New Method Determines Well Spacing in Unconventional Reservoirs
Present industry solutions to the challenge of well spacing involve expensive geomechanical Earth modeling or fracture-geometry monitoring that is time-consuming, data-intensive, and geography-specific.
Present industry solutions to the challenge of well spacing involve expensive geomechanical Earth modeling or fracture-geometry monitoring that is time-consuming, data-intensive, and geography-specific. The authors of the complete paper write that their developed methodology is quicker and pad-specific and requires no additional data collection than those routinely gathered. Other significant outcomes of this work flow include the ability to rank well performance and predict changes in geology and pressure regimes relative to an existing similarly landed producer.
Despite more than a decade of the shale revolution in the US, present well-spacing techniques are far from ideal. Given the spatial variation of reservoir properties, this problem becomes localized or pad-specific. Conducting many such complex reservoir studies across the acreage is not feasible because they are both cost- and data-intensive. In this paper, a novel methodology is described to estimate well spacing and characterize the reservoir using readily available completion data.
Most well-spacing studies try to characterize the extent of the hydraulic fracture half-length. During pumping of a typical stimulation job, after the rock is broken, the fluid volumes extend the fracture and leak off into the matrix. Once the pumping is stopped, the fluids continue to leak off into the matrix along with ongoing volume accommodation through fracture-tip extension. The fluid leakoff into the matrix will cease when pressure equilibrium is reached between the fracture and matrix. At the point when leakoff ceases, the affected matrix has received an additional influx of fluids from the stimulation job that results in a pressure increase that will dissipate beyond the affected matrix slowly. The novel idea presented in this paper solves for the affected matrix geometry. Thus, the affected matrix becomes the focus, and not the stimulated hydraulic fractures. The propped stimulated fractures simply provide a pathway for flow to occur from the affected matrix to the wellbore. The volume and the geometry of the affected matrix hold the key to determining the appropriate well spacing. The proposed method is to establish the relationship between completion design variables (e.g., volume pumped, cluster spacing) and the affected matrix geometry. The challenge in this approach is the computation of pressure increase for each stage because it is not cost-effective to wait for complete closure of the stimulated fractures or know when pressure equilibrium occurs. This paper describes a novel approach to compute the pressure increase for each stage.
To compute the pressure increase for each stage, two parameters routinely reported in a completion job are used:
- Instantaneous shut-in pressure (ISIP). This is measured at the end of injection, after frictional forces dissipate in the wellbore, perforations, and near-wellbore region.
- Change in ISIP with time. Usually, 5-, 10-, and 15-minute readings are provided from a stimulation job. Pressure measurements over longer times are preferred.
The method uses a collective view of all stages as a form of measurement rather than attempting to derive information from a single stage.
Field Applications and Interpretation
Pressure Regions/Fault Blocks. The relationship between ISIP and ISIP leakoff pressure can be used to determine if changes in pressure exist along the horizontal well. Consider the examples given in Fig 1. Wells A and B, landed in Wolfcamp A, had suspected faults identified from seismic. The challenges presented included the following:
- How accurate is the prediction of the faults from the seismic (i.e., do they exist)?
- Where exactly are they located in the horizontal?
- How do completions change with fault presence?
The ISIP diagnostic plot was able to confirm and identify the location where the fault occurred in those wells. The plot also reveals that three pressure regions have developed because of the faults. In this example, higher pressure was observed in Region 1, and Region 3 had the lowest pressure. The production performance of these wells corroborated the fault blocks and pressure interpretation. The plot also helps to identify stages (Stage A in Fig. 1) that exhibited abnormal leakoff compared with the rest, indicating high-permeability zones (natural fractures or a small local fault). Optimal well spacing and completion designs should be customized to the pressure regions and natural fractures encountered.
Effect of Lithology. Optimal completion designs should consider the effect of lithology. It is highly unlikely that the entire wellbore would be drilled in the same lithology. Though an all-encompassing design would suit the purposes for a single well, the effect of lithology on design variables should be studied when considering pad development. The leakoffs, and hence the affected reservoir matrix, could be different for different lithologies.
Identification of Underperformers/Relative Performance Ranking. Another important application of this plot is the ability to predict future well performance compared with a similarly landed producer, especially the ability to identify the underperformers. An example of this process and its applications is provided in the complete paper.
