Previous works have presented the results of successful simulations of fluid injection into naturally fractured shale using a Discrete Element Model (DEM). The simulations included coupled fluid flow-deformation analysis, failure type and extent calculations, as well as a series of parametric analyses. The parameters investigated included: 1) injection rate and its effect on the overall fracturing results, and 2) fluid viscosity, which had a significant influence on the ratio of tensile (mode 1) failure versus shear failure.
With the huge growth in the stimulation of naturally fractured formations such as fractured shales, it is clear that the industry needs new hydraulic fracturing simulation tools beyond the limits imposed by pseudo3D fracturing models. DEMs, in which both matrix block behavior and fracture behavior are explicitly modeled, offer one option for the specific modeling of hydraulic fracture creation and growth in a naturally fractured formation without, for example, the assumption of bi-planar fracture growth.
In this paper, we extend the previous works to quantify, for fractured shale gas plays, the effect of stress orientation, fluid viscosity, and rock mechanical properties in terms of changes in fracture aperture and transmissivity. Changes in fracture transmissivity directly correlate with improvements in well productivity - the primary goal of the stimulation.
The results of the study provide a means to improve shale completions by understanding the effects of the DFN orientation relative to the stress field, fluid viscosity, and rock mechanical properties on changes in fracture aperture, fracture transmissivity, and formation effective permeability, which directly relate to well productivity.
A method is presented that integrates a triple porosity model with sonic, density and resistivity logs for evaluation of tight gas formations. The interpretation takes into account results from petrographic work in the Western Canada Sedimentary Basin (WCSB), which indicates that tight rocks are comprised of different types of pores including (i) intergranular, (ii) slot + microfractures, and (iii) isolated non-effective porosities.
Seismic data are powerful in the exploration and production domains but a method that integrates seismic velocities and the observed triple porosity petrographic characteristics of tight gas formations is not available. This paper provides the theoretical foundation and development of equations for this integration along with examples using real data from tight gas formations in the WCSB.
The proposed method provides estimates of inter-well formation resistivity, porosity and water saturation to obtain estimates of original gas in place. The comparison between resistivity from seismic velocities and resistivity from well logs is good with a strong statistical effectiveness. Under favorable conditions, the partition between effective and non-effective porosity might be estimated.
The proposed methodology has significant potential for application in tight gas formations of the WCSB. The method can probably be extended to other regions around the world, which possess tight gas formations with similar characteristics to the ones described in this work.
During the life of producing wells, there comes a time when the well approaches its economic viability as a producing well. If the reservoir potential is sufficient to support the expenditure, many wells are candidates for recompletion, reperforation, or restimulation. This type of focus on the Barnett shale began in the late 1990s. Drilling activity dramatically increased during the ensuing years and now there are more than 14,000 wells that have been drilled, most of which are producing wells. A lot of these wells are potential candidates for restimulation (refrac) because their production rates have declined but still have significant reservoir potential. The completion techniques deployed in the Barnett evolved over time to where many wells have dozensof perforation clusters and hundreds of individual perforations. Generally, refracsare ineffectual unless the perforations can be temporarily isolated so that the energy of the subsequent fracturing treatment can be focused on individual portions of the reservoir. Additionally,refrac candidate wells often contain challenging wellbore environments that further complicate the ability to successfully refrac the wells. The use of biodegradable particulates to facilitate the temporary diversion and concentration of frac energy has increased the success of restimulation.
This paper discusses the recent development of techniques and materials being used in refracturing operations. Included are discussions of laboratory results of new and novel materials, along with case histories of refrac wells demonstrating application of such materials and techniques.
Analytical models to predict the performance of thermal recovery processes are useful tools for preliminary forecasting purposes and sensitivity studies and provide a better insight than simulation models into the physics of thermal processes. Classical models such as Marx and Lagerheim (1959), Willman (1961) and Farouq Ali (1971) are used extensively for steam-flood performance prediction. Several studies have been conducted to develop the theory for the estimation of the radius of the heated zone. This radius is important for computing the volume of recoverable oil, as well as to determine well spacing in steamflooding and cyclic steam stimulation. This work presents an analytical model to estimate the radius of heated zone in either conductive or conductive-convective heat transfer mechanisms, which mainly occur in Steam-Assisted Gravity Drainage (SAGD) and Cyclic Steam Stimulation (CSS) respectively. The heat flow equation was combined with mass and momentum convective transport equations in a porous medium, in an effort to correlate the temperature front velocity to the steam advancing front velocity. As the saturation front velocity is known from classical Buckley-Leverett transport equation, at each instant we investigated the transport distance of the heat front in a radial homogenous reservoir. The theoretical model takes gravity into account, but neglects the capillarity, and there is no longer the assumption of piston-like steam drive. CMG-STARS thermal simulation and COMSOL Multiphysics are used to compare and verify analytical model results. The improved model is superior to previous models used to calculate the radius of heated zone and the analytical results are in good agreement with the simulation results.
