Kornacki, Alan S. (Weatherford Laboratories Inc.) | Westrich, Joseph T. (Weatherford Laboratories Inc.) | Gong, Changrui (Apache Corporation) | Rodriguez, Lucia (Apache Corporation) | Etienne, Jeff S. (Apache Corporation)
It is difficult to obtain the end-members required to allocate commingled oil samples produced from a shale reservoir that has been hydraulically fractured – and then use those results to determine from which zones oil is actually being produced. We demonstrate that suitable end-member oils can be extracted from core plugs selected from a conventional core collected from a mudrock interval in the Wolfcamp Formation using water-based mud that consists of good oil-prone source-rock beds. End-member oils also can be extracted from most core plugs selected from an overlying conventional reservoir that contains oil expelled by the source-rock beds in the shale reservoir. Several standard geochemical source and maturity parameters confirm the results of a Hierarchal Cluster Analysis utilizing 24 HRGC peak-height ratios that shows the shale reservoir can be divided into five distinct zones that each contains a different type of oil generated at a different level of maturity by the kerogen organic facies present in that zone. HC fingerprinting results using >100 HRGC peak heights indicate that oil samples produced from two horizontal wells drilled near a stratigraphic boundary in the shale reservoir principally produce the kind of oil present in a distinct ≈40-foot zone immediately above those horizontal wellbores. An oil sample produced from a vertical well completed in the conventional reservoir produces a mixture of the type of oil extracted from a core plug selected in the lower part of that reservoir – plus the type of oil present in the same zone in the deeper shale reservoir that also accounts for most of the oil produced from the horizontal wells. We conclude that a significant amount of oil is not being produced from other zones in the shale reservoir, or from the upper part of the conventional reservoir. Implementing this technique has the potential to improve the effectiveness of hydraulic fracture programs by identifying the specific reservoir intervals from which oil is produced after a horizontal well is completed, and it increases the value of legacy and new conventional cores collected from shale-oil reservoirs using water-based mud.
The Lower Triassic Montney formation is one of western Canada’s most lucrative unconventional hydrocarbon plays. With a thickness in excess of 300m, aerial extent covering approximately 130,000 km2, the National Energy Board of Canada estimates the Montney Formation contains marketable resources of 449 TCF of gas, 14.5 billion barrels of NGL’s and 1.13 billion barrels of oil.
The Lower Montney Formation in the Peace River Arch region of northwest Alberta and northeast British Columbia was one of the first areas of the Montney exploited for hydrocarbon beginning in 1980. The initial discoveries were made in turbidite channel complexes, using 3D seismic and conventional vertical drilling and completion methods. The reservoir are highly porous (15%) and permeable (1-5 md) fine grained channel sands. The channel facies is very prolific but limited in aerial extent.
Beginning in 2006, with the arrival of horizontal drilling and multi-frac completion technology, it now became possible to economically exploit the thicker more aerial extensive, fan portion of the turbidite complex. These rocks are fine grained laminated siltstones with lower porosity (3-10%) and lower permeability (less than 0.1 md) but, extremely thick (30-70 m).
Detailed geological mapping reveals that the deposition and the thickness of the fan is structurally controlled by underlying Paleozoic fault systems that define the Fort St. John graben complex. Isotope geochemistry of the hydrocarbons reveals that they are internally sourced, and their distribution display a normal thermal maturity distribution with dry gas being found in the deepest portion of the basin, transitioning to wet gas condensate and light oil concentrated along the basin margin. The fan can be mapped into 3 cycles ranging in thickness from15m to 25m. Each cycle has a preferred landing zone which contains more highly porous and permeable rock which local operators target. Thru time the more proximal portion of the fan has been exploited but with the further advancement of drilling and completion techniques, more distal portions of the fan complex are economically being exploited.
An anomalous flowrate feature (often a "hump" or even a "spike") is characteristically observed at early-times during flowback performance in multi-fractured horizontal wells (MFHW) completed in ultra-low permeability (shale) reservoirs prior to the onset of a characteristic reservoir flow regime (i.e., linear or bilinear flow). The flowrate feature tends to occur in all fluid phases and this feature is thought to be attributed to the "clean-up" behavior following well stimulation and/or the phase behavior of the fluid as it flows along the well path.
