Cronkwright, David (University of Calgary) | Ghanizadeh, Amin (University of Calgary) | DeBuhr, Chris (University of Calgary) | Song, Chengyao (University of Calgary) | Deglint, Hanford (University of Calgary) | Clarkson, Chris (University of Calgary) | Ardakani, Omid (Geological Survey of Canada)
This paper was prepared for presentation at the Unconventional Resources Technology Conference held in Denver, Colorado, USA, 22-24 July 2019. The URTeC Technical Program Committee accepted this presentation on the basis of information contained in an abstract submitted by the author(s). The contents of this paper have not been reviewed by URTeC and URTeC does not warrant the accuracy, reliability, or timeliness of any information herein. All information is the responsibility of, and, is subject to corrections by the author(s). Any person or entity that relies on any information obtained from this paper does so at their own risk. The information herein does not necessarily reflect any position of URTeC. Any reproduction, distribution, or storage of any part of this paper by anyone other than the author without the written consent of URTeC is prohibited. Abstract Fluid distribution and fluid-rock interactions within the nano-/macro-porous pore network of tight oil reservoirs will affect both primary and enhanced oil recovery (EOR) processes. Focusing on selected samples obtained from the liquids-rich reservoirs within the Montney Formation (Canada), the primary objective of this work is to evaluate the impact of mineralogical composition on micro-scale fluid distribution at different saturation states: 1) "partially-preserved" and 2) after a series of core-flooding experiments using reservoir fluids (oil, brine) under "in-situ" stress conditions. Small rock chips (cm-sized), sub-sampled from "partially-preserved" (using dry ice) core plugs, were cryogenically frozen and analyzed using an environmental field emission scanning electron microscope (E-FESEM) equipped with X-ray mapping capability (EDS).
Gao, Yanling (State Key Laboratory of Petroleum Resources and Prospecting, China University of Petroleum) | Wu, Keliu (State Key Laboratory of Petroleum Resources and Prospecting, China University of Petroleum) | Yang, Sheng (University of Calgary) | Dong, Xiaohu (State Key Laboratory of Petroleum Resources and Prospecting, China University of Petroleum) | Chen, Zhongliang (State Key Laboratory of Petroleum Resources and Prospecting, China University of Petroleum) | Chen, Zhangxing (University of Calgary / State Key Laboratory of Petroleum Resources and Prospecting, China University of Petroleum)
In this work, we derive an effective mean free path (MFP) model for the confined gases in nanopores of shale gas reservoirs by taking into account the effects of the geometrical termination and the surface-gas interaction of the boundary. Among which, the effect of the surface-gas interaction is represented by a probability distribution function for the free flight directions of the gas molecules depending on the surface-gas potential strength ratio (εwf/εff). The validity of the model is verified by comparing the obtained MFP distribution with molecular dynamics (MD) simulation data in previous literatures. Results show that the effective MFP decreases with the increasing Knudsen number (Kn) as well as the increasing surface-gas potential strength ratio, and it is more sensitive to Kn; moreover, the reduction extent is more obvious in the center of the channel than that near channel wall region at both conditions.
With the improvement of manufacturing technology in recent years, the application of micro/nano scale devices is more and more extensive (Cao, et al., 2009; Giordano, et al., 2001; Arlemark1, et al., 2010), which has attracted considerable attention and interests of many experts and scholars. Compared with macro-scale apparatus, the micro/nano scale one has a much larger surface-to-volume ratio, which even shows a huge difference with several orders of magnitude according to Cao et al. (2009). In shale gas reservoirs, most of the pores are also nanoscale and the specific surface area is generally large (some even up to 103.7 m2/g), which are significantly different from that of conventional oil and gas reservoirs (Wu and Chen, 2016). The surface-related factors have a great impact on the flow of confined gas, and among these factors, the surface force or the surface-gas interaction strength plays an important role in the momentum and energy transport, and it cannot be neglected (Barisik and Beskok, 2012).