Understanding the morphology and growth of hydraulic fractures is essential in the development of unconventional tight gas and shale gas resources. In this paper, we report the use of a computational tool consisting of geologic data representation, geomechanical modeling and multiphase flow simulations to predict reservoir performance. A finite element-based geomechanical module is interfaced with a control-volume finite element discrete-fracture reservoir simulator. The hydraulic fracture geometry is generated by following the fluid injection pathway in the existing fracture network and in the matrix. The permeabilities in the fractures and in the matrix are adjusted dynamically in the simulator based on coupling of the geomechanical and flow attributes. The approach described is applicable to tight gas and shale gas reservoirs. There are over 3700 active wells in the Greater Natural Buttes field in Uinta Basin, Utah with a cumulative production of 1.8 trillion cubic feet of gas. Most of this production comes from several tight gas formations (matrix permeability of less than 0.01 md). The importance of hydraulic fractures and their interaction with natural fractures in shale gas reservoirs characterized by nano-Darcy permeabilities is well known. This methodology provides a means of mapping a complex (non-planar) network of hydraulically-activated/induced fractures. The initial conductivity distribution in the fractures and the relative permeability of the matrix along with the stress tensor and mechanical properties of the matrix and the fractures control the geometry of the hydraulic fracture system and the ultimate well performance. Prediction of complex hydraulic fracture morphology and injected water balance are important in the development of tight gas and shale gas resources. Multiphase flow simulations with complex fracture networks will help delineate complex effects such as water blocks, "permeability jail?? phenomena, etc.
Recovery factor (RF) is one of the most important parameters for economic justification in the petroleum industry. The RF is calculated from the ratio between expected ultimate recovery (EUR) and original oil-in-place (OOIP). However, many uncertainties exist in OOIP estimations, mainly due to inaccurate information about reservoir drainage area. In unconventional reservoirs, the estimations of drainage area and OOIP are more complicated due to unknown factors such as
fracture networks in the system, reservoir pressures, pore volumes, and in situ hydrocarbon properties.
We analyze here a case study of the Bakken Formation where the RF is still unclear and not frequently reported. RF for the Bakken formation from previous studies ranged from 0.7 - 50% (Price, 1984; Bohrer et al., 2008) derived from different approaches. In this paper, we use the material balance equation (MBE) approach to determine the deterministic (single value) and probabilistic (distribution) RF values in the Bakken Formation (the Antelope, Sanish and Parshall fields). We also perform a sensitivity analysis for the input parameters in our MBE approach.From the actual well production data, the EUR can be estimated from decline curve analysis (DCA) with knowledge of abandonment rate data. The DCA technique based on history data is used as forecast model for predicting the EUR and remaining volume. Once the RF and EUR are calculated, the OOIP value can be evaluated as the ratio between EUR and RF.
In the studied fields, RF result of the Parshall Field is highest because the produced gas-oil ratio is lower than the other two fields. From the EUR results, the high EUR area in the Antelope Field is located in the center of the field. For the Sanish and Parshall fields, high EUR areas are located in the east and west side, respectively. In addition, it appears that EUR has a direct relationship to the hydrocarbon pore volume per area (HCPV/area) distribution. For each field, the results of OOIP per well from the ratio between EUR and RF are presented, and these data can be compared with the volumetric OOIP for proper well spacing plan of future projects.
Many unconventional formations have large volumes of remaining oil still contained in the reservoir. Recovery Factor (RF) is such formations is difficult to assess mainly because the input data needed for calculations are poorly known and can have large uncertainties. Moreover, less production data in young fields cause uncertainties in estimations as well. For example, RF for the Bakken Formation from previous studies ranged from 0.7 - 50% (Price, 1984; Bohrer et al., 2008). We investigate here the Bakken Formation and analyze distributions of the RF given a possible range of the input parameters. We also analyze the effects of each of the input parameters on our calculations to determine RF sensitivity to different input values.
The Bakken Formation is a thin and deep unit from the Late Devonain to Early Mississippian (Meissner, 1978) covering 200,000 square miles (Houston et al., 2010) in the Williston Basin (Figure 1). The Williston Basin has large volumes of remaining oil still contained in the reservoir. The recovery of this remaining oil is related to the recovery factor (RF), thus affecting the expected net revenue of projects into the Bakken.
In unconventional reservoirs, such as tight sands or shale reservoirs, multi-stage hydraulic fracturing is required to make the reservoir productive. In such formations, the RF calculation is more complicated due to the complexity of reservoir drainage area estimation. The drainage area depends on boundaries and spatial distributions of naturally and hydraulically fractured networks, which are very difficult to determine. Thus, the RF result from the direct method contains many
uncertainties. For example, in the Bakken Formation, the current recovery factor is still ambiguous and not often reported. In 1992, the RF was estimated between 15% to 20% for horizontal wells (Reisz, 1992). In 2008, from reserves and RF analyses by county, RF was proposed to range from 0.7% to 3.7% (Bohrer et al., 2008). In 2009, the factor was recently estimated at approximately 6.1% to 8.7% (Clark, 2009).
