Next week, ResFrac will be coauthoring seven papers at the Unconventional Resources Technology Conference (URTeC). These papers include: operator case studies in the Haynesville, Marcellus, and Bakken, a study quantifying the effect of proppant uniformity on production and economics, a new procedure generalizing the Devon Quantification of Interference (DQI) method, a study on fracture design in the Cane Creek Play in Utah, and an excellent paper by a University of Texas PhD student on proppant flowback. We are excited to share them! Below is a quick teaser on each of the papers.
The schedule is:
- 4:15pm Monday – Augustin Garbino (UT Austin) presenting Modeling of Proppant Flowback to Quantify and Predict its Impact on Shale Gas Production under Different Drawdown Strategies – A Vaca Muerta Case Study
- 8:55am Tuesday – Craig Cipolla (Hess) presenting The Perfect Frac Stage, What’s the Value?
- 8:55am Tuesday – No’am Dvory (University of Utah) presenting Avoiding the Salts: Strategic Fracture Propagation Management for Enhanced Stimulation Efficiency in the Cane Creek Play
- 11:40am Tuesday – Chris Ponners presenting Interference Testing in Shale: A Generalized ‘Degree of Production Interference’ (DPI) and Developing New Insights into the Chow Pressure Group (CPG)
- 8:55am Wednesday – Jose Zaghloul (Continental) presenting Marcellus Field Development Optimization: Multi-Bench Evaluation in Bradford County, PA
- 9:20am Wednesday – Mark Pearson (Liberty Resources) presenting A Case Study of Bakken Development Optimization with Complex Constraints
- 11:40am Wednesday – Chad Jongeling (Chesapeake) presenting A Physics-Based Approach to Characterize Productivity Loss in the Haynesville Shale
A Physics-Based Approach to Characterize Productivity Loss in the Haynesville Shale
Chad Jongeling1, Jerrod Ryan1, Christopher Ponners2, Dominick Wytovich1, Joe Miller1, Josh Jackson1, Mark McClure2, Garrett Fowler2, 1. Chesapeake Energy Corporation, 2. ResFrac Corporation
11:40 AM Wednesday
This case study is differentiated by its use of field testing to constrain a key modeling uncertainty. In the Haynesville, strong deviation from linear flow is observed in rate-transient analysis (RTA). Deviation from linear flow corresponds to a loss of productivity, relative to ‘infinite acting linear flow.’ This is common in all shale wells, but it is especially strong in the Haynesville. There are various potential causes, but since the Haynesville is single phase gas, we can rule out multiphase effects.
In this study, the Chesapeake engineers performed interference tests at three different points in time during the first year of production. These tests make it possible to quantitatively estimate the fracture conductivity, and to track the progressive reduction in conductivity with drawdown. In addition, Chesapeake performed core analysis to measure changes in matrix permeability as a function of effective stress.
With these constraints – along with DFIT-derived stress profiles and ‘initial’ permeability estimates – they achieved a close match to production for the seven wells in the study area.
The strength of this case study is that it makes targeted use of field-data collection to resolve a key model input. This is exactly the kind of cost-effective diagnostic that can help clarify the conceptual model and constrain model inputs. In this case, they were able to unambiguously demonstrate why the RTA shows such strong deviation from linear flow.
With the calibrated model, they simulated a variety of frac design optimization scenarios. The results have informed subsequent fracture designs.
A Case Study of Bakken Development Optimization with Complex Constraints
Mark Pearson1, Stacy Strickland1, Larry Griffin1, Janz Rondon2, Dave Ratcliff2, Garrett Fowler2 1. Liberty Resources, Denver, CO, United States, 2. ResFrac, Palo Alto, CA, United States
9:20 AM Wednesday
This case study was performed in the Bakken. They built and calibrated a high-quality model, using data such as microseismic, sealed wellbore pressure monitoring, ISIPs, and production. They achieved a good match to the data across six wells, including both parent and some child wells, and then performed design optimization.
This paper has a particularly detailed description of the frac design and economics optimization. The paper shows sensitivities on well spacing, clusters per stage (with limited entry), stage length, job size, and drawdown schedule.
For economic optimization, they generated plots of Discounted Profitability Index (DPI) per DSU and NPV per DSU for a large number of prospective frac designs, and assuming either 5, 6, or 7 wells per DSU. The results demonstrate the competition between NPV and DPI. NPV is maximized with 7 wells per DSU, but DPI is maximized with 5 wells per DSU. If limited to maintaining the same historical capital expenditure per DSU, 6 wells per DSU maximizes NPV. For all three cases, the simulations show the range of frac, stage, and cluster designs that are optimal.