In the Dunvegan Kaybob South Pool, recent multistage fracked horizontal wells have revealed the presence of a light oil play enveloping a large legacy gas field, developed with vertical wells. The boundary between the oil and gas producing areas intersect structural contours a high angle within deltaic sandstones of the Cretaceous Dunvegan Formation. To address controls on this boundary, a multidisciplinary study of cores, core analysis data, well logs was completed and integrated with test and production data to identify controls on fluid production.
Legacy gas production is from relatively high permeability delta front sandstones, while oil dominated production occurs from lower permeability, fine grained pro-delta deposits. While wells within the legacy gas field produce very low volumes of oil, core fluid extractions reveal significant oil is also present within this portion of the reservoir, but is not mobile. The Dunvegan clearly demonstrates permeability as the main control on the anomalous fluid distributions, with several other tight sandstone plays showing similar relationships, although often more subtle, such as observed in the Cardium, Montney, etc.
The anomalous fluid distributions with higher gas saturations in higher permeability beds and higher oil saturation in lower reservoir quality beds contradict conventional capillary reservoir charge models. Thus, we propose late stage migration of predominantly gas related to the increase in gas generation post peak oil window due to increasing maturity of the kerogen during burial. These late generated gas fluids migrated from the deeper part of the basin preferentially within higher permeability strata and fractures, and displace the earlier emplaced oil resulting in reservoirs with high GOR. These counterintuitive observations with higher liquids production from lower reservoir quality, can significantly improve the play economics and allow better prediction of fluid distribution in many plays.
Although unconventional low permeability reservoirs form laterally continuous thick hydrocarbon accumulations, they often have variable liquid saturations vertically and laterally. While varying kerogen type and maturity are important controls. In several plays, fluid distribution shows a strong correlation with permeability, with higher gas saturations occurring in more permeable beds. The control of permeability on anomalous fluid distribution has been discussed for several clastic, low permeability unconventional light oil and liquid rich gas plays in the Western Canada Sedimentary Basin (e.g. Wood and Sanei 2016, Venieri and Pedersen 2017). In this study we present a study of a legacy gas pool producing from deltaic sandstone reservoirs of the late Cretaceous Dunvegan Formation (Figure 1). The pool is located within the deep basin of western Alberta, an area of pervasive hydrocarbon saturation charged by enveloping thermal mature organic rich mudstones and coals (Masters 1984). The Dunvegan Kaybob South Pool is comprised of a lowstand delta lobe of the southward prograding Dunvegan Delta (Bhattacharya 1993).
In self-sourced low-permeability reservoirs the efficiency at the interaction between the mudstone matrix and fractures is a key control on well performance. Commonly, the more heterogeneous (interbedded) the reservoir the more complex fracture network is naturally developed or can be achieved during stimulation. In this study, using observations from two different unconventional shale units, we demonstrate that mudstone stratigraphic heterogeneities are scale dependent, and thus capturing their expression at different scales is key to understanding the level to which facies arrangements can affect important petrophysical, geochemical and geomechanical properties. Characteristics from the Duvernay Formation in Alberta-Canada and the Woodford Shale in Oklahoma-USA were compared in this study; both units are Late Devonian in age and are organic-rich prolific reservoirs. Lithologies in the Duvernay mostly vary according to changes in carbonate content, whereas in the Woodford changes are according to quartz content. However, in both cases a systematic alternation of two distinct rock types is evident at the cm-scale in outcrops and cores: organic-rich and calcite-rich facies for the Duvernay, and mudstones and chert facies for the Woodford. By combining high-resolution geochemical and geomechanical data, two distinct trends were evident for both units, and illustrate that variations in organic contents, mineralogy and relative hardness can be grouped by the two main rock types for each unit. In the Duvernay, the calcite-rich facies occur as low-TOC beds, at the microscale these are dominated by pore-filling calcite cements. In the Woodford, chert beds present the lower TOC content and their microfabric consists of microcrystalline aggregates of biogenic/authigenic quartz. In both units, the higher porosity values correlate with the high-TOC beds with abundant interparticle porosity. As for mechanical hardness and natural fractures, the higher calcite and quartz contents positively correlate with stiffer beds which generally are more brittle and have more natural fractures. The interbedded character between high-TOC and low-TOC beds is common for both units but at different frequencies and thickness. Capturing the degree of interbedding using a heterogeneity index suggests that reservoir behavior might be depicted as a multi-layered model in which properties are affected by the thickness, permeability, storage capacity, stiffness and fracture frequency of each bed. Although sometimes neglected, the study of fine-scale variations in reservoir properties can provide significant criteria for the selection of optimal horizontal landing zones.
