- Email: [email protected]
Experimental and numerical simulation study on fracturing through interlayer to coal seam
Journal of Natural Gas Science and Engineering 21 (2014) 386e396
Contents lists available at ScienceDirect
Journal of Natural Gas Science and Engineering journal homepage: www.elsevier.com/locate/jngse
Experimental and numerical simulation study on fracturing through interlayer to coal seam D.Q. Li a, b, *, S.C. Zhang a, S.A. Zhang a a b
MOE Key Laboratory of Petroleum Engineering, China University of Petroleum (Beijing), Fuxue Road No.18, 102249 Beijing, China China United Coalbed Methane Corporation, Ltd., Anwai Street, Jia No. 88, Dongcheng Disrtict, 100011 Beijing, China
a r t i c l e i n f o
a b s t r a c t
Article history: Received 23 July 2014 Received in revised form 26 August 2014 Accepted 27 August 2014 Available online
Taking the technology of horizontal well fracturing through interlayer to coal seam as the simulation object, the true triaxial test system was employed for the first time in a fracturing stimulation experiment of layered combination specimens comprised of natural rock. The effects of in-situ stress, natural fracture and elastic modulus on hydraulic fracture propagation were studied. On this basis, the effects of different geological and engineering factors on fracture propagation in coal seams were studied quantitatively using the three-dimensional hydraulic fracturing numerical simulation model based on the fluidesolid coupling finite element method. The results indicate that the roof and floor in coal-bearing stratum with larger differences (5 MPa) between vertical stress and the maximum horizontal stress is preferential in the implementation of horizontal fracturing through interlayer to coal seam. The influence of a natural fracture on the propagation of a hydraulic fracture is mainly related to the width of the natural fracture, the injection pressure in the hydraulic fracture and the angle of approach. Under high injection pressure, the impact of natural fractures on hydraulic fracture propagation was significantly lessened. Depending on the developmental degree of a natural fracture, the effect of the approaching angle between the natural fracture and the hydraulic fracture will be smaller. The lower stress difference, elastic modulus difference and permeability difference between layers and higher pumping rates are beneficial in forming longer fractures in a coal seam. Permeability anisotropy characteristics of a coal seam and its roof and floor make the fracture geometry higher, wider and shorter. The experimental and numerical simulation study achievements provide a theoretical basis for effective implementation of this new technology in coalbed methane development. © 2014 Elsevier B.V. All rights reserved.
Keywords: Coalbed methane Horizontal well True triaxial Numerical simulation Propagation mechanism
1. Introduction The permeability of coalbed methane reservoirs in China is so low that it must be stimulated before the gas can be produced (Wright et al., 1995; Li et al., 2004, 2010; Shan et al., 2005; Meng et al., 2011). The application of horizontal well technology is limited in coal seams with depths below 1000 m due to substantial formation pressure, in-situ stress and poor borehole stability. In such instances, the initiation and propagation mechanism of hydraulic fractures near a wellbore is more complex. Over-pressure and fracture failure occur frequently in stimulation treatments (Zhao and Qin, 2010; Zhang, 2011; Lu, 2011). Therefore, it is crucial to develop new technologies to overcome the instability problems
* Corresponding author. Yanxiu Building 611, China University of Petroleum (Beijing), Fuxue Road No.18, Changping District, Beijing 102249, China. Tel./fax: þ86 10 89733323. E-mail address: [email protected] (D.Q. Li). http://dx.doi.org/10.1016/j.jngse.2014.08.022 1875-5100/© 2014 Elsevier B.V. All rights reserved.
