Lamberghini, Lucia (Pan American Energy) | Parra, Daniel (Pan American Energy) | Alonso, Ezequiel (Pan American Energy) | Sorenson, Federico (Pan American Energy) | Espina, Cristian (Pan American Energy) | Viglione, Lucas E. (DeGolyer and MacNaughton) | Clem, Daniel (DeGolyer and MacNaughton) | Ilk, Dilhan (DeGolyer and MacNaughton)
This study presents a case study example from a tight gas reservoir system in the Neuquén Basin, Argentina. Performance based-reservoir characterization is the main premise of this work integrating multiple data sources (geology, petrophysics, completions and production). For well performance analysis, an integrated procedure, which included production diagnostics, decline curve analysis and model-based analysis, was followed. Using the results from well performance analysis of 41 wells, multi-well numerical simulations were performed to investigate future development scenarios - particularly in the context of well configuration (multi-fracture horizontal versus vertical wells) and well spacing (various cases were considered).
The Lindero Atravesado block, operated by Pan American Energy (PAE), is located in the Neuquén basin in Argentina, 1,000 km from the city of Buenos Aires city and 30 km from the city of Neuquén (Fig. 1). The area comprises 530 km2 (approximately 131,000 acres) and has a total of 229 wells drilled, with 94 of these wells targeting the Grupo Cuyo tight gas formations of interest; Lajas (main target) and Punta Rosada (secondary target). The rock charge occurred vertically and was generated by the same energy of the generation and expulsion of the hydrocarbon from the source rock, Los Molles, which was deposited in a marine platform environment and is in direct contact with the multiple tight stacked reservoirs of Lajas and Punta Rosada formations. A stratigraphic column of Lindero Atravesado block is illustrated in Fig. 2.
Production started in the block in February 1974 from conventional formations, Sierras Blancas, Quintuco and Lotena reservoirs, which are shallower than the Grupo Cuyo tight sands formations. Since then, a total of 133 development wells targeting these reservoirs were drilled to develop the area, reaching a peak gas production of 5.77 MMm3/d (approximately 204 MMscfd) in November 1989. By 2012, the field was in a mature stage of development with very limited additional drilling activity or activity.
Production trends from hydraulically fractured reservoirs show impediments of gas flow via the partially invaded fracturing fluid into the formation due to viscous and capillary forces and formation damage due to clay swelling. The occurrence of this clay swelling on the surfaces of hydraulic fractures (HFs) highly depends on the fracture network complexities which can be predicted using a geomechanical model. Clay-swelling-induced conductivity damage is primarily a function of rock mineralogy, fracturing fluid composition, formation brine salinity, and fracturing fluid access to a realistic complex fracture network. In this work, we introduce a mechanistic approach to model clay swelling and water imbibition in various rock mineralogies including the Barnett (clay-rich), Eagle Ford (calcite-rich), and Marcellus Shales. Clay swelling counteractively influences production through favorable water-gas replacement and undesirable permeability reduction. We used a coupled multi-phase reactive-transport simulator to comprehensively model this process. Here, the main clay swelling mechanisms are denoted as the ion hydration and the expansion of the electrostatic double layer. We used the calculated volume expansion of clay materials exposed on the fracture surface to modify the fracture and matrix permeabilities. We acquired a complex fracture network mechanistically using a well-established poro-elastic cohesive zone model integrated with a wellbore model and an implicit intersection model. Subsequently, we developed an embedded discrete fracture model (EDFM) in order to evaluate the production performance of this complex HF network after clay swelling damage and water imbibition.
The simulation results indicate that the degree of clay swelling varies in different shale formations. On the basis of the clay content and the mineralogies considered in this work, we observed a significant expansion in electrostatic double layer for the Barnett Shale following the fracturing fluid injection. Nevertheless, this effect was lower in the Eagle Ford and Marcellus Shales. The presented approach provides the capability to mechanistically model the clay stability and water imbibition to approximate their impact on the production performance.