Clarkson, Christopher R. (University of Calgary) | Yuan, Bin (University of Calgary) | Zhang, Zhenzihao (University of Calgary) | Tabasinejad, Farshad (University of Calgary) | Behmanesh, Hamid (NCS Multistage) | Hamdi, Hamidreza (University of Calgary) | Anderson, Dave (NCS Multistage) | Thompson, John (NCS Multistage) | Lougheed, Dylan (NCS Multistage)
The dominant transient flow regime for multi-fractured horizontal wells producing from low-permeability and shale (unconventional) reservoirs has historically been interpreted to be transient linear flow (TLF) in the framework of classical diffusion (CD). Recently, observed deviations away from this classical behavior for Permian Basin Wolfcamp shale (oil) wells have been attributed to anomalous diffusion (AD). The objective of the current study is to systematically investigate other potential causes of deviations from TLF.
The conventional log-log diagnostics used to identify flow regimes do not account for reservoir complexities such as multi-phase flow and reservoir heterogeneity. Failure to correct for these effects when they are occurring may result in misdiagnosis of flow regimes. A new workflow is therefore introduced herein to improve flow regime identification when reservoir complexities are exhibited, and to provide a more confident diagnosis of AD behavior. The workflow involves the correction of log-log diagnostics for complex reservoir behavior through the use of modified pseudo-variables (pseudo-pressure and pseudo-time) after the complex reservoir behavior is identified. Although reservoir heterogeneity is an accepted cause of deviations from TLF, the impact of multi-phase flow has not been investigated in detail. Therefore, in this study, corrections to pseudo-variables for multi-phase flow, a known reservoir complexity exhibited by Wolfcamp shale wells, are presented. Pressure-dependent permeability is also accounted for in the pseudo-variable calculations, although its impact is demonstrated to be relatively minor in this study.
Application of the new workflow to a simulated case and a Wolfcamp shale field case demonstrates the following: 1) multi-phase flow, and in particular the appearance of a mobile gas phase after two-phase oil and water production, results in deviations from classical TLF behavior when data is analyzed using conventional (uncorrected) diagnostics; 2) this deviation has characteristics similar to that expected for sub-diffusion; 3) application of the modified diagnostics to a simulated case that includes multi-phase flow results in the “true” flow regime signature of TLF being observed; 4) application of the modified diagnostics to a field case exhibiting evidence of multi-phase flow reduces the deviation from TLF.
Unconventional resources such as Bakken shale have made a significant impact on the global energy industry, but the primary recovery factor still lingers from 5% to 15 %. Over the past ten years, a number of pilot tests for both gas and water injection or their cyclic injection have been implemented to improve oil recovery in the Bakken Formation. The available public data show that the injectivity is not a problem, but only a small increase in production. The obvious reason is unexpected early breakthroughs even with a relatively low reservoir permeability of around 0.03 mD. Lots of experimental and simulation studies have been conducted to investigate different mechanisms behind these improved oil recoveries. However, no one has succeeded to clarify this early breakthrough.
In this study, a simulation reservoir model, including two wells, is developed, whose properties are based on public data. In terms of hydraulic fractures for each well, their geometry and conductivities are evenly built. Furthermore, our geomechanical module is applied to capture the evolution of stress field and rock failure, where a Barton-Bandis model and a Mohr–Coulomb failure criterion are applied to model tensile and shear failure, respectively. Our simulation model coupled with the geomechanical module is then implemented to explain the performance of injection pilot test.
The results of this initial study clearly show the new fractures (frac-hits) induced by water injection connect the injection and production wells, resulting in the early water breakthrough. The stress field has also been altered by the production process to favor the formation of these fractures. This study highlights the importance of geomechanics during an IOR process; identifies the reasons for the early breakthrough and provides an insight view about how to improve oil production in the Bakken Formation.
