OR/15/066 Fracture initiation
|Cuss, R J, Wiseall, A C, Hennissen, J A I, Waters, C N, Kemp, S J, Ougier-Simonin, A, Holyoake, S, and Haslam, R B. 2015. Hydraulic fracturing: a review of theory and field experience. British Geological Survey Internal Report, OR/15/066.|
This chapter introduces the mechanisms responsible for the formation and initiation of hydraulic fractures following perforation of the well casing. The section titled Fracture propagation then concentrates on the propagation of these fractures.
In this section we introduce the basic concepts of rock mechanics relevant to the stress state that shale at depth will be subjected.
As introduced in Chapter 2, rocks at depth are subject to a complicated, heterogeneous, stress field. The vertical component of stress is related to the weight of overlying rock, which is partially transmitted into the horizontal sense (the Poisson effect). Additional horizontal stresses are created by erosion (Goodman, 1989), tectonic activity arising from lithospheric resistance to plate motion, rock anisotropy, and geological discontinuities. The result is a complex stress-field, which is described locally by an orthogonal set of normal (s) and shear (t) stresses (Figure 5a). It should be noted that the principal stress components corresponding to maximum (σ1), intermediate (σ2), and minimum (σ3) stress do not necessarily correspond with vertical (z) and horizontal (x, y) directions, as exaggerated in (Figure 5b). It is often simplified that the maximum stress component (σ1) corresponds with the vertical direction.
Another aspect to consider when discussing the stresses acting on a rock is the pressure from the fluid held within the pore space. This acts in the opposite direction to the normal stress, the result is an effective pressure which can be expressed as:
𝜎′ = 𝜎𝑛 − 𝑢
where σ’ is the effective stress, σn is the normal stress and u is the pore pressure acting on the rock.
When acting on a plane, stress can be split into a normal and shear component (Figure 5c). Normal stress (σn) acts perpendicular to a plane whereas the shear stress (τ) acts parallel to this plane. Normal compressive stress tends to inhibit sliding along a plane; shear stress tends to promote sliding. Normal stresses are considered to be positive if they are compressive and negative if they are tensile. Shear stresses are labelled according to their sense of shear.
The stresses that a rock is subjected can result in deformation; which may be recoverable (elastic) or permanent (plastic or inelastic) when stress is relieved. In many regions, the upper crust is subject to shear stresses approaching the frictional strength of favourably orientated faults (Engelder, 1992). This results in a state of limiting equilibrium within the crust with rocks at depth close to the point of failure according to the frictional characteristics of the rock. Deformation can present itself in rocks in many forms; for instance faults, fractures, joints, compaction bands, mineral alteration, cementation, grain crushing or porosity reduction.
Elastic deformation is often the initial response of geological materials to an applied stress and strain. This deformation is fully recoverable and non-permanent once the load is removed (Figure 6). The elastic moduli of Young’s Modulus (E), Poisson’s ratio (ν), Bulk Modulus of Compressibility (K) and the Shear Modulus (G) describe a rock’s stiffness, translation of strain in one principal direction into the other principal directions, resistance to volume change, and resistance to shear deformation respectively. The elastic moduli allow predictions of deformation state for a given stress condition. Wholly elastic responses are rare in geological materials and may only occur at very low strain rates. This is mainly due to the natural heterogeneity of rocks and the stress field. However, knowledge of the elastic properties allows the initiation of fractures to be predicted.
All materials have a limit at which permanent (inelastic or plastic) deformation occurs (Figure 6a); often referred to as yield. Rock deformation can either be plastic, ductile or brittle. The mode of deformation which may occur is governed by the stress state, material properties, temperature and hydraulic conditions.
Brittle behaviour represents a near-instantaneous stress reduction (Figure 6b) involving some combination of fracture and frictional sliding, and is common in rocks at low pressures and temperature. Usually less than 1% elastic strain occurs before failure and results in fault or fracture formation. Stable frictional sliding along fractures requires less energy than fracture initiation (μdynamic< μstatic), resulting in a stress-drop. Failure is observable on a wide range of scales, from microscopic to regional scale, as observed in the upper crust from microcracks to continent scale strike slip zones. Low porosity, well indurated argillaceous rocks, such as shale, behave in this manner depending on the rate of strain.
Certain rock types display a stress-strain relationship where yield-stress is achieved, but a peak-stress is never attained due to continual work-hardening (Figure 6c). This is what occurs when deformation becomes increasingly difficult as the strain increases. These conditions are said to be fully mechanically ductile (Jaeger & Cook, 1979). Ductility can be viewed as rock ‘flow’, with rupture occurring, if at all, after at least 10% shortening. Poorly indurated argillaceous rocks may behave in this manner. The transition from brittle to ductile behaviour occurs as confining pressure increases and therefore mode of deformation is dependent on depth and the physical properties of the rock.