of deep coal seam wellbores in order to exploit deep coal seam gas efficiently. In this article, we propose a novel stimulation technology: drill horizontal wells in the roof or floor of a coal seam and communicate with the coal seam through hydraulic fracturing. This new technology will not only effectively overcome the adverse effects of the instability of a deep coal seam, but can also eliminate the influence of downhole string on coal mining afterward. The key to this technology is whether hydraulic fracture can penetrate into coal seam from the roof or floor and propagate efficiently in coal seam or not. Several theoretical and experimental studies have been carried out in the past to investigate the effect of formation mechanics characteristics and interfacial properties between layers on the vertical propagation of hydraulic fracturing in layered formation. Simonson et al. (1976) concluded that a formation with higher Young's modulus can halt a fracture from a lower modulus layer based on a linear elastic fracture mechanism. Conversely, a
D.Q. Li et al. / Journal of Natural Gas Science and Engineering 21 (2014) 386e396
hydraulic fracture in a higher modulus formation can penetrate the interface to a lower modulus layer. Hanson et al. (1978, 1980) found in a similar way that the vertical propagation of a fracture may be halted whether the fracture is extended from a lower modulus layer to a higher modulus layer or in the opposite direction. Fung et al. (1987) argued that Young's modulus ratio between the layers (even increases up to 5) have no noticeable effect on the vertical propagation of a fracture using semi-analytical methods. They found that contrasts in elastic modulus between layers are not the main factor influencing the height containment of a fracture. Smith et al. (2001) came up with a similar conclusion. However, he further concluded that Young's modulus could affect the net pressure in a fracture. The net pressure in a layer with higher Young's modulus was larger, which would encourage fracture propagation in a vertical direction. Daneshy (1978) first studied the effect of interfacial property on the vertical propagation of fractures by experimenting on layered rock samples. He argued that fractures could easily penetrate the interface and propagate into other layers if the cementing strength between layers was high enough. With a weak cementing strength, fracture containment is possible within the layer and is associated with slippage at the interface. The conclusion was later confirmed by the experimental work of Anderson (1981), Defuel and Clark (1981) and Abass (1990). According to the study of Defuel and Clark (1981), in-situ stress had control effects on the vertical propagation of hydraulic fractures. With the increase of minimum horizontal principal stress in an adjacent layer, hydraulic fracture was difficult in penetrating into the adjacent layer. Warpinski et al. (1982) and Settari (1988) also agreed that stress contrast between layers was the most important factor affecting the height containment. Meanwhile, modulus contrast and interface cementing strength were not sufficient enough to stop the growth of a fracture. Thiercelin et al. (1989) researched the effect of fracture toughness on the vertical propagation of fractures. His study revealed that the layer with low fracture toughness could promote the propagation of fracture and the fracture might be limited in the layer with higher fracture toughness. However, Hsiao et al. (1987) proposed fracture toughness had a limited effect on the propagation of fracture based on experiments because variation ranges of fracture toughness is rather small. Rahim and Holditch (1995) studied the effect of formation mechanics characteristics (in-situ stress, Young modulus and fracture toughness) and fracturing fluid properties (apparent viscosity and volume of fracturing fluid) comprehensively on the vertical extension of fracture. It turned out that formation mechanics properties had a significant impact on fracture vertical propagation. An adjacent layer with high in-situ stress could limit the fracture in pay formation. For specific reservoirs, vertical propagation of a fracture was mainly affected by apparent viscosity and the volume of the fracturing fluid. However, the research achievements mentioned above were based on the assumption that hydraulic fractures can penetrate into an interlayer interface and an adjacent layer. It was not clear under what conditions such hydraulic fractures can be formed, nor was the effect of natural fracture on the propagation of hydraulic fracture in layered formation considered. What's more, changes of fracture parameters in adjacent layers were not investigated. Therefore, this article will clarify these mechanisms and their rules. The conditions for forming a vertical fracture in the roof or floor of a coal seam is studied through a true triaxial hydraulic fracturing experiment with layered samples made of natural rocks. The effects of natural fracture and elastic modulus between layers on the propagation rules of hydraulic fracture in layered medium were also investigated. In addition, three-dimensional numerical simulation of horizontal well fracturing in the roof and floor of a