The guiding principle of this work is that this anomalous flowrate feature can be represented by decaying skin effects, a changing wellbore storage effect, or a combination of both decaying skin effects and changing wellbore storage effects. The goal of this work is to provide a proof-of-concept which considers the simplified case of a vertical well with a single vertical fracture to develop a series of time-dependent skin and wellbore storage models that can effectively be used to characterize the early-time flowrate behavior observed in practice. For this study, we forced a constant wellbore flowing pressure constraint, and while we recognize that this constraint is not truly met in practice, we believe that this approach can serve as a base model for diagnostics/interpretative analyses.
Based on the work developed by Fair (1981) and Larsen and Kviljo (1990), our procedure is to couple a series of time-dependent wellbore storage and skin effect models with a set of "power law" reservoir flow models (i.e., linear flow, bilinear flow and a generalized power-law flow model). Specifically, we combine the time-dependent wellbore storage and skin effect models with the constant rate solution reservoir flow models, then apply the convolution integral to produce the constant pressure condition — all in the Laplace domain. In order to generate various scenarios of production performance, we use the Gaver-Wynn-Rho Algorithm implemented in Mathematica to numerically invert the Laplace domain solutions into the real time domain. A generalized workflow is provided to demonstrate the addition of time-dependent wellbore storage and skin effects to any prescribed reservoir model.
Using the various wellbore storage and skin time-dependent models proposed in this work, we observe that each of these models, individually and in combination, provide behavior indicative of early-time flowrates observed in the field. In short, we demonstrate that each time-dependent model has unique characteristics, which could, in concept, allow for characterization of flow behavior in the fracture prior to the onset of an undistorted "reservoir" flow geometry (i.e., formation linear or bilinear flow).
McAndrew, James (Air Liquide Delaware Research and Technology Center) | Cisternas, Pablo (Air Liquide Delaware Research and Technology Center) | Pruvot, Antoine (Air Liquide Delaware Research and Technology Center) | Kong, Xianhui (Air Liquide Delaware Research and Technology Center) | Tong, Songyang (Air Liquide Delaware Research and Technology Center)
Foam fracturing fluids provide a means to reduce water consumption by replacing most of the water used in a fracturing job by industrial gases as N2 or CO2. Foams also have the potential to provide increased well productivity, based on improved proppant placement, faster fluid cleanup and reduced reservoir damage. We have built a new experimental apparatus to measure the transport of proppant by foams under high pressure (2000-3000psi) and developed a computational fluid dynamics (CFD) model to enable extension of the results to real fracture geometries. Our laboratory results show that proppant transport by water is dominated by bed transport, whereas transport by foams is primarily in suspension. Both simple foams (without thickener added) and thickened foams transport proppant substantially further than water.
Oil and gas sources that rely on hydraulic fracturing have remained active even in the face of dramatically lower prices. This is further evidence, if it were needed, of the tremendous economic significance of fracturing technology. Nevertheless, improvements in the productivity of fractured wells, and management of their environmental impact, are both urgently needed in order to continue to meet stringent cost per barrel (or per cubic foot) requirements. The current rigorous market is the right time to implement technological improvements that can lead to higher production per well.
In parallel with productivity improvement, it is necessary to manage the overall “footprint” per well, i.e. the consumption of resources, most notably water and proppant. Progress in these areas also reduces truck traffic to and from the wellsite. These advances are win-win solutions, reducing producer cost and environmental impact in a single move.
The idea of using foam fracturing fluids to reduce water consumption and disposal requirements, by replacing most of the water used in a fracturing job by N2 or CO2, is not new. However, foam fluid implementation is generally limited to underpressured or highly water-sensitive reservoirs. This is largely attributable to logistical issues in N2 and CO2 supply, lack of availability of pumping equipment and capable crews, and perceived higher cost. In this article we will briefly discuss how foam fracturing can reduce total fluid consumption, i.e. not simply replacing water by N2 or CO2, then introduce our new laboratory apparatus designed measure the transport of proppant by various fracturing fluids, and finally review our initial results comparing foams with water. These results support the proposal that foam fracturing fluids can provide improved proppant placement and therefore improve well productivity, as well as reduced water consumption.