Pore pressure effects
As described in the section State of stress the fluid within pores exerts a pore fluid pressure (u), which acts in the opposite direction to the confining pressure (s), forming an effective pressure (δ):
𝝈′ = 𝝈 – 𝑢
Stresses within shale are therefore described in terms of the effective stress as it dictates deformation. The effect of pore pressure is best shown by Figure 7. The principal stresses are plotted as a circle in Mohr’s space, along with the Coulomb failure criterion in the compressional deformation field and a tensile failure criterion in the tensile stress field. The failure criteria are used to predict the stress conditions when permanent deformation will occur. The addition of pore fluid pressure can be seen to move the Mohr circle to the left. Therefore, under a static boundary condition (i.e. no change in rock stress), the addition of pore pressure can change the likelihood of deformation. Figure 7a represents the case of fracture reactivation. A pre-existing plane of weakness, oriented at θ to the stress field, is shown on the Mohr diagram. The addition of pore fluid pressure moves the Mohr circle to the left until this plane of weakness intercepts the Coulomb failure criterion, resulting in shear deformation in the compressional deformation field. Figure 7b shows the example of hydrofracture. The addition of pore fluid pressure has resulted in the Mohr circle intercepting the tensile fracture criterion at a stress of T0. This results in the formation of tensile hydrofractures.
The injection of fluid at a high pressure acts to move the effective stress of the rock to beyond the failure envelope; therefore failure occurs and fractures form. Gas is then free to flow out of the shale unit and into the well. This report will mainly focus on the mechanical properties of shale rocks; however there is clearly a large interaction between the mechanical properties and hydraulic properties of rocks which must be considered. As shown, pore pressure can either result in the reactivation of pre-existing discontinuities (faults, fractures, joints, etc) or the formation of new hydrofractures. In practice a combination of both is likely to occur, with shear deformation occurring along pre-existing microfractures, the formation of new hydrofractures and the extension of new hydrofractures at the tips of pre-existing fractures.
Stress concentration around a borehole
It has been stated that shale at prospective depth is under a state of stress dictated by the weight of the overburden and tectonic forces that might also act on the sedimentary basin. If a circular hole is made in a stressed plate, the stress distribution around the hole will be changed (Timoshenko & Goodier, 1970), i.e. as a borehole is drilled, the rock surrounding the hole must carry the force previously carried by the removed rock. This can be described as a conservation of energy. Although mass is removed (δ M ≠ 0) energy, and therefore stress, is not, and the condition of Laplace (Ñ2 δx + δy = 0) still has to be met. The solution for the stress modification around a circular opening in a uniaxially loaded infinite plate was given by Kirsch in 1898 (Timoshenko & Goodier, 1970). The Kirsch solution is easily modified to consider biaxial loading and the effect of pressure within the hole (Jaeger & Cook, 1979):
where δr is radial stress, δθ is circumferential or hoop stress, trθ is shear stress, r is the radius of the bore, u is pore pressure, R is distance from centre of bore to point where stresses are being calculated, θ is the angle made between R and δH. This solution results in a stress concentration around the periphery of the bore with regions where stress is increased and regions where it is decreased. It can be shown that tensile stresses form in the direction of the maximum far-field stress direction. Therefore a complex stress is formed around a borehole, which if greater than rock strength, results in compressional or extensional failure in and behind the borehole wall.
The boundary conditions are such that hoop stress (δθ) at the bore-surface varies from 3 δh–δH when θ = 0 to 3 δH–δh when θ = ½ π. Thus, an area of tensile (negative) stress is created in the maximum far-field stress direction (θ = 0). When u is zero, tensile stresses are absent from all points if 3δh > δH. Tensile stresses are created in the bore-surface if u > 3δh–δH with radial tensile failure possible.
The Kirsch solution in a biaxial stress-field (as given in Jaeger & Cook, 1979) is applicable to a borehole aligned with one of the principal stress directions and tends to be considered for a vertical borehole. The stress-field is greatly complicated by deviating the wellbore or by drilling horizontally. Hossain et al. (2000) give the solution to the stress-field as:
where δ is normal stress, t is shear stress, β is wellbore deviation, Ψ is wellbore inclination, δx, δy, δz are stress in the direction of the borehole with z parallel to the well, and δH, δh, δv are the principal stresses.
As stated in the section Basic concepts, the far-field stress is usually markedly distinct from homogeneous and shale at depth is subjected to a triaxial stress-field. What the numerical solutions introduced above show is that the stress field around the well is greatly complicated by this far-field stress and the stress concentrations created in the borehole wall. Changes in pore fluid pressures, such as during hydraulic stimulation, are most likely to create tensile fracturing in the direction of the maximum horizontal principal stress.
Hydraulic fracturing (tensile Mode I fracturing) can occur:
- Naturally, due to the tectonic regime and changes in the effective stress conditions (hydrofractures)
- Artificially, due to drilling activities (drilling-induced tensile fractures)
- Artificially, generated around a tunnel or borehole due to changes in the in situ stress conditions.