387
coal seam was also conducted in order to study the changes of fracture parameters in the coal seam quantitatively. The research results will provide a theoretical basis for the application of this technology in the effective development of coalbed methane. 2. True triaxial experiment of hydraulic fracturing 2.1. Experiment apparatus Experiment apparatus used to conduct hydraulic fracturing simulation is a true triaxial simulation system (Zhou et al., 2008; Guo et al., 2014). This simulation system, as shown in Fig. 1, is composed of a large-sized pressure bearing system, a hydraulic intensifier system, a fluid injection system, a data acquisition system and other auxiliary devices. 2.2. Experiment design The stress parameters and experiment program are designed as shown in Table 1. To study the effects of natural fractures on the propagation of hydraulic fractures, simulated wellbore should be avoided so as not to intersect with natural fractures. Therefore, natural fractures strike are analyzed, parameters of natural fractures are accurately measured, and the location of the borehole section are also measured accurately before drilling a simulated wellbore. Large natural fractures exist in the roof and floor blocks of sandstone 2, limestone 1 and sandstone 4. Natural fracture lengths are 9.60 cm, 21.30 cm and 20.80 cm, respectively with corresponding widths 0.05 mm, 0.01 mm and 0.05 mm, respectively. The angles of approach between natural fractures and vertical hydraulic fractures are 75 , 30 and 15 , respectively. The natural fractures are relatively wide in coal samples M1, M5, M10 and M11. The natural fracture in sample M5 is parallel to the bedding plane with a width and length of 0.07 mm and 19.90 cm, respectively. Two crossed natural fractures exist in sample M1 with widths of 0.09 cm and 0.07 mm, respectively. And the corresponding lengths of natural fractures are 9.41 cm and 16.92 cm, respectively. In sample M10, the width and length of the natural fracture are 0.10 mm and 2.49 cm, respectively. In sample M11, the width and length of the natural fracture are 0.28 mm and 1.28 cm, respectively. In order to study the influence of the difference in modulus between layers on the growth of a hydraulic fracture, a limestone block was designed as a contrast sample. Young's modulus of
Fig. 1. Schematic plot of a true triaxial hydraulic fracturing simulation system. In the figure: 1-roof and floor rock; 2-coal; 3-cement; 4-sealed injection tube; 5-hydraulic cylinder; 6-pressure plate; 7-hydraulic cylinder system; 8-fluid injection pump; 9-air compressor; 10-data acquisition processor.
388
D.Q. Li et al. / Journal of Natural Gas Science and Engineering 21 (2014) 386e396
Table 1 Hydraulic fracturing simulation experimental program on combination samples. Combination sample no.
Rock combination
M1 þ sandstone 1 þ M2 M3 þ sandstone 2 þ M4 M5 þ limestone 1 þ M6 M7 þ sandstone 3 þ M8 Sandstone 4 þ M9 M11 þ sandstone 5 þ M10 Sandstone 6 þ M12
#1 #2 #3 #4 #5 #6 #7
Triaxial stress MPa
sn
sH
sh
8 8 8 12 12 12 12
7 7 7 7 7 7 7
5 5 5 5 5 1 1
Completion method
Displacement ml/min
Fracturing fluid
Open Open Open Open Open Open Open
25 25 25 25 25 25 25
Clean Clean Clean Clean Clean Clean Clean
hole hole hole hole hole hole hole
water water water water water water water
þ þ þ þ þ þ þ
2%KCl 2%KCl 2%KCl 2%KCl 2%KCl 2%KCl 2%KCl
Fig. 2. Preparation of standard sample. (a) Preparation rock; (b) Drill hole; (c) Fix simulated wellbore; (d) Cement wellbore; (e) Prepare for rock cementation; (f) Complete cementation; (g) Cast by Concrete; (h) Form standard sample.
sandstone and limestone are tested with samples of coring direction perpendicular to the bedding plane. Tested results are averaged 32.08 GPa and 66.18 GPa, respectively. 2.3. Specimen preparation Coal samples are collected from No. 3 coal seam in Duanshi Coal Mine in the southern part of Qinshui Basin belonging to a high-rank anthracite. The limestone and sandstone used as the roof and floor of the coal seam come from an outcrop near the Duanshi Coal Mine. The standard single rock size is 200 mm 90 mm 90 mm. A central 10 mm eyehole parallel to the bedding plane was first drilled in the roof and floor rock to a depth of 60 mm. Steel tube with a length of 140 mm was then
fixed to the hole to simulate the wellbore, leaving a 30 mm open hole section. The non-open hole section was cemented by a specific chemical glue. Then, the roof and floor rock was cemented with a coal sample with cement face parallel to the bedding plane. To obtain the standard 300 mm 300 mm 300 mm cubic samples, the cemented and dried combination sample was then cast with concrete. A detailed sample of the preparation process is shown in Fig. 2. Due to the heterogeneity of rock structure, the injection pressure curve shows fluctuation response during the fracture propagation progress (Guo et al., 2014; Yang et al., 2012). In order to determine whether hydraulic fractures have penetrated the interface into the coal seam, a 20 mm 20 mm rectangular space in the middle of the cement face was reserved (Fig. 3). 2.4. Experiment procedure
200 mm concre
90 mm
coal wellbore
90 mm
roof/floor
300 mm
90 mm
90 mm
coal
300 mm 300 mm Fig. 3. Schematic plot of standard sample with dimensions.