Well Spacing. The most important application of this diagnostic plot is to help determine the optimal well spacing for a given completion design. At given optimal cluster spacing (or aided by matrix permeability), the affected matrix can be represented, as in Fig. 10 of the complete paper, by an elliptical cylinder. While the stage length represents the height of the cylinder, the semimajor axis and minor axis are represented by the affected matrix half-length and stimulated fracture height, respectively.
The computation of affected matrix half-length involves a multidisciplinary approach that must take into account the uncertainties associated with measurements and certain assumptions.
Dealing With Uncertainties Through Stochastic Approach. The main sources of uncertainties for well-spacing estimation include the following, as well as others detailed in the complete paper:
- Fluid volume that remains in the fractures long after fracture closure at zero leakoff
- Accuracy of reservoir pressure prediction
- Fracture-height growth is required to compute the affected matrix half-length
- Uncertainties in fluid compressibilities estimated from pressure/volume/temperature because of incorrect fluid sampling
- The change in water saturation in the affected matrix because of fluid influx
- The affected matrix half-length computation assumes that the fluids have imbibed through the entire height of the elliptical cylinder model. The uncertainty of this assumption should also be considered.
- Reservoir properties, water saturation, and porosity are estimated through petrophysical models from vertical wells and then extrapolated to horizontal but are prone to uncertainty
- Accuracy of estimating final equilibrium pressure
To handle these uncertainties, a stochastic Monte Carlo approach is prescribed. These variables are modeled with a distribution for each stage, and the resulting median values are studied.
The methodology was applied in several Wolfcamp A wells to determine the effect of the completion design. These results, along with graphs illustrating the findings, are detailed in the complete paper.
The ISIP diagnostic plot is an effective tool in identifying stages experiencing high leakoff because of high permeability created by natural fractures or low-pressure sinks. In some cases, however, the interpretation is not always obvious.
A need exists to integrate multiple diagnostic tools to crosscheck and construct a valid interpretation. Two such diagnostic plots include the relative change of ISIP plotted along the horizontal and the average treating pressure against the mechanical specific energy from drilling data. The anomaly in this diagnostic plot is useful in confirming stages that might have encountered a local fault or a high-permeability streak.
Transfer Function. A transfer function is determined from clusters of stages that undergo similar leakoff into the matrix. Once the best-fit model slope is established, assigning a transfer function slope for each stage is a manual process and prone to uncertainty. In the future, machine-learning processes could be used to automate clustering and perform linear regression and assignment.
Symmetrical Assumption of Affected Matrix Volume Around the Wellbore. The matrix volume modeled from the ISIP diagnostic plot is symmetrical around the wellbore. While stress shadowing between stages is captured in the ISIP measurements, asymmetrical fracture growth caused by stress shadowing or pressure depletion from long-term production is not differentiated. This is a limitation of this methodology, and optimal cluster spacing should be realized from investigating ISIP from subsequent stages.
Effect of Clay and Interbedded Carbonates. Diagnostic plots show that Wolfcamp A reservoir featured the least data scatter compared with Wolfcamp C and D. Wolfcamp A contains the lowest clay and fewer interbedded carbonates, Wolfcamp C contains more clay and moderate interbedding, and Wolfcamp D features substantial interbedding. Further studies need to be performed in understanding the resulting scatter caused by clay and interbedding.
A methodology has been introduced to estimate well spacing from readily available completion data. The emphasis is placed on the fluids pumped during stimulation rather than the fracture half-lengths. The application of the methodology helps in local reservoir characterization and identifying the future underperformers relative to a known producer. The predicted underperformers will help set expectations on reserves and budget. The methodology also guides the selection of optimal completion design with respect to landing zone and lithology for a proposed well-spacing program.
This article, written by JPT Technology Editor Chris Carpenter, contains highlights of paper URTeC 2584, “A New Methodology To Determine Well Spacing in Unconventional Reservoirs: Delaware Basin Study,” by Dicman Alfred and Andrew Lundy, Scala Energy, prepared for the 2020 Unconventional Resources Technology Conference, originally scheduled to be held in Austin, Texas, 20–22 July. The paper has not been peer reviewed.