Hydrofactures may be large features, or a linked, permeable, dilatant fracture network. These changes may be induced by the development of disequilibria pore pressure conditions or by changes in the tectonic load. For example, a reduction in the minimum compressive stress (δ3), induced by extension during regional uplift, may result in the formation of dilatant shear fractures. Hydrofractures occur under conditions of low differential stress when pore fluid pressure reduces the minimum effective horizontal stress below zero to the tensile strength of the rock.
In extensional basins, where the minimum compressive stress (δ3) is significantly less than the maximum compressive stress (δ1), hydrofractures are invariably vertical to semi-vertical in orientation and form perpendicular to δ3. For hydrofractures to develop in preference to shear fractures, the following conditions must be satisfied:
where uf is pore fluid pressure required to initiate hydrofracture, δ1 and δ3 are maximum and minimum horizontal stresses respectively and T0 is the tensile strength of the cap-rock (Hubbert & Rubey, 1959; Sibson, 1995). These conditions can occur in highly overpressured systems undergoing continual subsidence, or during exhumation when rapid denudation, without re-equilibration of overpressure, results in tensile failure. Brittle shale will increase its permeability by developing dilatant fractures, whereas ductile shale is able to undergo plastic deformation without increasing permeability (it will contain non-dilatant, sealing fractures). The tendency to dilate will be a function of the mechanical properties of the rock, effective pressure and shear zone geometry. At a given effective pressure, a stronger (over-consolidated or cemented) rock is more likely to dilate than a weaker one.
Considerable research has been conducted in connection with the engineering of wells to investigate the generation of artificial hydraulic fractures in order to determine in situ stress. The hydrofrac (HF) test measures in situ stress down a borehole by increasing the pore fluid pressure in an isolated segment until tensile hydraulic fracturing is initiated, identified by a drop in pore fluid pressure. Breakdown pressure (uc) is defined as the borehole pressure necessary to initiate hydraulic fracturing. There are two classical HF criteria to establish equations between uc and in situ horizontal principal stresses (Song et al., 2001); one is based upon elastic theory for impermeable rocks (Hubbert and Willis, 1957); the other upon poroelastic theory and considers the poroelastic stress induced by fluid permeation into rocks (Haimson & Fairhurst, 1967). This has been extended to include the characteristics of the bore during pressurisation (Detournay & Cheng, 1992):
|Hubbert & Willis (1957):|
|Haimson & Fairhurst (1967):|
|Detournay & Cheng (1992):|
where u0 is initial pore pressure in the rock formation, Thf is the hydraulic fracturing tensile strength, and η and γ are the poroelastic parameter and dimensionless pressurisation rate respectively, given by:
where α is the Biot parameter (Biot & Willis, 1957), ν is the Poisson ratio, A is borehole pressurization rate, λ is the microcrack length scale, c is the diffusivity coefficient, and S is stress.
Basic fracture mechanics
To fully understand the physics behind hydraulic fracture development and propagation in shale rocks we must first understand the basic mechanics which underlies fracturing in geological materials. This first requires the understanding of the modes in which fractures form and then the basic theory which governs the behaviour observed. Discontinuities originate from the build-up and concentration of stress at the tips of natural weaknesses and heterogeneities (USNCRM, 1996). These natural heterogeneities are a result of the mechanical properties of the rock and the rocks response to lithostatic (uplift, erosion and weathering), tectonic and thermal stresses, together with variations in fluid pressures. The mechanics that underpin fracture processes derives from classic work by Griffith (1921) and Irwin (1958).
The Griffith theory is based upon the linear elastic theory, which states that the stress at a tip of a narrow fracture is infinite. As a crack grows this requires two new surfaces to be created which in turn creates what Griffith calls a surface energy, C, expressed as:
where C is the surface energy, E is the materials Young’s Modulus and γ is the surface energy density. Failure occurs when free energy attains a peak value at a critical crack length, beyond which the fracture energy will decrease and the crack length will increase.
Irwin (1958) further developed the Griffith model as the classic model only accounts for pure brittle materials such as glass. In ductile materials a plastic zone develops at the crack top, as load increases the plastic zone increases in size until the crack grows in length. This plastic zone acts to provide a resistance to the crack growth. Irwin split the energy into two parts, the stored elastic strain energy which is released as the crack grows (thermodynamic driving force) and the dissipated energy which includes plastic dissipation and the surface energy, therefore:
where γ is the surface energy, Gp is the plastic dissipation, G is the total surface energy. When applied to Griffith’s theory:
where, a is the microcrack length, E is the materials Young’s Modulus and σf is the stress at fracture. Irwin further developed this to calculate the magnitude of energy available for fracture by taking into account the asymptotic stress data displacement fields around a crack front:
where K is the stress intensity factor. The magnitude of this depends on geometry, size, location, and load distribution. The stress intensity factor is directly proportional to the applied load on the material. It is possible to determine the minimum value of K which is required to propagate the crack; this minimum value is referred to as the critical stress intensity factor Kc. Using a combination of Irwin and Griffith fracture mechanics it is possible to determine the shape of the stress field and the magnitude, using the stress intensity factor.