Abstract and Introduction
Advanced hydraulic fracturing technologies and increasing oil prices gave an impulse to shale oil resources development in the U.S. in the past decade. Recent drop in energy prices spurred technological improvements and research in understanding of the resource to enable efficient production in changing economic conditions. Exploring the variability in individual well production and identifying the relationships between well performance and major input factors, such as geology and completion practices are essential to 1) describe technological progress and success; 2) reveal areas for further improvements; and 3) predict production for all locations of future wells. The latter is particularly important if one wants to make projections about the future shale resource development.
In this study, we focus our attention on the Bakken shale play recognizing two producing horizons, Middle Bakken and Three Forks. We apply various statistical and machine learning techniques that give insight into the factors that drive productivity. The calibrated models help us extend that knowledge to estimate productivity in undrilled locations. First, we study current productivity trends, explore the existence of functional relationships between operator’s completion practices, geologic rock properties, time and oil production. Then, machine learning techniques are used to investigate how data availability affects predictive power of the models. Finally, we decide on the most suitable model to make predictions about well productivity for all locations of future wells.
In our analysis 1) we find evidence of spatial and temporal heterogeneity in the data, which indicates that analysis of the entire Bakken Formation for all time periods could lead to suboptimal results; 2) we account for the presence of vertical variability of geologic resources, which, if neglected, could lead to erroneous conclusions; 3) we use statistical methods such as model-based recursive partitioning to identify spatial and temporal productivity regions, which not only shed light on the interplay between oil production, geologic resources, and operators’ decisions, but also aid us in conducting an economic analysis for creating potential future oil-production scenarios; and 4) we use machine learning technique to predict productivity in undrilled locations.
Hydraulic fracturing provides the petroleum industry with unconventional means for extracting hydrocarbons, but the method can be made more efficient to enable better production by answering some not so simple questions: where did the injection fluid go and how much of the volume was propped? Ground-based controlled source electromagnetics (CSEM) provides a means to generate, record, and interpret electromagnetic signals responding to temporal changes in the subsurface electrical conductivity, using instruments deployed on the surface. Since the injection of fluid into rock for hydraulic fracturing alters the conductivity of the formation in the vicinity of the wellbore, CSEM is a geophysical technique that is well-suited for monitoring completions. In this paper we show results from two CSEM surveys. The first CSEM survey acquired data from the Anadarko basin during hydraulic fracture operations from a lateral well while the second CSEM survey is based on data obtained during the completions of a vertical well in the Delaware basin/ Northwest shelf. Both of the surveys provided unique challenges. The data interpretations from each survey help to influence future completion strategies by providing valuable information to the well development teams, for example fracture asymmetry, fracture half lengths, and unexpected fracture behavior.
Ground-based controlled-source electromagnetics (CSEM) is an emerging geophysical technique that can be used to monitor the movement and extent of injection fluid during hydraulic fracture operations (1,2). The electromagnetic response of a conductive zone containing fluid to its energization by a surface-deployed CSEM source is dependent upon the electrical conductivity difference between the fluid-rich zone and the background geological formation (3). The fact that CSEM is sensitive to the electrical conductivity of the fluid allows properties such as hydraulic fracture half-length and fracture asymmetry to be studied dynamically. Additionally, it has been shown that the presence of a steel well casing increases the sensitivity to deep targets (4,5). We now present two case studies, the first involving a horizontal well in the Anadarko Basin and the second involving a vertical well in the Delaware basin Northwest shelf.
This study was based on core plugs from multiple formations within the Appalachian Basin. Data for integration include bulk rock porosity, LECO TOC measurements, and multi-scale SEM-based digital rock analysis (DRA) to show a comprehensive view of the rock properties within each of the five different formations; Rhinestreet, Genesee, Burket, Hamilton, and Marcellus. Applications include upscaling micro-pore scale DRA reservoir properties to GRI crushed sample and bulk rock TOC, leading to improved reservoir quality estimation. The primary goal was to provide an in depth look at porosity-typing at SEM-scale to better evaluate depths for further analysis and “sweet spots”.