Jin, Lu (Energy & Environmental Research Center) | Hawthorne, Steven (Energy & Environmental Research Center) | Sorensen, James (Energy & Environmental Research Center) | Pekot, Lawrence (Energy & Environmental Research Center) | Kurz, Bethany (Energy & Environmental Research Center) | Smith, Steven (Energy & Environmental Research Center) | Heebink, Loreal (Energy & Environmental Research Center) | Bosshart, Nicholas (Energy & Environmental Research Center) | Torres, Jose (Energy & Environmental Research Center) | Dalkhaa, Chantsalmaa (Energy & Environmental Research Center) | Gorecki, Charles (Energy & Environmental Research Center) | Steadman, Edward (Energy & Environmental Research Center) | Harju, John (Energy & Environmental Research Center)
The Bakken petroleum system is one of the largest unconventional plays in the United States, with over 10,000 wells drilled in the past 10 years. The main target of this drilling has been two of the non-shale low-permeability units: the Middle Bakken Member and the Three Forks Formation. Although well logs and core data show that there is significant oil content in the two shale members—the Upper and Lower Bakken, the oil transport behavior in these source rocks is still not well understood.
The Energy & Environmental Research Center (EERC) conducted a series of experiments to investigate the rock properties of the two shale members, how fluids flow through them, and how flow may be induced. Twenty shale cores were collected (eight Upper Bakken samples and twelve Lower Bakken samples) from six wells in three North Dakota counties to ensure the samples were representative of the shales in the most productive areas of the Bakken. Six primary mineralogical components were detected in the samples through x-ray diffraction (XRD) analysis. High-pressure mercury injection tests showed that pore throat radii are less than 10 nanometers for most pores in both the Upper and Lower Bakken samples. Such small pore sizes yield high capillary pressure in the rock and make fluid flow difficult. Total organic carbon (TOC) content was measured, and kerogen was characterized by Rock- Eval/TOC pyrolysis, which indicated considerable TOC present (10–15 wt%) in the shales. However, oil and gas are difficult to mobilize from organic matter using conventional methods.
Field experience has shown that hydrocarbon extraction with supercritical CO2 is effective for extracting hydrocarbons (up to C20+) from conventional reservoirs. Additionally, laboratory experiments indicated that supercritical CO2 interacts with the oil associated with the organics, solvating the oil so that it can be extracted at reservoir temperature and pressure. A systematic experimental procedure was carried out to reveal the potential for extracting hydrocarbons from the shale samples under typical Bakken reservoir conditions (e.g., 5000 psi and 230°F). Results from 20 samples showed that supercritical CO2 enables extraction of a considerable portion (15%– 65%) of the hydrocarbons from the Bakken shales within 24 hours. The results may be used to improve modeling and forecasting the effects of CO2 enhanced oil recovery (EOR) and suggest the possibility for increasing ultimate recovery, and possibly CO2 storage, in some areas of the Bakken Formation.
It is better to know what you don’t know than not know anything at all. For operators, this means using data analytics to understand shortcomings and successes in their own operations as well as competitors. Unfortunately, public data sources aren’t always maintained to the same standards as internal data, making field analysis difficult and accurate recommendations inconsistent or impossible. Leveraging multidisciplinary data analytics from raw public data such as digital well logs and production and completion data can help deliver necessary insights to understanding key successes and shortcomings of unconventional plays.
A case study of the Wattenberg field will be presented in this session, demonstrating why public data cannot be used in its raw format and the exponential value gained from a cross-discipline analytical process. Within the field, four geological horizons are targeted through horizontal drilling – the Niobrara (A-C Chalks) and Codell formations. The Niobrara consists of alternating chalk and marl units, whereas the Codell Formation consists of a clay-rich sandstone; both were deposited within the Interior Cretaceous Seaway. The petroleum system is overpressured and is believed to be self-sourced from the organic-rich marl intervals.
This study analyzed 1,100 digital well logs to generate a surface-based geological model that delineates where horizontal wells were drilled. In addition, completion and production data from over 4,500 wells were compiled, with type curves generated based on sub-region within the field, operator and vintage to normalize for geological variabilities. Highlights of this work include: geological parameters for optimal targets; differing estimated ultimate recoveries (EUR) on a lateral foot basis as operators transition away from the core of the play; optimal completion design; and changes in wellhead liquids percentages across the play. Results can be directly traced to conclusions such as higher proppant loading and longer well lateral lengths yield materially better well performance. In general, data accessed through public sources allows for larger sample sizes; however, it’s through a technically-sound methodology that the data can be analyzed at a granular level, illustrating the effectiveness of using a multidisciplinary approach.