Dependent variables of hydraulic fracturing
The theory introduced above shows that the initiation of fracture formation is dependent on a number of factors. These include: the orientation and size of the borehole; orientation and magnitude of the stress-field; pressure and rate of increase of the hydraulic fracture fluid; pore fluid formation of the shale; elastic properties of the shale, including elasticity, poroelasticity, and tensile strength; and the crack properties, such as stress intensity factor and surface energy. These parameters will dictate where a fracture is initiated and consequently, the direction of fracture propagation.
It should be noted that the theory above is based on a homogeneous elastic medium. Shale is a complex heterogeneous material with strong directional variation in many properties that need to be considered when predicting where fracture initiation will occur. It is also important to consider that the starting shale is not pristine; the action of perforation will create weaknesses within the shale surrounding the well. Perforations need to be directed with respect to the stress-field so that they are in phase with the anticipated fracture direction (Hossain et al., 2000). Perforations have been shown to reduce longitudinal fracture initiation pressures when preferentially oriented (Hossain et al., 2000).
At the tip of a microcrack, the concentration of stress results in the creation of many small microcracks in a non-linear process zone. Microcrack communication lengthens the features, and propagates the process zone into the rock mass in the direction of the maximum compressive stress trajectory. Several processes control or influence discontinuity propagation, including elastic strain accumulation, crystal-plastic processes, diffusion processes, phase transformations and reactions, and fluid processes.
The propagation of fractures is clearly related to the stress state of the rock. The state of stress is often heterogeneous and this therefore has an effect on the mechanics of fracture formation and the type of fracture which may form. At all but the shallowest depths within the Earth, the far-field stress components S1, S2 and S3 are compressive, and in most locations they are of different magnitudes (Gay & Weiss, 1974). Two distinct discontinuity types exist in compression, namely shear and extension (Griggs & Handin, 1960). Three displacement modes act on an ideal, flat, perfectly sharp discontinuity (Figure 8):
- Tensile or opening (mode I)
- In-plane shearing or sliding (mode II)
- Anti-plane shearing or tearing (mode III)
The injection of high pressure fluid along with proppant materials during the hydraulic fracturing process will result in predominantly Mode I open fractures forming; these will form parallel to the maximum principal stress. However, Ferril et al. (2012) state that there will also be an element of Mode II fractures forming, as shearing of the rock mass will also take place during fracturing; this may result in a Mode IV fracture type which can be called a hybrid fracture.
It is important to understand the direction and magnitude of these hydraulic fractures in order to be able to predict productivity and the development of the fractured disturbed zone. Hydraulic fractures are predominantly tensile or opening (Mode I) fractures, meaning they will propagate perpendicular to the minimum principal stress (σ3). Therefore in order to predict the orientation of these fractures it is essential to have a detailed knowledge of the stress field within the target area. The stress state is considered to be a major factor that can influence rock deformation (Warpinski et al., 1982; Busetti et al., 2014; Ferril et al., 2012). Reservoir depth and the tectonic stress regime are considered the major influencing factors, as well as the in situ pore pressure. Shale plays in the United States are situated at a range of depths and stress states, as summarised by Table 5.
|Sample Group||Depths (m)||In situ Stress (MPa)|
| Barnett – 1
Barnett – 2
|2,600||σv = 65 u =30 σ' = 35|
| Haynesville – 1
Haynesville – 2
|3,450||σv = 85 u =60–70 σ' = 15–25|
|Eagle Ford – 1 Fort St. John||3,800|| σv = 90 u =65 σ' = 25|
σv = 25 u =10–12 σ' = 13–20
Physical properties of shale
Chapter 3 highlighted the variability seen in shale in terms of properties such as mineralogy, kerogen content and tectonic setting. The location and initiation pressure of fractures requires knowledge of these parameters.
As discussed in Shale variability, shale mineralogy can vary greatly, but will predominantly contain a significant portion of clay minerals along with a combination of quartz, feldspars and carbonate minerals. The mineralogy of a shale unit is often used to predict the way in which it may deform under a certain stress field. It is generally agreed that a large portion of quartz and carbonate minerals will mean the shale is likely to deform in a brittle manner. Whereas, a larger clay content will more likely result in more plastic deformation. Therefore, hydraulic fracture treatments are more likely to be targeted in areas with higher quartz and carbonate contents as they will be more likely to fracture. This does not mean a shale with a high clay content will not be a productive formation, for example the Barnett Shale has areas with a clay content of up to 39% (Sone & Zoback 2013), this is also the case for the Haynesville field. This rule can also apply to the quartz content, which is as low as 11% in some parts of the Eagle Ford shale (Sone & Zoback 2013); this is however compensated by a carbonate content of up to 78%. As well as mineralization, the degree of cementation (or induration) can also influence whether brittle or ductile deformation is likely. More indurated shale will behave in a brittle manner. Even plastic clays will hydrofracture if the rate of pressurization is high enough.
