Part 1: What Are the Challenges
Importance of hydraulic fracturing to industry
The shale hydrocarbon reservoirs compromise a significant portion of the total unconventional resources in Canada. Duvernay and Montney are the principal western Canadian shale plays which have been estimated to have over 600 trillion cubic feet of natural gas and 5 billion barrels of oil of technically recoverable resources. Other major shale plays in Canada are the Horton Bluff Shale of New Brunswick and Nova Scotia and the Utica Shale of Quebec. These abundant shale resources have provided an excellent opportunity for the Canadian oil and gas industry to play a vital role in the global crude oil supply.
The recovery of the shale resources has become economically viable in the past few decades only thanks to the technology known as hydraulic fracturing or “fracking.” In Canada, hydraulic fracturing method was first successfully implemented in Pembina Oil field, Alberta (an example of a conventional reservoir) in 1935. Since then, hydraulic fracturing has been extensively used to stimulate the hydrocarbon resources from tight shale reservoirs. Estimates suggest over 160,000 oil and gas wells are hydraulically fractured in Canada.
Hydraulic fracturing is a process of injecting a large volume of water, chemical, and sand at a high rate and pressure into a target formation, causing deformation and failure (fracture) of the rock mass and increasing the permeability in a region surrounding the injection point (Fig. 1a). The advent of the new technologies combining multi-horizontal drilling with hydraulic fracturing has significantly increased the hydrocarbon production from shale formations by providing greater access to difficult hydrocarbon area (Fig. 1b).
New challenges are emerging
The widespread use of the massive multi-stage hydraulic fracturing in shale resources has brought new challenges to the shale gas industry which are detrimental to the success of the fracturing job. Parent-child well interference (, or frac hit) and induced seismicity are perhaps the two biggest challenges of the shale gas industry. For example, the likelihood of the frac hit occurrence has been significantly increased in the past decade as more horizontal wells are fractured and completed near the previously depleted parent wells with significantly altered stress level (referred to as the pressure sink). The frac hit can potentially damage well casing and tubing, obstructing the oil flow toward production (due to the capillary effects), and subsequently reduce the life of a shale asset.
The number of the induced seismicity events associated with shale fracking and reservoir stimulations has also drastically increased in the recent years. The energy/work required for a typical hydraulic fracturing injection job is approximately 100 gigajoules (equivalent to an energy content of 100,000 cu ft of natural gas). This significant energy when dissipated can drastically change the stress state and pore pressure along the pre-existing natural fractures/faults initiating slips events and subsequent seismic energy release. It is worthwhile to mention that the overpressurized shale reservoirs are reportedly (e.g., some area of Duvernay formation) more prone to induced seismicity events.
The application of the geomechanics in frack design can reduce the operational costs of the shale resource developments leading to the optimization of the fracturing jobs and the mitigation of the potential environmental risk of reservoir containment breach and induced seismicity. However, the accelerating growth of the innovation in fracturing completion technologies has indubitably outpaced our understanding of geomechanics of the hydraulic fracturing process. It is, therefore, necessary to develop powerful simulations tools to improve our knowledge of hydraulic fracturing behaviour.