In order to simulate the horizontal well hydraulic fracturing, the direction of wellbore is parallel to the direction of minimum horizontal stress. To prevent mechanical shear failure of the sample caused by the unbalanced loading of triaxial stresses, the threedimensional stresses were first loaded to the value of the minimum horizontal stress at the same time, and then the maximum horizontal stress and vertical stress were increased to the maximum horizontal stress value slowly. Lastly, the vertical stress was again slowly increased to the designed value to finish the three-dimensional stress loading. For the convenience of fracture observation, a red agent was mixed with the fracturing fluid to enhance detection of the hydraulic fracture.
D.Q. Li et al. / Journal of Natural Gas Science and Engineering 21 (2014) 386e396
Fig. 4. Fracture morphology before and after hydraulic fracturing. (a) Views before fracturing. (b) Views after fracturing.
The pump was stopped when the injection pressure remained stable and the fracturing fluid overflowed from the chamber. After stresses were unloaded, the blocks were removed to reveal the fracture geometry. The experiment was completed and all data was saved.
2.5. Experiment results and analysis 2.5.1. Characteristics of hydraulic fracture propagation in roof or floor (1) Hydraulic fracture morphology under different stresses The results of this experiment revealed that when the difference between vertical stress and maximum horizontal stress was small (1 MPa), hydraulic fractures inside the roof or floor rock were horizontal (Sample #1 in Fig. 4) or vertical (Sample #3 in Fig. 4), but when the difference between vertical stress and maximum horizontal stress was larger (5 MPa), hydraulic fractures inside the roof or floor rock were generally vertical (Sample #5, #6 and #7 in Fig. 4), which indicated that when vertical stress was close to maximum horizontal stress, the control of stress conditions on fracture morphology was weak, and that the hydraulic fractures could be horizontal or vertical. When the difference between vertical stress and maximum horizontal stress was significant, the control of stress conditions on fracture morphology was enhanced, and vertical fractures were generally formed. It can be clearly seen that when developing coalbed methane with fracturing technology through interlayer to coal seam of a horizontal well in the roof or floor of a coal seam, coal bearing strata with a larger difference between vertical stress and horizontal stress should be preferably selected, to ensure a vertical hydraulic fracture can be formed and
389
Natural fracture;
Hydraulic fracture.
that it penetrates the roof or floor of a coal seam to enter the coal seam. (2) Impact of natural fracture on propagation of hydraulic fracture. As shown in Figs. 5e7, obvious natural fractures exist in (Table 1) sandstone 2 in combination sample #2, limestone 1 in combination sample #3 and sandstone 4 in combination sample #5. When preparing samples, the openhole section was avoided so as not to intersect with natural fractures. It was confirmed by the injection pressure curve that the hydraulic fracture was not initiated from a natural fracture. However, the impact of natural fracture on hydraulic fracture propagation in sandstone 2, limestone 1 and sandstone 4 is different. Hydraulic fracture in sandstone 2 was initiated from the rock body, but it extends along the natural fracture surface after intersecting with the natural fracture in the process of propagation (Fig. 5(b)). Hydraulic fracture of limestone 1 also encounters natural fracture, and a natural fracture opens locally (Fig. 6(f)), but the hydraulic fracture finally penetrates the natural fracture and propagates in the original direction. Natural fracture almost has no impact on hydraulic fracture propagation. Hydraulic fracture in sandstone 4 (Fig. 7) fails to intersect with the natural fracture in the process of propagation, thus, natural fracture has no impact on hydraulic fracture propagation. Based on comprehensive analysis, it argues that strike of natural fracture does impact the propagation direction of hydraulic fractures and the degree of impact of natural fracture on hydraulic fracture is connected to the development scale of natural fracture and the injection pressure of hydraulic fracture. If the opening width of a natural fracture is high (0.05 mm) and the injection pressure of hydraulic fracture is low (10e15 MPa), the hydraulic fracture also changes direction and propagates along the
390
D.Q. Li et al. / Journal of Natural Gas Science and Engineering 21 (2014) 386e396
Fig. 5. Contrast plot of combination sample #2 before and after fracturing. (a) View of natural fracture before hydraulic fracturing. (b) Hydraulic fracture after fracturing.