Modeling of the petrophysical properties in tight carbonate reservoirs is important for better characterization of reservoir quality. This study aims to investigate the relationship between reservoir properties (porosity and permeability) and ultrasonic P-wave velocity of carbonate samples from different sedimentary lithofacies. In this context, different carbonate lithofacies of a wide range of rock lithification were selected in order to examine and model their reservoir properties and ultrasonic velocity under multiphase confining pressure simulation. The carbonate lithofacies were selected from the Miocene Dam Formation carbonates in Eastern Saudi Arabia. Three lithofacies were selected for this purpose, those are, (I) The stromatolite boundstone, (II) The burrowed wackestone, and (III) The quartz fossiliferous wackestone-packstone. The nature and field setting of these lithofacies reflects wide range of sedimentary texture, lithification, and therefore reservoir properties. Besides porosity and permeability models, ultrasonic P-wave velocity models were generated for each carbonate lithofacies under a wide range of confining pressure. In addition to the petrography, further investigations were used using SEM, and micro-CT scan imaging in order to demonstrate the effect of lithofacies patterns on the petrophysical properties. The petrographic analysis, SEM, and micro-CT scan revealed that the sedimentary fabric, composition, internal structure, and diagenesis have influenced the porosity and permeability patterns of the different lithofacies, and therefore the ultrasonic p-wave velocity. The homogeneous composition of the burrowed wackestone in terms of grain sorting, shape and stability clarify its steady petrophysical relationships, however, its weakest lithification has affected this relationship in higher confining pressure (3MPa and 6MPa). In contrast, the highly lithified stromatolite boundstone and quartz fossiliferous wackestone-packstone revealed better relationship with higher confining pressure. This study showed that integration of petrophysical properties and ultrasonic P-wave velocity under multiphase confining pressure can provide guides that might lead to better understanding and prediction of the tightness of carbonate reservoir with less degrees of uncertainty.
Permeability of fracture filled or not with proppants under in situ effective stress condition is a key parameter for optimization of proppant recipe used in hydraulic fracturing job. Furthermore, fracture permeability impacts results of numerical modelling of Stimulated Rock Volume geometry and prediction of its evolution during gas or oil production. Total E&P has implemented recently an experimental set-up allowing creating shear fracture on a cylindrical plug of shale rock in a conventional triaxial and measuring permeability tests to water and to gas according a specific protocol simulating the change of in situ effective stress and pressure gradient in the near and far field of perforation. After creating shear fracture, we filled the fracture with proppants used in frac job, and we performed measurements of permeability to gas under various effective stress condition. The hydraulic aperture of fracture with different concentration of proppants is then determined.
When the fracture is filled with proppants, significant decrease (-20%) of the fracture hydraulic opening is clearly observed when confining pressure increases from 10 to 300 bars for the fracture filled with 1 layer of 30/50 mesh ceramic proppants. Compared to the fracture without proppants, the relative decrease of hydraulic opening (e/e0) is 5 times less for fracture with one layer of proppants. When the fracture was filled with 2 layers of same proppants, the amplitude of change of fracture opening is 3 times less compared to that with one layer of proppants. Embedment of proppants on fracture surface is observed. The fracture filled with natural sand is much more sensitive to confining pressure compared to the fracture filled with ceramic proppants 30/50 mesh.
We present in this paper the protocol, the calibration of the experimental set-up and the main results of fracture permeability to water and to gas obtained on a Vaca Muerta shale sample, filled with different quantities of proppants.
The potential of gas production is first and foremost determined by geochemical and petrophysical factors such as total organic carbon content, thermal maturity, porosity and permeability. However, the productivity is strongly dependent on the fracture network connectivity and permeability since the shale matrix has extremely low permeability, and acts as a seal to many conventional reservoirs. Many geomechanical parameters of shale control the hydraulic quality of stimulated reservoir volume (SRV) created by hydraulic fracturing.
