Optimization of Duvernay Fracturing Treatment Design Using Fully Compositional Dual Permeability Numeric Reservoir Simulation

Abstract

Abstract This paper discusses how a fully-compositional, dual permeability numeric reservoir simulation was used to model flow from a Duvernay well pair after fracture stimulation. A geologic model based on well log data, referenced to coordinates on the Universal Transverse Mercator (UTM) coordinate system, utilizing North American Datum 83 (NAD83), Zone 11 UTM, was constructed. A sub-grid of this, containing permeability and porosity data, was imported as the grid for the dynamic reservoir simulator. The subject wellbore location was loaded into the model using UTM coordinates for the wellhead location and the wellbore trajectory survey. Microseismic event locations were also imported, relative to the treatment wellbore, measured during the time the fracturing treatments were conducted. Microseismic data was used to determine which grid blocks to include in a stimulated reservoir volume (SRV) for each fracturing treatment within the reservoir simulation. Within each SRV a dual permeability model was used to capture the flow in both the "primary" hydraulic fracture conduit as well as a "secondary" set of fractures (natural fractures which were also stimulated during the hydraulic fracturing process). Permeability and width for both fracture systems were independently defined. Based on in-situ stresses, it was assumed that the hydraulic fractures grew perpendicular to the wellbores with the SRVs branching into other directions because of the opening of natural fractures through the reservoir. The objective for creating two fracture systems was to simulate complex hydraulic fracturing geometry in communication with a primary fracture system, which then connects to the wellbore. It was assumed that the native unstimulated natural fractures had a negligible to no effect on flow from the native reservoir. The impact in differences to production between the primary and secondary fracture systems was evaluated by studying ranges of values for the primary and secondary fracture system's permeabilities and widths as well as secondary fracture system spacing. Using a fully compositional simulator, an equation of state (EOS) was developed and applied to properly simulate phase behavior under dynamic downhole conditions. This is essential to properly model multiphase flow effects on relative permeabilities and resulting production. Results were then compared between using this dual permeability model to those obtained with a model using only single permeability with transverse (bi-wing) hydraulic fractures (no stimulated reservoir volume) to determine which better simulated actual production for this particular well. An underlying assumption in most, if not all Duvernay well stimulations, is that creation of an enhanced secondary fracture system (often considered a network or SRV) is essential to economic production. The assumption that slickwater stages are required is largely derived from the success using slickwater to fracture stimulate dry gas wells. The use of a compositional reservoir simulator allows studying of the contribution of the secondary fracture system with liquids rich production. The matching of simulation results to actual production through changes in primary and secondary fracture system conductivity values allows investigation of the contribution each fracture system makes to both early and later time production. Through a better understanding of the contributions of both primary and secondary fracture systems in liquids rich production, operators can much better understand how to optimize a fracturing treatment design while helping minimize costs.