The initiation of hydraulic fractures is dependent on the elastic properties of the shale; these can either be derived from wireline geophysics in situ or through laboratory experimentation. The ability to derive elastic properties from non-intrusive wireline logs make deriving understanding on hydraulic fracturing based on elasticity favourable.
Young’s modulus, E
The Young’s Modulus (E) is a measure of material stiffness and is therefore a key parameter in terms of hydraulic fracture propagation. Gale et al. (2007) quote values that can range from 4 to 61 GPa; which shows considerable variation. This variation may be related to variations in mineralogy. It is generally observed that as the clay and kerogen content decreases E will increase (Sone & Zoback 2013; Josh et al., 2012). A high silica or carbonate content is likely to result in a higher E (Jarvie et al., 2007; Ding et al., 2012). High values of E are likely to result in longer fracture lengths, as found by a study on the Woodford Shale by Tran et al. (2014). Josh et al. (2012) conducted two laboratory experiments on two separate facies, one with high clay content (~60%) and no obvious laminations and another which had a moderate clay content and a more well-developed fabric. The shales had an E value of 1–3 GPa and 9–11 GPa respectively. This highlights the role that clay content and fabric can have on the elastic moduli, however this study did not investigate properties perpendicular and parallel to the bedding. Sone & Zoback (2013) showed that the anisotropy of E increases with the clay and kerogen content. Despite the importance of E in terms of fracture propagation there is a relatively small amount of data openly available.
Poisson’s ratio, ν
The Poisson’s ratio is a measure of translation of strain in one of the principal directions into the other principal directions. One correlation that can be made between this and the Young’s Modulus is that they are negatively correlated. Therefore, brittle shale will have a low value for the Poisson’s ratio. So the same principle applies with respect to mineralogy that a high brittle mineral content will result in a low Poisson’s ratio (Rickman et al., 2008; Tran et al., 2014; Ross & Bustin, 2008).
There is a relatively small amount of values available for the Poisson’s ratio of shale. Sone & Zoback (2013) do show, however, that ν will exhibit a degree of anisotropy, with the greater values being generally parallel to bedding. There was, however, no obvious correlation between anisotropy and the clay and/or kerogen content.
Shear modulus, G
The Shear modulus (G) is a measure of the resistance to shear deformation. As stated previously, the predominant mode of fracturing that occurs during stimulation is tensional; however there may be an additional element of shear. Despite this fact there is a paucity of available values for G. It is, however, possible to calculate a shear modulus using values of E and v. The values of the elastic moduli reported in Gale et al. (2007) for the Barnett Shale and Austin Chalk give shear modulus values of 12.7–13.8 GPa and 17.1–21.8 GPa for each shale respectively.
Josh et al. (2012) state that many models often assume a constant value for the shear modulus, however it may be more likely to be anisotropic and heterogeneous throughout a shale formation. Sayers et al. (2015) and references within assign a shear modulus to the main mineralogical components, giving Quartz 44 GPa, Calcite 29 GPa, clay minerals 6 GPa and kerogen 3.2 GPa. This would therefore result in a heterogeneous distribution of the shear modulus as the mineralogy varies throughout the formation. Johri & Zoback (2013) assume a shear modulus of 30 GPa for their model, although the relationship of this assessment is unclear to the values calculated from Gale et al. (2007) and the values for the individual constituents. This illustrates that there is likely to be a high degree of variability in G.
A more direct approach to predicting the initiation of hydraulic fractures is the measurement of strength parameters. These can be recorded from true tensile tests, indirect tensile tests, or by compression tests. Ture tensile tests are rare in rocks, even more so in shale, with indirect or compression testing more common.
Unconfined compressive strength, qu
The uniaxial compressive strength test (UCS) is a standard rock mechanics tests conducted on unconfined, prepared, core samples loaded axially until failure. The UCS test yields the unconfined compressive strength (qu), which is used as a comparative parameter in most rock mechanics applications.
Josh et al. (2012) present data from a weak and a strong shale; they state that the weak shale would not be considered for shale gas exploration. The weak shale had a qu of 8 MPa and had a clay content of approximately 60% whereas the strong shale had a clay content of approximately 30 % and a qu of 44 MPa. Davey et al. (2012) report values of 117 MPa and 136 MPa for the average qu of the Montney Shale in two separate boreholes. Sone & Zoback (2013) inferred qu using the internal angle of friction and intercept from triaxial tests. For samples of Barnett, Haynesville, Eagle Ford and Fort St. John shale qu ranged between 100 and 250 MPa. A negative correlation between qu and the clay and/or kerogen content and positive correlation between qu and E was observed. It should be noted that considerable variation in qu is reported by Josh et al. (2012), Davey et al. (2012) and Sone & Zoback (2013) for shale of between 8 and 250 MPa (a factor of 30). In terms of rock strength characterization (Waltham, 1994) shale would range from a weak rock to a strong rock. Whilst the UCS test is a commonly performed test, it does not directly give insight into the hydraulic tensile properties of shale at depth. It does, however, show that strength is greatly variable in shale and that the pressure at which hydraulic fractures will form will greatly vary depending on the properties of the shale at the point of stimulation.