Fig. 6. Effect of natural fracture on hydraulic fracture of limestone 1 in combination sample #3.
natural fracture (such as sandstone 2, Fig. 5), even with a high angle of approach (75 ). If the opening width of a natural fracture is low (0.01 mm) and the injection pressure of a hydraulic fracture is high (30e40 MPa), the hydraulic fracture will penetrate the natural fracture and propagate along the original direction (such as limestone 1, Fig. 6). It also reveals that compared with the development degree of natural fracture, the impact of the angle of approach between natural fracture and hydraulic fracture on fracture propagation is less. Blanton (1982), Warpinski and Defuel (1987) and Zhou et al. (2008) argued that under the condition of a low horizontal stress difference (3.44 MPa), hydraulic fracture will propagate along natural fracture surfaces no matter how great the angle of
Natural fracture;
Hydraulic fracture.
approach. In the experiment above, the horizontal stress difference was only 2 MPa, but the natural fracture was penetrated by the hydraulic fracture. Thus it can be seen that the developmental degree of a natural fracture and the injection pressure both have an important influence on hydraulic fracture propagation as well as insitu stresses. In the process of the experiment, fracturing fluid displacement was set at 25 ml/min, but the injection pressure curve of sample #1 and #3 showed a significant difference, and the injection pressure in the limestone block was obviously higher than that of the sandstone block. In other words, the pressure inside the fracture in high Young's modulus layer was higher and was able to promote the propagation of the fracture, which further confirmed the conclusions proposed by Smith et al. (2001).
D.Q. Li et al. / Journal of Natural Gas Science and Engineering 21 (2014) 386e396
Fig. 7. Hydraulic fracture morphology of sandstone 4 in combination sample #5 and relative position of natural fracture and simulated wellbore. draulic fracture.
For the fracturing technology of a horizontal well through interlayer to coal seam, under the premise that a vertical fracture can be formed, one should avoid formations with a natural fracture with parallel bedding as much as possible. If this technology is implemented in this type of formation, it is suggested that the operating injection rate should be increased so as to ensure the success of this venture. 2.5.2. Response characteristics of hydraulic fracture through interlayer to coal seam In the process of hydraulic fracture in sandstone or limestone through interlayer to coal seam, it propagates from a high modulus
Fig. 8. Hydraulic fracturing injection pressure curve of combination sample #1.
391
Natural fracture;
Hy-
layer to an interface space and then to a low modulus layer. Fluid injection pressure in the fracture shows an obvious fluctuation by decreasing first and increasing later, as shown as the rectangular frame in Fig. 8. Therefore, whether or not hydraulic fracture has entered a coal seam can be judged by pressure fluctuation characteristics of the treatment pressure curve.
2.5.3. Characteristics of hydraulic fracture propagation in coal seam In addition to in-situ stress, hydraulic fracture propagation in coal seam is also affected by the bedding plane and cleat. According to experimental results, the degree of impact of natural fracture in a coal seam on hydraulic fracture propagation was associated with strike and the developmental scale of a natural fracture as well as the injection pressure of hydraulic fracture. When injection pressure was similar (about 10 MPa), hydraulic fracture was more easily affected by natural fracture and propagates along it when encountering natural fracture with a larger opening width. For example, hydraulic fracture inside a coal seam in combination sample #1 is diverted to two split fractures along natural fractures after it enters the coal seam from sandstone 1 and intersects with the natural fracture (Fig. 4); hydraulic fracture in combination sample #6 propagates along the natural fracture at the bottom surface of M10 to form a vertical fracture, and at the same time a secondary fracture extends along the natural fracture at the top surface of M11 (Fig. 4). In comparison, no obvious natural fracture exists in the coal samples of combination samples #5 and #7, and the hydraulic fracture propagates along the main fracture plane to form a single fracture (Fig. 4).