Realistic geomechanics modeling of complex hydraulic fracturing
The simulation of the hydraulic fracturing process in the shale reservoirs is an intricate task. The shale formations exhibit invariable geological complexities on both micro and macro-scales. The application of the traditional bi-wing planar hydraulic fracturing models for simulation of the shale fracturing can be misleading because the foundations of these models are based on the simplifying assumptions limiting their applicability. Simple fracture geometries, homogeneous and non-fractured rocks, elastic rock behaviour, and a priori know fracturing path are among the most common assumptions. In practice, propagation of a hydraulic fracture is more complex due to: (1) the presence of the discontinuities in rock masses, such as joints, natural fractures, and horizontal lamination; (2) heterogeneity in the mechanical properties of rocks including the Young’s modulus, Poisson’s ratio; and (3) the non-uniformity in the in-situ stress field (Fig. 2). The main challenges of the realistic fracturing modelling in shale reservoirs are:
Shale fracturing and creation of the Stimulated Reservoir Volume
It is widely recognized that the hydraulic fracturing in the tight shale formations does not lead to the creation of a single fracture, rather it creates a complex interconnected network of fractures commonly referred to as the Stimulated Reservoir Volume (SRV). The creation of the SRV (complex fracture pattern) has been so far understood to be dominantly associated with the shear reactivation and subsequent permanent dilation along the pre-existing discontinuities and natural fractures (as evidence by microseismic imaging). However, the recent observations supported by core studies sampled from previously hydraulically fractured well reveals that the creation of a complex fracture network does not necessarily depend upon the presence of the natural fractures, shear dilation and self-propped fracture reactivation. A hydraulic fracture can create new subparallel simultaneous multiple fracture propagations spanning over a large area of reservoir area (up to 5m on each side of the main fracture) (Fig. 3). These subparallel fractures can curve, intersect and coalesce to form new fractures creating a complex fracture swarm. The behaviour of the fracture swarm can be very complex at the discrete level. However, it has been shown that at some distance away from the fracturing details, the global/effective behaviour of the system does not depend upon the behaviour of the parts. In this case, the fracture can be represented by continuum smeared zone of enhanced permeability characterized by an internal length scale (Fig. 4). The presence of the natural fractures, stress anisotropy, and other heterogeneity (in mechanical and flow properties) can further complicate this complex fracturing behaviour.
I should also mention that the occurrence of fracture swarm (, or smeared fracture) is not only limited to the shale fracturing. The study of the dikes and sills (examples of natural hydraulic fractures) in magmatic rocks also confirms the existence of the simulated regions on either side of the dike extending up to 20 m in length. This large lateral extension relates to the slow propagation mechanism of the high viscous magma generating tensile stress far beyond dike tips (Fig. 5).
Influence of the layering and stress heterogeneities
In the dominant hydraulic fracturing simulations, we commonly assume that the formation properties and the in-situ stress field remain unchanged through the entire reservoir. However, in the real world, the rock masses may contain discontinuities, such as faults, joints, and bedding planes primarily caused by the important geological events of tectonic and diagenesis nature. The discontinuities often give rise to the non-uniformity and sharp contrasts in the mechanical properties of rocks, such as Young’s modulus and Poisson’s ratio. For example, the stiffer layers characterized by higher Young’s modulus and Poisson’s ratio tend to accommodate higher tectonic stress, whereas the softer rocks are often tectonically relaxed. As a result, fractures growth can be significantly restricted or favoured, for instance, when the hydraulic fracture tends to grow into a stiffer or softer layer, respectively. The difference of 2-5 MPa is reported to be sufficient for the fracture containment in the laboratory experiments. For the practical hydraulic fracturing cases, the combined effect of the formation lithology, the nature of the interface barrier (i.e., horizontal laminates, bedding planes), the geological and tectonic history, and the injection parameters determines the condition of the fracture arrest and containment [Sarvaramini et al., IJSS, 2018].
Unconstrained and Large fluid leak-off effects on SRV growth
One of the main challenges in mathematical modelling of SRV growth in shale reservoirs is to develop an appropriate model accounting for the massive fluid leak-off into the formation. The dominant hydraulic fracturing model often assumes that fluid leak-off volume into the surrounding rocks is small (or, negligible). During fracturing of shale formation using the slickwater where we expect a large SRV, the leak-off and related pore fluid diffusion take place over a broader range of scales, from 1-D to 2 or 3- D invalidating the conventional leak-off models (Fig. 6). The inaccurate representation of the fluid leak-off in hydraulic fracture models may lead to the overestimation of the fracture half-length and underestimation of the global SRV permeability.