Based on these results, the operator selected the best design, based on their weighting of objectives and other constraints.
Marcellus Field Development Optimization: Multi-Bench Evaluation in Bradford County, PA
Roberto Wagner1, Jose Zaghloul2, Travis Glauser1, Chance Morgan1, Patrick Crump1, Ashley Mercer1, Matt Mantell1, Dave Ratcliff3, Garrett Fowler3, Mark McClure3, 1. Chesapeake Energy Corporation, 2. Continental Resources, 3. ResFrac Corporation
8:55 AM Wednesday
This study investigated alternative well configurations and frac design in a sweet-spot area of the Marcellus. The legacy design places wells at 1100’ spacing in the basal part of the Lower Marcellus. The study evaluated new designs with either: (a) wine-racked upper/lower Marcellus developments, or (b) three-bench wine-racks across Upper, Lower, and Basal Lower Marcellus. Within each configuration, they considered various wine-rack patterns and frac designs.
Petrophysical measurements, modeling, and field diagnostics suggest that, in this area, the Cherry Valley interval (between the Upper and Lower Marcellus) is not sufficiently thick to prevent fracture growth up into the Upper Marcellus. It acts as a partial hindrance. Some fractures propagate through, and others do not. The result is partial drainage of the Upper Marcellus. Placing wells in the Upper Marcellus increases overall recovery, but also increases CAPEX.
The economic optimizations come down to commodity price. At low prices, they find that the legacy design with wells only in the basal Marcellus is the most efficient. However, with higher prices, the multibench development becomes preferred.
The Perfect Frac Stage, What’s the Value?
Craig Cipolla1, Ankush Singh2, Mark McClure2, Michael McKimmy1, John Lassek1, 1. Hess Corporation, 2. ResFrac Corporation
8:55 AM Tuesday
This paper asks – what is the economic value of fracturing uniformity? Over the years, the industry has worked hard to identify designs that maximize the uniformity of fluid and proppant placement. Plug and perf completion has become preferred to sliding sleeve/openhole completion because it forces more fractures to form, at a tighter spacing. With plug and perf, limited-entry completion has become preferred to improve uniformity. Most recently, downhole imaging has been used to measure erosion, quantify uniformity, and optimize perforation design to account for proppant inertia, gravitational effects, and shot erosion. Considering these efforts, this paper asks – what is the upside?
To address, we used a ResFrac model that has been calibrated to a variety of outstanding Hess datasets that they have gathered over the past several years, including offset pressure observation wells, fiber optic, and tracer. Using this model, we ran scenarios in which we enforced ‘high,’ ‘medium,’ and ‘low’ values for uniformity. The ranges were selected to represent plausible values for a modern plug and perf design. Thus, even the simulation with ‘low’ value of uniformity had better performance than would be expected from a design with no limited entry, or a design with sliding sleeves and openhole.
The ‘high uniformity’ case had roughly 10% greater production than the ‘medium’ case. The ‘low’ uniformity case had roughly 10% lower production than the ‘medium’ case. In terms of net-present value, the difference between low and high was roughly \$2.5M. Overall, the paper concludes “Every 0.1 improvement in UI is predicted to increase first year production about 2.5% and add about \$0.3 million in NPV.”
These results can be helpful for operators evaluating changes to improve uniformity. For example, an operator might consider self-orienting perforation guns to improve phasing consistency, or an operator might implement a real-time optimization scheme to maximize uniformity. Prior to implementing these changes, modeling could be used to evaluate their likely impact (and assess if they are worth trying). After implementing, downhole imaging can help them evaluate the uniformity before and after such changes. The results of this paper relate those observations to production and economics.
Modeling of Proppant Flowback to Quantify and Predict its Impact on Shale Gas Production under Different Drawdown Strategies – A Vaca Muerta Case Study
Agustin G. Garbino1, D. Nicolas Espinoza1, Mark McClure2, Marcelo Pellicer3, Iñaki Barrangu3, Sebastian Olmos4, Federico Salicioni4, 1. The University of Texas at Austin, 2. ResFrac Corporation, 3. Pan American Energy, 4. Tecpetrol SA
4:15 PM Monday
This outstanding paper was written by Agustin Garbino, as a part of his PhD at UT Austin in the Department of Petroleum and Geosystems Engineering. ResFrac has the ability to model proppant flowback during production, based on the correlation from Canon et al. (2003). However, until this paper, ResFrac’s proppant flowback capability has not been tested in a rigorous case study. Augustin gathered field data from Pan American and Tecpetrol in the Vaca Muerta and used it to calibrate and validate the model.