392
D.Q. Li et al. / Journal of Natural Gas Science and Engineering 21 (2014) 386e396
Fig. 9. Local enlarged view of hydraulic fracture at top surface of M5 in combination sample #3.
When the developmental scale and width of a natural fracture are larger, it has greater impact on hydraulic fracture propagation. But when the injection pressure of the hydraulic fracture is higher (average 32 MPa), this impact is weakened. An obvious open natural fracture exists in the back of M5 in combination sample #3, and when the hydraulic fracture intersects with this natural fracture, it dilates the natural fracture to form a T-shaped fracture. But at the same time the hydraulic fracture still propagates along the original direction and finally forms a complex fracture (Fig. 4). The bedding plane and cleat in the coal sample may cause tortuosity of the hydraulic propagation path, as shown in Fig. 9. As a consequence, when coalbed methane is developed with interlayer to coal seam hydraulic fracturing technology, the deep strata should be selected in order to avoid the adverse impact of a natural fracture in the bedding direction.
vp v2 u ¼m 2 vx vy
(2)
In this paper, we consider the solid as a porous media. The fluid flow in porous media obey Darcy's law:
1 vpw grw vw ¼ k$ nw grw vx
3. Three-dimensional numerical simulation of fracture propagation Based on the finite element method coupling seepage, stress and damage, 3-D numerical simulation on the fracturing of a horizontal well through interlayer to coal seam was conducted. In the simulation, coupling of the stress field and the seepage field was fully considered in the process of fracture growth. Formation geologic parameters are obtained from the No.3 coal seam and its roof and floor in a depth of approximately 1000 m at Shizhuang block at the southern part of Qin shui Basin. 3.1. Governing equation The strong form for the solid of the initial boundary value problem is as follows:
sij;j þ bi ¼ 0 in U u ¼ u on Gu s$n ¼ t on Gt s$nc ¼ p on Gcrack
Here s is the stress of solid skeleton, bi is the body force, u is a given displacement on displacement boundary, n is the normal of traction boundary, t is a given surface traction on traction boundary. nc is the normal of crack surface and p is the water pressure action on the crack surface. For an incompressible Newtonian fluid, in order to simplicity, we neglect the gravitational forces, and assume that the flow in the fracture is one dimension flow, then the governing equation of fluid motion can be written as:
(1)
(3)
where vw is the seepage velocity (fluid flow velocity relative to the solid skeleton), k is the vector of permeability coefficient, g is the acceleration of gravity, rw is the density of water, pw is the pore pressure, nw is the ratio between the volume of pore water and the total volume.
3.2. Simulation model The 3-D model is built by finite element software ABAQUS. Fig. 10 is the schematic plot of the model. The dimensions of the model are 300 m, 200 m and 80 m in the direction Y, X, Z respectively. The coal seam and the roof and floor are assumed to be fully bonded. The thickness of the roof, coal seam and floor is 40 m, 6 m and 34 m, respectively. To observe the interaction between adjacent hydraulic fractures, two potential fracture propagation faces with cohesive unit properties were set (Zhang et al., 2010). The fracture faces are perpendicular to the minimum horizontal stress direction, which initiates and extends along the XeZ face, forming
Table 2 Parameters in the model.
Fig. 10. Sketch plot of model.
No.
Items
Unit
Coal seam
Roof/floor
1 2 3 4 5 6 7 8 9 10 11 12
Elastic modulus Poisson's ratio Permeability Average leak off coefficient Vertical stress Maximum horizontal stress Minimum horizontal stress Fluid gravity Saturation Pore pressure Pumping rate Fracture space
GPa
5 0.2 0.685 0.0062 30 26 16 9800 1.0 9.0 8 100
30 0.16 1.37 0.0165 35 30 22 9800 1.0 9.0
103 mm2 m/min0.5 MPa MPa MPa N/m3 MPa m3/min m
D.Q. Li et al. / Journal of Natural Gas Science and Engineering 21 (2014) 386e396
393
Fig. 11. Displacement vector distribution at the end of simulation time.