Overcoming the challenges of the realistic fracture modeling in shale formation
The development of a mathematical model to simulate the evolution of the stimulated volume is singularly challenging. At the discrete level there are strong fabric issues (oriented joint sets, perhaps faults, and weak bedding planes), different joint properties, Biot coupling, advective-conductive heat transfer, and flow in joint arrays with changing apertures. Complex discrete interaction laws for joints involve sliding Mohr-Coulomb friction with joint dilation (i.e. aperture increase), cohesion loss related to sliding and to extensional displacements across joints, strongly non-linear block contact stiffness behaviour that deviates from Hertzian behaviour, and loss of contacts during hydraulic fracturing when some natural fractures become non-contacting. Furthermore, the rock blocks delineated by the joints in the stimulated volume possess anisotropic properties, and large-scale heterogeneity also exists (Sarvaramini et al., ARMS, 2018).
To date, the simulation of the complex fracturing deformation formations is mainly carried out using the Discrete Fracture Network (DFN) and Discrete Element Method (DEM). DEM and DFN treat the jointed rock masses as an assemblage of the interface and blocks. However, modelling of each fracture within the jointed rock mass with the micro-resolution of each fracture in the large-scale reservoir simulations is computationally expensive and often prohibitively complex task given the uncertainties in natural fracture distribution and reservoir parameters. Also, the DEM and DFN models do not account for the irreversible dissipated energy/work to create new surfaces- a requirement for the hydraulic fracturing process. It has recently been shown that it is possible to build up-scaled continuum models that can, in an average sense, capture the behaviour of naturally fractured rock masses (Sarvaramini et al., IJSS 2018, Sarvaramini et al., JAM 2018, Sarvaramini et al. , ARMA 2018, Sarvaramini et al. , ARMS 2018).
Dr. Sarvaramini has recently developed a new paradigm (STimsIM software) for modelling of the HF with the ability of integration into the reservoir simulation programs. Dr. Sarvaramini initially started STimsIM development at the University of Waterloo with collaboration with Dr. Robert Gracie at Civil and Environmental Engineering Department and Dr. Maurice Dusseault at the Earth and Environmental Science Department. In this new paradigm, an equivalent poro-elasto-plasto continuum zone of enhanced permeability represents the Stimulated Reservoir Volume. The details of the complex fracturing, nucleation and micro-events of void coalescence in the discrete level are not explicitly modelled; instead, their aggregate effects are translated into continuum constitutive laws for permeability and plasticity degradation, and internal physical parameters representing natural fractures.
STimsIM is a fully coupled 3D simulator developed in the framework of the Finite Element Method (FIM). This software is a substantial step forward over existing commercial HF software, and it provides new insights into the geomechanics of the fracturing in the naturally fractured rock masses.
In the future, we will showcase the strength of the developed approach by considering examples of the hydraulic fracturing and SRV evolution for typical injection of the fluid into the ultra-low permeable shale formation.
This article was contributed by Erfan Sarvaramini, Ph.D.
Erfan Sarvaramini, and Maurice Dusseault, Mohammad Komijani, and Robert Gracie. A Non-local Plasticity Model of Stimulated Volume Evolution During Hydraulic Fracturing. International Journal of Solids and Structures. 2018 Oct 2.
Erfan Sarvaramini, and Maurice Dusseault, and Robert Gracie. Characterizing the Stimulated Reservoir Volume during Hydraulic Fracturing-Connecting the Pressure Fall-off Phase to the Geomechanics of Fracturing, Journal of Applied Mechanics, 2018.
Erfan Sarvaramini, and Maurice Dusseault, and Robert Gracie. Pre- and Post-Fracturing Analysis of a Hydraulically Stimulated Reservoir. In 52nd US Rock Mechanics/Geomechanics Symposium. American Rock Mechanics Association, Seattle, USA, June 2018.
Erfan Sarvaramini , Robert Gracie, and Maurice. Dusseault. Hydraulic Fracture Stimulated Zone Analysis with a FEM Upscaled Model. 10th Asian Rock Mechanics Symposium, The ISRM International Symposium, Singapore. November 2018
Maurice Dusseault. Geomechanical aspects of shale gas development. Rock Mechanics for Resources, Energy and Environment. 2013.
Paul Delaney, Pollard David, Ziony Joseph, and McKee Edwin H. Field relations between dikes and joints: Emplacement processes and paleostress analysis.” Journal of Geophysical Research: Solid Earth 91.B5 (1986): 4920-4938.