Early in the development of a shale gas resource, optimal well spacing remains unknown as wells are sparsely drilled to hold leases by production. Developing the acreage requires operators to select locations, specify drilling plans, and design completions for multi-stage horizontal wells to maximize the operating metrics as defined by the company.
This paper builds on our earlier work which presented sensitivity analysis for optimal well spacing with respect to permeability, fracture spacing and half-length under the assumption of uniform and symmetric completion configurations. The well spacing sensitivity to heterogeneity in completion configurations (i.e., non-uniform fracture half-length and asymmetric fracture spacing) are presented in this paper using forward and stochastic modeling approaches.
Forward modeling results show a strong bias towards the longest repeated fracture half-length in determining the optimal well spacing. Higher reservoir permeability abates the impact of fracture heterogeneity. Fracture modeling, constrained by production logs, temperature logs, and/or microseismic, can be used to aid in the identification of the longest repeated half-length.
This paper demonstrates the challenges associated with stochastic modeling of well performance. Examples from synthetic and field case studies are presented to illustrate uncertainty in reservoir and completion parameter determination. The spacing optimization workflow used captures this uncertain range to effectively determine the impact on recovery factor and Net Present Value (NPV). The importance of the quantity of production history needed to determine optimal well spacing is also presented. Results reveal that with increasing heterogeneity longer production history is required for reliable determination of optimal well spacing. Finally, a field case study applying a production log and fracture modeling is examined to identify the impact of non-uniform fracture spacing and fracture half-length heterogeneity.
These conclusions, via the application of deterministic and stochastic modeling on production from field cases and synthetic wells, will aid operators in answering the multi-billion dollar question: how many wells should be placed in a given area? The workflow described in this paper not only can answer this question but also help us to understand how to maximize economic return and the ultimate gas recovery.