Zhang, Zhenzihao (University of Calgary) | Clarkson, Christopher (University of Calgary) | Williams-Kovacs, Jesse (University of Calgary) | Yuan, Bin (University of Calgary) | Ghanizadeh, Amin (University of Calgary)
Quantitative flowback analysis can be used to obtain early hydraulic fracture property estimates which, in turn, can be used to guide stimulation and well operations decisions on future wells/pads. Most quantitative studies of flowback data have primarily utilized rate and pressure data to derive fracture/reservoir properties. However, salinity data contains important additional information that can be used to constrain flowback modeling. In this work, salt transport modeling is combined with a previously-developed frac-through-flowback model based on the dynamic drainage area (DDA) concept in order to constrain the reservoir matrix and hydraulic fracture property estimates. The mechanisms of salt mixing, dispersion/diffusion and advection are captured in the salt-transport model.
In previous work, an integrated model comprised of the following components was developed: 1) hydraulic fracture propagation and proppant transport model; 2) leakoff model; and 3) flowback model. The integration of these components has proven useful for a) constraining hydraulic fracture property estimates (e.g. fracture half-length) and b) modeling the initial pressure and saturation conditions in the fractures and reservoir at the start of flowback. The inclusion of salt transport modeling during flowback to match salinity profiles also helps to constrain matrix and fracture property estimates. For this purpose, salt mixing, dispersion/diffusion, and advection during hydraulic fracturing treatment, subsequent shut-in, and flowback are modeled using a finite-difference-based salt-transport model coupled with a black-oil simulator. The salt transport model was validated against the analytical solution for a diffusion-advection problem, while the black-oil simulator was verified with CMG®-IMEXTM. The coupled salt-transport/black-oil simulator was then tested against a field case to demonstrate its practical applicability.
Water salinity during flowback was precisely matched using the coupled salt-transport/black-oil simulator for the field case. The matrix permeability evaluated using the frac-through-flowback model was constrained by history matching the flowback salinity data using the developed simulator. Laboratory-derived stress-dependent properties served to reduce the number of simulation runs performed during history-matching. The distribution of chemical species in the reservoir was tracked using the new model. Advection and advection-related dispersion were found to be the dominant mechanisms affecting flowback salinity, contrary to the findings of some previous studies.
Dong, Xiaohu (China University of Petroleum Beijing) | Liu, Huiqing (China University of Petroleum Beijing) | Wu, Keliu (China University of Petroleum Beijing) | Liu, Yishan (China University of Petroleum Beijing) | Qiao, Jiaji (China University of Petroleum Beijing) | Gao, Yanling (China University of Petroleum Beijing) | Chen, Zhangxin (University of Calgary)
The presence of nanopores in tight and shale rocks has been confirmed by numerous studies. Due to the pore-proximity effect, the confined behavior of fluids in nanopores differs significantly from that observed in PVT cell. Currently CO2 huff-and-puff has been used to unlock the tight and shale reservoirs. Because of the high adsorption selectivity of CO2, after the injection of CO2, the original fluid density and composition of hydrocarbons in nanopores has been changed. In this paper, the PR-SLD model is applied to investigate the confined behavior of pure CO2/hydrocarbon fluids and their mixtures in nanopores. The Lee’s partially integrated 10-4 potential model is used to represent the solid-fluid interaction. For mixtures, a group contribution method is used to estimate the binary interaction parameters of CO2/hydrocarbon mixture. Thereafter, from the results of density distribution across the nanopore, the adsorption amount of fluids can be derived. Based on this model, a prediction process for the behavior of pure CO2 and hydrocarbon fluids (of methane and ethane) and their mixtures is performed. Results indicate that the adsorption selectivity of CO2 is much higher than CH4 and C2H6. And the density of pure CO2 in nanopores is higher than that of CH4 and C2H6. For binary mixture, because of the difference of interaction energy, the mole fraction of CO2 molecular is gradually increased from pore center to pore surface, and that of the hydrocarbon molecular is reduced from pore center to pore surface. The composition difference between bulk fluids and adsorbed fluids of CO2-C2H6 mixture is lower than that of CO2-CH4 mixture. For ternary mixture, the mole fractions of CO2 and C2H6 are always increasing from pore center to pore surface, and the mole fraction of CH4 is decreased from pore center to pore surface. Compared the original pure hydrocarbon mixtures, the addition of CO2 further increases the density of bulk fluids and adsorbed fluids. This study sheds some important insights for the behavior of confined fluids in nanopores and provides sound guidelines for the application of CO2 huff and puff in tight and shale reservoirs.