Tensile strength, T0
The tensile strength may be considered one of the more important physical parameters of shale formations due to the tensile nature of hydraulic fractures. The tensile strength is traditionally calculated in the laboratory using the indirect tensile test, or Brazilian test. A cylinder of rock is loaded perpendicular to the long axis between two flat plates. Although compression is applied to opposite sides of the sample, this results in tensile stresses at the centre of the sample, resulting in the formation of a tensile fracture.
Sierra et al. (2010) present data from a shallow monitoring borehole which intersected Woodford Shale, the maximum depth being approximately 65 metres. Brazilian tests were carried out parallel and perpendicular to bedding, the results showed the tensile strength to be anisotropic and heterogeneous throughout the borehole. The perpendicular tensile strength was in the region of 10–15 MPa and 5–10 MPa parallel to bedding. The lower values were found to correlate to regions with a high clay and kerogen content. Sone & Zoback (2013) support this theory, stating shale with a high clay content are likely to have lower tensile strength. Slatt (2011) also report anisotropy in tensile strength, quoting 7.1 MPa and 12.6 MPa for parallel and perpendicular to bedding respectively. Areas of high clay content may be more likely to have a well-developed lamination due to the physical properties of clay minerals; these laminations are areas of weakness and may be contributing to a lower tensile strength. Tran et al. (2014) present the same experimental data as Sierra et al. (2010) on the tensile strength of the Woodford Shale. However, they also compare the anisotropic values of the tensile strength with the carbonate content. A correlation is reported for tensile strength and carbonate content, although it should be noted that considerable variation is observed in the data that could be interpreted as showing no variation with carbonate content. Keneti & Wong (2010) also present anisotropy data for the Montney Shale, using samples from between 2,318–2,320 metres. Perpendicular strength ranged from 6–15 MPa whereas the parallel tensile strength ranged from 0.3–2.8 MPa.
Few published datasets are available for tensile strength in shale formations. Those reported above show a large variation between 0.3 and 15 MPa (a factor of 50), with considerable anisotropy of around 2 to 5. It is clear that anisotropy is strong in shale and it is expected that this will play a major role in the initiation and propagation of hydraulic fractures. The main controls on anisotropy are related to clay content and the degree of lamination of the shale. Fractures are likely to propagate in the direction of least resistance, so that may be likely to be parallel to bedding if the shale is strongly laminated.
The role of mineralogy
As shown above, mineralogy plays a considerable role on the initiation and propagation of hydraulic fractures in shale. The physics governing fracture and rupture are related to mineral bonds and therefore it is unsurprising that the mineralogy plays such a key role. Recent technological advances, especially in imaging techniques, now mean that it is possible to describe the microstructure of shale, which occurs at a micro- to nano-metre scale. X-ray Computer Tomography (CT), high resolution micro CT and dual beam Focused Ion Beam Scanning Electron Microscopy are imaging techniques which allow the pore system and petrology of shales to be described more accurately. Petrophysical approaches to shale gas reservoirs have been described by Jacobi et al. (2008) and Parker et al. (2009) using laboratory and wireline log-based methods to identify organic matter, porosity, permeability and mechanical properties. Rickman et al. (2008) combined mineralogy and geomechanics with petrophysics to optimise the hydraulic fracturing program; they concluded that this needs to be done for each shale separately due to the heterogeneous nature of shale. Britt & Schoeffler (2009) bring together the mineralogy (clay content) and geomechanical conditions of various producing shale formations to recommend mineralogical and elastic property cut off points, below which shale would no longer be considered prospective from a brittle fracturing perspective. Despite these recent advances in techniques there still remains a large paucity in data which quantifies the rock mechanics properties that control the fracturing process. One of the major reasons for this limited data may derive from the difficulty in accessing quality, preserved core material for testing. Josh et al. (2012) consider this the most important issue with regard to experimental geomechanical testing in shale. This preservation of core material is essential to reducing uncertainty in experimental data through reducing the effects of drying, chemical & biological degradation and reduce the influence of de-stressing material prior to mechanical testing.
Knowledge gaps and recommendations
This chapter has described the state of understanding of the initiation of hydraulic fractures during stimulation. The following statements on current knowledge, knowledge gaps and recommendations can be made:
- Shale is a highly variable and heterogeneous material vertically and laterally. Both variability and heterogeneity need to be better understood and incorporated into numerical models to describe the behaviour of shale with respect to hydraulic fracturing.