We observed good agreement between bulk rock data and digital rock analysis with respect to organic matter volume. Multiple resolution imaging was key in matching porosity visible in SEM with porosity from GRI crushed rock analysis. DRA quantifies the effective porosity, which can be used for better reservoir estimation, while the remaining porosity not visible with SEM is primarily occupied by clay-bound water. While the GRI method measures total porosity, only DRA can give separate results for organic porosity and intergranular porosity.
With the advance of technology in multi-fractured horizontal wells, production from shale plays has improved significantly, making them become viable sources of oil and gas production. While these unconventional resources bring great benefits to the industry, production prediction from these wells has proven to be challenging.
In this paper, our goal is to present a statistical method for the prediction of production from liquid-rich shale and other ultra-low permeability reservoirs. This method starts from the learning process from multiple wells with sufficiently long production histories. Functional principal component analysis (FPCA) is applied to extract key features of production patterns. Then, principal component functions obtained from the training production data set are used as the basis to construct a linear production model from which we can predict production from other wells. Multiple test wells are selected for validation to compare predicted results to true production data. The approach in this paper is driven by production data and it has several advantages over empirical models used in decline curve analysis.
Production forecasting has significant consequences in investment decision making and is also a major component of reserves estimation required for reports to regulatory agencies. With the aid of the approaches proposed in this work, we can improve reserves estimation, field management and evaluation of project economics.
Unconventional resources are becoming an increasingly significant source of hydrocarbon supply in the global energy market (Holditch 2010). The great economic success in developing these resources have been largely driven by the advances in technologies such as multistage horizontal well drilling and multistage fracturing. However, the lack of sufficient knowledge in physical properties and the physics controlling production from shale formation limits our ability to model and forecast with confidence production and reserves from these important plays. Advances in the industry’s ability to forecast future production more accurately impacts financial forecasts, perceived asset values and accuracy of reserves disclosed to the public.
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).
Key field scale observations from comparative analysis of microseismic data saw significant differences in microseismic focal mechanisms, event distribution and event magnitudes that indicated more complex fracture systems and improved stimulation of natural fracture networks occurred in ball-activated completion systems. Commensurate disparities in production studies that compared the two systems further demonstrated production increases greater than 40% BOEPD. These observations from numerous basins give rise to the question of why these events occur. To answer that question, this paper explores the importance of timing of the applied stress in the context of a dynamic stress field through comparison of Ball Activated and Plug-n-Perforate completion systems and quantifies the geomechanical impacts of the operations of each system type. These are done using a FEM modeling software. Further modeling and investigation of stress prior to any fracturing operations related to the two completions systems are explored in terms of their effects on energy transmitted to the formation and consequently its effect on fracture initiation and propagation potentials using a software platform that employs the Material Point Method (MPM).
Data used in the first model is from the Bakken Shale in North Dakota. Modeling of hydraulic stimulation parameters, fracture dimensions and the resulting modified stress state are performed from two simulations. Geomechanical modeling of the initial reservoir stress state and elastic properties of the rocks are considered in combination to determine the extent of shear fracture development through time. Dynamic stress state simulation for each system begins with a model of in-situ stress conditions and rock mechanical properties prior to hydraulic stimulation. The stress state is then perturbed by placement of multiple hydraulic fractures in adjacent stages of the well. The altered stress state around each fracture, commonly referred to as a stress shadow is then modeled throughout the rock volume using energy balance criteria related to net pressure decline. The effects of stimulation timing inherent to each completion system are considered and commensurate changes in elastic properties of the rocks in the altered stress state are quantified. The variation in intensity of the new stress state is then mappable as stress shadows for each hydraulic fracture. Longer-term considerations of each energy state are compared to determine the potential of the rock mass to undergo inelastic strain in the form of shear fractures and tip extension once new volume is no longer being added to the system.
This technique further enables fracture design considerations such as optimal timing of treatment delivery and determination of stage spacing where shear fracture development is promoted without impinging on fracture development in successive stages.