A method has been developed for the analysis of pressure falloff data following a single-stage treatment in a multi-stage fracture stimulation. The basic premise is that the greater the permeability contacted by the fracture stimulation, the greater the rate of pressure falloff will be. This can be done with as little as 15 minutes of falloff data, but with a “zipper” style completion, the surface pressure falloff of a given fracture stage may be monitored for several hours for no incremental cost while an offset well on the same pad is being stimulated. The initial falloff data is collected well before fracture closure, so proppant is not yet a factor – the pressure decay is influenced by the total fracture system of that stage.
This analysis has been performed on approximately 30 wells, each with about 20 stages, including two wells equipped with fiber optic sensing. The pressure decay follows a straight line on a plot of pressure versus logarithm of time. The slope of that line is the decay exponent, and a large exponent is indicative of greater connected permeability or fracture complexity.
The development of this technique is in its early phases, but thus far a good correlation has been observed between the pressure decay exponent and microseismic activity, as well as between pressure decay and the Young's modulus of the rock being stimulated. In a multi-cluster “plug and perf” completion equipped with fiber optic cable, a positive correlation was observed with the number of clusters being treated. When the same hydraulic fracture stimulation was executed in similar rock types, very consistent results were obtained, suggesting a valid and repeatable relationship. The final validation of this technique will be possible when compared against production logging results.
The prospect of a low cost, or even free, analytical technique in an environment where anything beyond a gamma ray curve is often a luxury, is particularly exciting. This assessment technique could be used for optimization of perforation cluster design and location, landing zone, and fracturing fluid optimization. The authors invite other operators to try this technique and discuss their observations.
Recent studies have indicated that Huff-n-Puff (HNP) gas injection has the potential to recover an additional 30-70% oil from multi-fractured horizontal wells in shale reservoirs. Nonetheless, this technique is very sensitive to production constraints and is impacted by uncertainty related to measurement quality (particularly frequency and resolution), and lack of constraining data. In this paper, a Bayesian workflow is provided to optimize the HNP process under uncertainty using a Duvernay shale well as an example.
Compositional simulations are conducted which incorporate a tuned PVT model and a set of measured cyclic injection/compaction pressure-sensitive permeability data. Markov chain Monte Carlo (McMC) is used to estimate the posterior distributions of the model uncertain variables by matching the primary production data. The McMC process is accelerated by employing an accurate proxy model (kriging) which is updated using a highly adaptive sampling algorithm. Gaussian Processes are then used to optimize the HNP control variables by maximizing the lower confidence interval (μ-σ) of cumulative oil production (after 10 years) across a fixed ensemble of uncertain variables sampled from posterior distributions.
The uncertain variable space includes several parameters representing reservoir and fracture properties. The posterior distributions for some parameters, such as primary fracture permeability and effective half-length, are narrower, while wider distributions are obtained for other parameters. The results indicate that the impact of uncertain variables on HNP performance is nonlinear. Some uncertain variables (such as molecular diffusion) that do not show strong sensitivity during the primary production strongly impact gas injection HNP performance. The results of optimization under uncertainty confirm that the lower confidence interval of cumulative oil production can be maximized by an injection time of around 1.5 months, a production time of around 2.5 months, and very short soaking times. In addition, a maximum injection rate and a flowing bottomhole pressure around the bubble point are required to ensure maximum incremental recovery. Analysis of the objective function surface highlights some other sets of production constraints with competitive results. Finally, the optimal set of production constraints, in combination with an ensemble of uncertain variables, results in a median HNP cumulative oil production that is 30% greater than that for primary production.
The application of a Bayesian framework for optimizing the HNP performance in a real shale reservoir is introduced for the first time. This work provides practical guidelines for the efficient application of advanced machine learning techniques for optimization under uncertainty, resulting in better decision making.