- It is recommended that recovered core material from exploration wells is well preserved to maintain the stress state, reduce the effects of drying, chemical and biological degradation so that consistent datasets can be recorded, which should allow correlation of parameters to be determined.
- A lack of relevant data exists for shale in North America recorded from well preserved core material. Research has been conducted using a number of approaches, this hampers comparison studies. It is recommended that full disclosure of experimental protocols and data be made.
- A complex stress field is created around deviated wells in shale. The complexity of stress can be described for a perfectly elastic medium, the complexity of shale variability and anisotropy need to be incorporated so that a better understanding of where fracture initiation is likely to occur.
- Little research has been conducted on the effect of perforation on the mechanical properties of shale; it is recommended that this is undertaken to understand fracture initiation in shales.
- Little research has been conducted on quantifying tensile and/or hydraulic fracturing properties in the laboratory; it is recommended that this is undertaken to understand shale behaviour during hydraulic fracturing (for example, to confirm relationships between composition and rock behaviour), and also to upscale understanding from the laboratory to reservoir-scale models.
- It is clear that mineralogy plays a major control on the initiation of fractures in shale. More research is required in order to quantify the influence of different mineral constituents on the overall mechanical properties of shale. A better understanding of where and how fractures are initiated is also required.
- Goodman, R E. (1989). Introduction to Rock Mechanics. New York, John Wiley and Sons.
- Engelder, T. (1992). Stress Regimes in the Lithosphere. Princeton University Press. pp.351.
- Jaeger, J C, and Cook, N G W. (1979). Fundamentals of Rock Mechanics. (3rd edition) London, Chapman and Hall. Now Jaeger, J C, Cook, N G W, and Zimmerman, R W. (2007). Fundamentals of Rock Mechanics. (4th edition) London, Chapman and Hall.
- Timoshenko, S P, and Goddier, J N. (1970). Theory of Elasticity. (3rd edition). Tokyo: McGraw-Hill Kogakusha.
- Hossain, M M, Rahman, M K, and Rahman, S S. (2000). Hydraulic fracture initiation and propagation: roles of wellbore trajectory, perforation and stress regimes. Journal of Petroleum Science and Engineering 27 pp.129–149.
- Hubbert, M K, and Rubey, W W. (1959). Role of fluid pressure in mechanics of overthrust faulting: 1 — Mechanics of fluid-filled porous solids and its application to overthrust faulting. Geological Society of America Bulletin, 70, pp.115–166.
- Sibson, R H. (1995). Selective fault reactivation during basin inversion: potential for fluid redistribution through fault-valve action. In: Buchanan, J.G., and Buchanan, P G, eds., Basin Inversion: Geological Society of London, Special Publication, 88, pp.3–19.
- Song, I, Suh, M, Won, K S, and Haimson, B. (2001). A laboratory study of hydraulic fracturing breakdown pressure in table-rock sandstone. Geosciences Journal, 5, pp.263–271.
- Hubbert, M K, and Willis, D G. (1957) Mechanics of hydraulic fracturing. Journal of Petroleum Technology, 9, pp.153–168.
- Haimson, B C, and Fairhurst, C. (1967). Initiation and extension of hydraulic fractures in rocks. Society of Petroleum Engineering Journal, 7, pp.310–318.
- Detournay, E, and Cheng, A. (1992). Influence of pressurization rate on the magnitude of the breakdown pressure. In: Tillerson, J R, and Wawersik, W R, eds., Rock Mechanics. Rotterdam, AA Balkema, pp.325–333.
- Biot, M A, and Willis, D G. (1957). The elastic coefficients of the theory of consolidation. Journal of Applied Mechanics, 24, pp.594–601.
- U.S. National Committee for Rock Mechanics (1996) Rock Fractures and Fluid Flow: Contemporary Understanding and Applications. National Academy Press, Washington, D.C. pp. 551.
- Griffith, A A. (1921). The phenomenon of rupture and flow in solids, Philosophical Transactions of the Royal Society of London, A221, pp.193–198.
- Irwin, G R. (1958). Fracture. In: Flügge, S., ed., Handbuch der Physik. V. VI Elasticity and Plasticity Berlin, Springer, pp.551–590.
- Gay, N, and Weiss, L E. (1974). The relationship between principal stress directions and the geometry of kinks in foliated rocks. Tectonophysics, 21, pp.287–300.
- Griggs, D T, and Handin, J W. (1960). Observations on fracture and a hypothesis of earthquakes. Chapter 1. In: Griggs, D T, ed., Rock deformation — A symposium. Memoir — Geological Society of America Boulder, CO, United States, Geological Society of America (GSA), pp. 347-364.
- Ferrill, D A, Morris, A P, Smart, K J, and McGinnis, R N. (2012). Stress Management: Applications of structural geology and geomechanics to energy exploration and production. The Journal of Petrotech, 7, pp.32–40.
