natural fractures along a planned well, within a pad, or at a basin scale in a predictive manner. Prior knowledge of the in-situ 3D natural fracture network, and its spatial variability and anisotropic stress profile is key to optimizing the overall completion and development strategy of an unconventional resource in a cost-effective way.