In this study we present a geomechanical analysis workflow using microseismic focal mechanisms to investigate the dynamic response of the reservoir during and after stimulation. Focal mechanisms are derived using full waveform fitting techniques, and the ambiguity in identifying the true fracture plane is resolved by simply choosing the nodal plane that aligns with the developing hydraulic fractures. A global stress inversion of the fracture plane solutions is done to estimate the orientations and relative magnitudes of the principle stresses. Friction laws are then used to constrain for each event a suite of geomechanical parameters (failure potential, dilation tendency, and excess pore pressure) in order to identify fracture populations likely to control fluid flow, those that required different stimulation pressures in order to contribute to flow, and the mechanical conditions that favored out-of-zone growth and reactivation of geohazards. Additional observations, such as net wellbore pressure measurements and geophysical logs, are used to calibrate the model as well as to further understand the geological, geomechanical and treatment-related variables affecting the overall stimulated rock volume. The method is applied and discussed in the case of a microseismic event catalogue obtained during the stimulation of two horizontal wells landed in the Eagle Ford, where large variations in fracture patterns as well as the reactivation of a large macroscopic fault zone was observed.
The state of stress of the reservoir is one of the dominant factors controlling the reservoirs response to stimulation as well as the effectiveness of the treatment design. For instance, the orientation and magnitude of the maximum horizontal stress (SHmax) strongly affects the stimulated range of fracture orientations and in turn the geometry of the stimulated zone (i.e. localized versus distributed fracturing). The hydraulic horsepower, which takes into account the reservoir stress states and pressures, may be sufficient to stimulate parts of the reservoir with a specific state of stress, but any variations in the stress state can result in adverse effects such as damaging nearby wells (“frac hits”), out-of-zone growth, and large-magnitude earthquakes. Furthermore, the reservoir stress state can also impact the hydraulic conductivity of stimulated fractures (Barton et al., 1995).
Elastic properties such as Young’s modulus (E) and Poisson’s ratio (ν) are key controlling factors in the design of fracture stimulation in unconventional reservoirs such as gas/oil shales and tight gas sands. Well logs are capable of measuring parameters that in some cases can be transformed to static moduli for inputs into frac design and to infer mechanical heterogeneity and stratigraphy (at log resolution). Due to the high cost, plug availability, and the presence of natural fractures, it is not possible to conduct triaxal tests over 100’s of feet of core to adequately measure mechanical heterogeneity and stratigraphy. Here, we present results of Index methods of strength and modulus measurement over 100’s of feet of core. We have chosen to use two commercially-available options, the Impulse Hammer and the Leeb Rebound Hardness tester.
The Impulse Hammer drops a small hemispherical mass into the core. From a combination of Hertzian contacts and impulse, the Plain Strain modulus, E*= E/(1- ν2), can be determined from the force-time curve during the impact. Rebound Hardness is measured from the Leeb method in which a small ball is shot into the core and the ratio of the rebound velocity to impact velocity is measured, giving a hardness. It is possible to empirically transform this hardness to Unconfined Compressive Strength (UCS) or confined strength in some cases. Each individual measurement represents relatively-quick deformation of a small volume (~10 mm3) of material; leaving a small pit in the sample on the order of 10’s of microns. Measurement is considered a quasi-static measurement and the strain magnitudes are similar to those obtained when loading a plug sample in a conventional rock mechanics test.
Measurements are taken every cm, to assess fine-scale heterogeneity in unconventional reservoirs, along core lengths of many meters. This technique therefore spans small-scale up to large-scale and provides values that can be compared to more macroscopic measurement of heterogeneity and stratigraphy from well logs and into upscaling methods.
We propose azimuthal plane-wave destruction seismic diffraction imaging workflow (AzPWD) to efficiently emphasize small-scale features associated with subsurface discontinuities such as faults, channel edges, fracture swarms, etc., and to determine their orientation by properly accounting for edge diffraction phenomena. We apply the workflow to characterize unconventional tight-gas sand reservoir in the Cooper Basin in Western Australia. Channels and faults are emphasized in comparison to a conventional reflection image. Extracted orientations of edges provide valuable additional information, which can be used by the interpreter to locate finer scale features and distinguish them from noise.