- Warpinski, N R, Schmidt, R A, and Northrop, D A. (1982). In-situ stresses: The predominant influence on hydraulic fracture containment. Journal of Petroleum Technology, 34, pp.653–664.
- Busetti, S, Jiao, W, and Reches, Z E. (2014). Geomechanics of hydraulic fracturing microseismicity: Part 1. Shear, hybrid, and tensile events. AAPG Bulletin, 98, pp.2439–2457.
- Sone, H, and Zoback, M. (2013). Mechanical properties of shale gas reservoir rocks — Part 1: Static and dynamic elastic properties and anisotropy. Geophysics, 78, pp.381–392.
- Gale, J F, Reed, R M, and Holder, J. (2007). Natural fractures in the Barnett Shale and their importance for hydraulic fracture treatments. AAPG Bulletin, 91, pp.603–622.
- Josh, M, Esteban, L, Delle Piane, C, Sarout, J, Dewhurst, D N, and Clennell, M B. (2012). Laboratory characterisation of shale properties. Journal of Petroleum Science and Engineering, 88–89, pp.107–124.
- Jarvie, D, Hill, R J, Ruble, T E, and Pollastro, R M. (2007). Unconventional shale-gas systems: The Mississippian Barnett Shale of north-central Texas as one model for thermogenic shale- gas assessment. AAPG Bulletin, 91, pp.475–499.
- Ding, W L, Chao, L, Chunyan, L, Changchun, X, Kai, J, Weite, Z, and Liming, W. (2012). Fracture development in shale and its relationship to gas accumulation. Geoscience Frontiers, 3, pp.97–105.
- Tran, M, Chen, S, Rafael, S P, Abousleiman, Y N, and Slatt, R M. (2014). A Geomechanics Approach to Evaluate Gas Shale Frackability: A Case Study with the Woodford Shale. AAPG Annual Convention and Exhibition, Search and Discovery Article #50913 (2014).
- Rickman, R, Mullen, M, Petre, J E, Grieser, W V, and Kundert, D. (2008). A Practical Use of Shale Petrophysics for Stimulation Design Optimization. All shale Plays Are Not Clones of the Barnett Shale. In: SPE Annual Technical Conference and Exhibition, 21–24 September, Denver, Colorado, USA.
- Ross, D J K, and Bustin, R M. (2008). Characterizing the shale gas resource potential of Devonian–Mississippian strata in the Western Canada sedimentary basin: Application of an integrated formation evaluation. AAPG Bulletin, 92, no.1, pp.87–125.
- Sayers, C M, Guo, S, and Silva, J. (2015). Sensitivity of the elastic anisotropy and seismic reflection amplitude of the Eagle Ford Shale to the presence of kerogen. Geophysical Prospecting, 63, pp.151–165.
- Johri, M, and Zoback, M D. (2013). The Evolution of Stimulated Reservoir Volume during Hydraulic Stimulation of Shale Gas Formations. In: Unconventional Resources Technology Conference, Society of Exploration Geophysicists, American Association of Petroleum Geologists, Society of Petroleum Engineers, pp.1661–1671.
- Davey, H. (2012. Geomechanical Characterization of the Montney Shale Northwest Alberta and Northeast British Columbia, Canada. PhD Thesis, Colorado School of Mines.
- Waltham, A C. (1994). Foundations of Engineering Geology. Glasgow, Blackie Academic & Professional, 88 pp.
- Sierra, R, Tran, M, Abousleiman, Y N, and Slatt, R M. (2010). Woodford Shale mechanical properties and the impacts of lithofacies, in 44th US Rock Mechanics Symposium, Salt Lake City, Utah.
- Slatt, R M. (2011). Important geological properties of unconventional resource shales. Central European Journal of Geosciences, 3, pp.435–448.
- Keneti, A, and Wong, R C. (2010). Investigation of Anisotropic Behaviour of Montney Shale Under Indirect Tensile Strength Test. In Canadian Unconventional Resources and International Petroleum Conference, 19–21 October, Calgary, Alberta, Canada.
- Jacobi, D, Breig, J, LeCompte, B, Kopal, M, Hursan, G, Mendez, F, Bliven, S, and Longo, J. (2008). Effective geomechanical and geochemical characterisation of shale gas reservoirs from the wellbore environment: Caney and the Woodford Shales. In SPE Annual Technical Conference and Exhibition. Society of Petroleum Engineers.
- Parker, M, Buller, D, Petre, J E, and Dreher, D T. (2009). Haynesville shale-petrophysical evaluation. SPE Rocky Mountain Petroleum Technology Conference. Society of Petroleum Engineers.
- Britt, L K, and Schoeffler, J. (2009) The geomechanics of a shale play: what makes a shale prospective? SPE Eastern Regional Meeting. Society of Petroleum Engineers.