The flowback correlation has an empirical fitting parameter, which was used to match actual observed flowback data. For two separate wells, the model was able to match water and gas production, in addition to produced proppant mass per day.
Next, a series of sensitivity analysis simulations were performed, testing the effect of drawdown. As expected, more aggressive drawdown led to greater proppant flowback. Interestingly, in the base case simulations, the effect of that flowback on production was limited. The fracture pack conductivity was sufficiently high that the loss of some proppant near the well had only a minor impact on production.
The sensitivities were repeated with changes to fracture conductivity. With lower conductivity, the effect of flowback on production became more significant. In simulations with time-dependent conductivity loss, the loss of proppant due to flowback had a larger effect at late time than at early time. The fracture conductivity is a quantity that can be measured using interference tests.
Interference Testing in Shale: A Generalized ‘Degree of Production Interference’ (DPI) and Developing New Insights into the Chow Pressure Group (CPG)
Chris Ponners1, Mohsen Babazadeh2, Craig Cipolla3, Karan Dhuldhoya2, Qin Lu2, Ripudaman Manchanda4, Daniel Ramirez Tamayo4, Steve Smith4, Mojtaba Shahri5, Mark McClure1; 1. ResFrac Corporation, 2. ConocoPhillips Company, 3. Hess Corporation, 4. Exxon Mobil Corporation, 5. Apache Corporation
11:40 AM Tuesday
This paper was written as part of a collaborative effort between ResFrac, ConocoPhillips, ExxonMobil, Hess, and Apache. We had two goals: (a) improve our understanding of the ‘Chow Pressure Group’ (CPG) metric, and (b) extend the ‘Devon Quantification of Interference’ (DQI) procedure from Almasoodi et al. (2023) to account for arbitrary wellbore configurations and heterogeneity of fracture geometry along strike and between clusters.
Regarding the CPG, in previous work, we’d observed that larger values of CPG do, in fact, correlate with greater well-to-well interference, but that there is significant scatter in the relationship. In this paper, we tried to explain why. To briefly summarize – the CPG starts low and increases over time; with stronger well-to-well communication, the low-to-high progression occurs sooner. As a result, since CPG is typically measured at a particular time (24 or 48 hours), stronger communications means that the measured CPG is higher. Overall, we conclude the CPG is capable of qualitatively assessing communication between wells, but that because it does not account for variability in reservoir properties, it cannot be used in a precise way to quantitatively assess interference.
Regarding the DQI method from Almasoodi et al. (2023), we develop an algorithm to calculate the ‘degree of production interference’ (DPI), or the ‘fractional production loss’ (FPL), for wells in any configuration. The new procedure accounts for wine-rack configurations, anisotropy of fracture conductivity, and fracture height growth barriers. Results from multiple interference tests, at different distances, can be aggregated to form a ‘conductivity versus distance’ curve. Counterintuitively, we show that because proppant is injected from both wells, the ‘measured’ conductivity versus distance curve is different from the ‘actual’ conductivity versus distance curve. Finally, we show how to account for variability in conductivity among different fractures connecting the wells.
Overall, the result is a procedure for mapping depletion along and between fractures and between wells. While not as rigorous as a full modeling study, this ‘generalized DQI’ procedure enables rapid, realistic assessment of well spacing and wine-rack configurations, based on interference tests and reservoir properties.
Avoiding the Salts: Strategic Fracture Propagation Management for Enhanced Stimulation Efficiency in the Cane Creek Play
No’am Zach Dvory*1, John David McLennan2, Ankush Singh3, Brian James McPherson1; 1. Civil & Environmental Engineering & Energy and Geoscience Institute, The University of Utah, USA, 2. Chemical Engineering & Energy and Geoscience Institute, The University of Utah, 3. ResFrac Corporation
8:55am Tuesday
This modeling study investigates fracture design optimization in the Cane Creek Play in southeastern Utah. The target interval is overlain by a halite interval. It is desirable to avoid fracturing into this layer, to avoid inefficiently placing proppant, and also, to avoid salt influx that can clog the proppant pack. The authors identify tweaks to the fracture design that can be used to limit height growth and maximize economic performance.
