This paper provides a rigorous, physics-based workflow to construct typical well production profiles (aka type wells or type curves) for multi-fractured wells in resource plays. Our approach solves the problem of scaling wells with different matrix properties (notably permeability) and completion characteristics to common reference conditions; this scaling allows the analyst to significantly increase the number of analog wells used to construct the type well, increasing the statistical significance of the final product. We present our proposed workflow and illustrate its application to a 130-well data set from the DJ basin.
In our workflow, the analyst first matches production data (as infrequent as monthly rates) to the Eleiott type curve, a modification of the familiar Wattenbarger type curve for hydraulically fractured horizontal wells. This match provides b-factors for early transient and later boundary-dominated flow (BDF) segments in two-segment Arps decline models. In addition, the type-curve matches provide estimates of effective average matrix permeability and fracture half-length for each well. Given these results and other available reservoir and completion properties, the analyst can scale each well to a common set of reference conditions, including permeability, fracture half-length, fracture spacing, and lateral length. These wells can be placed into separate bins, with the major sorting parameter being the transient b-factor. The reference well production profile is the desired type well for these reference conditions.
For future wells not yet drilled, most parameters (such as lateral length, fracture spacing, and fracture length) will be specified in the well design; only permeability will be beyond the control of the technical team, and the permeability distribution thus becomes the major source of uncertainty in future wells. This known permeability distribution provides the opportunity to construct P10, P50, mean, and P90 type wells to use to predict the performance of future wells (or existing wells with short production histories). We can rescale the type well for the different permeability values and for the design parameters for undrilled wells and generate a customized type well for those wells.
Our workflow can be implemented readily on a spreadsheet, with no need for commercial software. It takes into account the variations in lateral length, hydraulic fracture properties, and matrix permeability, and allows the user to base type well construction on a relatively larger number of wells in a given bin, reducing
uncertainty in the result through increased sample size and thus increased statistical significance while honoring the complex physics of fluid flow in unconventional reservoirs as reflected implicitly in individual well production profiles.