In SAGD, steam is injected into a bitumen bearing oil sands formation. Steam temperature ranges from about 200 to 260°C and at these temperatures, bitumen undergoes aquathermolysis yielding acid gases such as hydrogen sulphide and carbon dioxide. SAGD simulation models in the literature often account for spatial heterogeneity of the geology and oil composition and heat transfer, multiphase flow, gas solubility effects, and viscosity variations with temperature, however, none account for the chemistry of SAGD. Here, we consider aquathermolysis reactions to understand the reactive zones in the SAGD process and how the process generates acid gases via aquathermolysis. The results show that SAGD is both a physical and chemical-reactive process.
Steam generated at the surface is currently the most common technology for in-situ thermal stimulation of heavy oil, however it has significant limitations due to heat loss that make it uneconomic for some 2 trillion barrels of deeper heavy oil resources. Downhole steam generators (DHSGs) have the ability to unlock these deeper deposits and allow steam to be used where it was previously not economical. A brief review of recent advances in DHSG technology and a description of potential applications using thermal oil recovery modeling and simulation are presented here. The DHSG technology described can be particularly suitable for recovering a number of known heavy oil deposits around the world that have experienced low recovery efficiencies through primary and secondary (water flood) production methods. These deeper resources typically require thermal stimulation to reduce viscosity in-situ for improved production volumes. Cost estimates and economic projections were developed based on a full green field project using production estimates obtained from STARS, a widely accepted heavy oil simulation program. Simulations and associated economic studies for the fields modeled, assuming the deployment of DHSGs, showed viable economics for all cases at both USD $75/bbl and USD $100/bbl 2009 WTI oil prices. The technological constraints of deep heavy oil production may be removed by further development of DHSG technology including direct-fired, umbilical-supported downhole natural gas fired systems that can be operated without maintenance for long periods of time. DHSG technology development activities are presented and specific issues relating to previous unsuccessful efforts are discussed. Finally, a new generation of DHSG technology is introduced and key design and operating problems with proposed solutions are presented. The uniqueness of the downhole environment is discussed along with the special requirements imposed on a downhole combustion system as a result. Technical challenges include tool positioning and inaccessibility, installation and operational safety, pressure management, and flame stability & control. Advances from the aerospace industry are also discussed as they pertain to robust combustion system development and downhole tool design which help enable a broad range of operating pressures and flow rate turndown. The demonstrated potential for DHSG technology to create optimum steam and EOR gas mixes that are tailored for a given reservoir may provide the basis for a new generation of thermal in-situ EOR technology to unlock deep heavy oil resources as well as those in environmentally challenged locales such as offshore or under arctic or permafrost environments.
As interest in exploiting shale gas/oil reservoirs with multiple stage fractured horizontal wells increased, complexity of production analysis and reservoir description have also increased. The main objective of this paper is to present and demonstrate type curves for production data analysis of shale gas/oil wells using a Dual Porosity model.
Dual Porosity model is based on Bello and Wattenbarger's (2010) mathematical model where hydraulic fractures act as a secondary porosity system where matrix is the primary porosity system. Samandarli et al. (2011) showed application of this model on history matching and forecasting of shale gas wells with multiple fractures by doing regression on effective fracture and matrix permeability and half length. This type of regression is as rigorous as simulation, however much faster than it. On the other hand for "quick look interpretation?? having type curves will make the production analysis even more convenient for practical purposes.
With this method production of shale gas/oil well can be matched with developed type curves which vary with effective permeability. Once the production data is matched with one of the type curves, effective permeability and match points are recorded. By using dimensionless equations developed for Dual Porosity model fracture half-length can be determined. In order to use type curves, good estimate for effective porosity and matrix permeability should be predetermined.
Type curves developed in this paper were applied to synthetic and field data examples. Early results show that method works well in determining effective fracture half-length which is the most important parameter in evaluation of stimulated reservoir volume (SRV).
This paper discusses a modeling technique that can be used to predict the gas production from Marcellus Shale wells. The fundamental concept is to model the relationship between pressure and production over time using inflow performance relationships and volumetric calculations. The model can be adjusted to fit the production values and then used to predict production on that well or on offsetting wells based on the pressures that are expected.
The introduction of this paper provides a general description of the characteristics and production volumes that are typical of horizontal Marcellus wells.
The theory separates the reservoir into two systems; the stimulated reservoir and the matrix reservoir. From these systems inflow performance relationships, decline curves, and volumetric calculations are used to build the model. Over time, the inflow performance curves will change. With this model, an inflow performance curve can be generated at any point in the life of the well.
By creating a model for a well, a prediction can be made of the well's production if conditions change. This could be extremely helpful in modeling pipeline and compression systems or evaluating wells under changing conditions. It can also be used to estimate potential production from curtailed wells. Over time, the model may also be used to indicate if damage or enhancements have occurred or if chemical, mechanical or stimulation treatments were effective.
The technique could easily be applied to other similarly designed wells. This could include other shale wells, CBM wells or wells in naturally fractured reservoirs.
This paper presents a theory that is a "first run?? concept, based on previous work, theories, and observations. The technique presented can be expanded and refined by future researchers.