The degree of natural fracture development in otherwise low-matrix-permeability unconventional reservoirs is a known controlling factor in oil and gas producibility. The design of a successful well for unconventional reservoir development can use the knowledge of subsurface discontinuities distribution. Unconventional reservoirs may exhibit high structural variability, which is hard to characterize using discrete wells network. 3D reflection seismology allows extracting additional information about the subsurface with significantly denser spatial sampling intervals. However, conventional images of the subsurface have low spatial resolution and are dominated by continuous and smooth reflections, which carry the information associated with only large-scale heterogeneities.
Diffraction imaging is capable of emphasizing small-scale features associated with subsurface discontinuities such as faults, channel edges, fracture swarms, pinch-outs, etc. (Fomel et al., 2007) in comparison to conventional reflection images, where those features are often masked by higher magnitude reflections (Klem-Musatov, 1994). Schoepp et al. (2015); Burnett et al. (2015); Sturzu et al. (2015); Tyiasning et al. (2016); Klokov et al. (2017); Merzlikin et al. (2017); de Ribet et al. (2017); Pelissier et al. (2017); Koltanovsky et al. (2017) illustrate this statement and employ diffraction images as a source of additional information for interpretation (Grasmueck et al., 2015). Diffraction phenomena consideration in 3D (Hoeber et al., 2010; Keller, 1962) allows to distinguish between tip and edge diffractions (Moser, 2011; Klokov et al., 2011; Bona and Pevzner, 2015). Alonaizi et al. (2013) produce D-volumes describing diffractivity distribution by computing semblance along diffraction travel-time hyperboloids and determine edge diffraction orientation by scanning along different azimuths. Merzlikin et al. (2016) propose azimuthal plane-wave destruction seismic diffraction imaging workflow (AzPWD). AzPWD workflow extends plane-wave destruction diffraction imaging framework (Fomel et al., 2007) to account for edge diffraction orientation and allows us to efficiently extract these orientations based on scanning of different azimuths.
Although material balance time was originally proposed by Blasingame and co-workers to model wells in boundary-dominated flow, later investigators proposed that logarithmic rate vs. material balance time lead to the same conclusions as logarithmic rate vs. time plots during transient flow for smoothly changing rates and are superior for identifying and analyzing data in boundary-dominated flow. However, application of this idea to field data often shows different curve shapes for the two methods of plotting during transient flow. The purpose of this study was to identify the cause of this discrepancy and to suggest an approach to flow regime identification in which we might have more confidence.
Our hypothesis was that the cumulative production used in calculating material balance time will not be consistent with the rates during transient flow for multi fracture horizontal wells (MFHW), completed in ultra-low permeability reservoirs, when (1) fracture damage was present (more common than not) and (2) pressure drawdown and production varied during the early months of a well’s production history. To test our hypothesis, we analyzed simulated production data in several ways: (1) production at constant BHP, no fracture damage; (2) variable BHP, no fracture damage, BHP values known as functions of time; (3) variable BHP, no fracture damage, BHP values unknown; (4) constant BHP with fracture damage; (5) variable BHP with fracture damage and known BHP; and (6) variable BHP with fracture damage and unknown BHP.
We found that our hypothesis was correct, and that, during transient linear and bilinear flow, the slopes of logarithmic plots of rate vs. time and rate vs. material balance time can be quite different. Linear and bilinear flow can be identified with a slope of (-1/2) and (-1/4) on logarithmic rate-time plot for smoothly decreasing rates. However, they can be delayed or disguised significantly on logarithmic rate-MBT plots with large fracture skin and varying BHP.
We conclude that a simple logarithmic rate-time plot leads to a more definitive identification and analysis of the transient flow regime in MHFWs than the logarithmic rate-material balance time plot. On the rate-time plot, transient linear and bilinear flow can be identified with a (-1/2) and a (-1/4) slope which may be delayed or distorted on the rate-material balance time plot. For boundary-dominated flow identification and analysis, the material-balance-time plot remains the preferred method.