(1) The pore pressure distribution at both sides of the fracture are asymmetric. Fig. 11 shows the final displacement vector graph of three layers. As depicted, the pore pressure between the two fractures increased as a result of the stress superposition which hinders the fracture propagating, and instead, grows towards the outer part. (2) Hydraulic fracture propagates effectively and penetrates the coal seam. The simulation fracture height in the coal seam (40 me46 m at height in model) is 31.67 m, which proves that the fracture initiating at the roof propagates effectively and grows through the coal seam to the floor. (3) The injection pressure curve presents an obvious response of penetrating layers. For better analysis, injection pressure curves at the injection point, the top point of the coal seam and the top point of the floor in the simulated first 20 s are plotted in Fig. 12. The curves show when the hydraulic fracture is breaking into the coal seam from the roof, the pressure at the injection point increases sharply while the breaking pressure of the coal seam decreases immediately. When the fracture penetrated the coal seam and reached the floor, the pressure at the top point of the coal seam increased while the pressure at the top point of floor decreased. So during the process of interlayer to coal seam fracturing, there will be an energy breakthrough effect when a fracture goes from one layer towards another layer. It demonstrated that the injection pressure will increase in the fracturing layer, and after it was fractured, the injection pressure will gradually decrease to reach a normal propagation pressure.
Fig. 12. Injection pressure curves at the first 20 s.
transverse fractures. The initiating point is located at the upper layer, 4.41 m away from the coal seam. For the sake of improving the accuracy of simulation results, LGR (Local grid refinement) was used to define grids of hydraulic fractures and the coal seam. The parameters used in the model are listed in Table 2.
3.3. Simulation results and analysis According to the simulation results, the following conclusions can be drawn: Table 3 Simulation results under different stresses. Vertical stress difference, MPa
Half fracture length, m
Fracture width, cm
Fracture height, m
Breakout pressure, MPa
Treatment pressure at 3600 s, MPa
5 8 10
108.6 102.86 97.15
1.25 1.24 1.18
31.67 30.77 27.50
24.3 24.9 27.6
21.4 22.2 23.4
Table 4 Simulation results under different elasticity modulus. Elastic modulus difference, GPa
Half fracture length, m
Fracture width, cm
Fracture height, m
Breakout pressure, MPa
Treatment pressure at 3600 s
25 20 15
64.26 70.01 81.42
2.51 2.36 2.05
60.34 53.61 47.83
26.7 27.0 27.2
21.7 22.0 22.4
394
D.Q. Li et al. / Journal of Natural Gas Science and Engineering 21 (2014) 386e396
Table 5 Simulation results under different permeabilities. Ki/Km
Half fracture length, m
Fracture width, cm
Fracture height, m
Breakout pressure, MPa
Treatment pressure at 3600 s, MPa
2 3 4 5
108.6 74.29 51.43 45.72
1.251 1.245 1.213 1.178
31.67 31.67 31.67 27.51
24.3 22.2 20.7 20.7
21.4 20.9 20.6 20.1
modulus contrast increases the breakout pressure in the barrier and the fracture length in the coal seam, but decreases the fracture width and height. Therefore, higher elastic modulus in coal seams and lower modulus contrast between layers are more beneficial in forming a longer fracture in coal seams.
Fig. 13. Relationship curve between fracture width and length under different injection rate.
3.4.3. The effect of permeability contrast on fracture propagation Keeps the vertical permeability to horizontal permeability ratio of the barrier and coal seam consistent, simulates and compares the fracture propagation rules when vertical permeability ratios of barrier to coal (Ki/Km) are 2, 3, 4, and 5. Simulation results in Table 5 show that a larger permeability contrast between the outer barrier and the coal seam results in lower treatment pressure and shorter fracture length in the coal seam. Because of an increase in permeability of the barrier, fracturing fluid loss increases, so the growth ability of the fracture in the coal seam is weakened.
3.4. Parametric study on fracture propagation Based on a standard simulation model established in 3.2, the effects of multiple factors including in-situ stress, Young's modulus, permeability and fluid injection rate on fracture propagation, were quantitatively studied. 3.4.1. The effect of in-situ stress contrast on fracture propagation Maintain stresses on the coal seam constant with parameters in Table 2 and maintains the three-dimensional stress proportions of the roof and floor. The in-situ vertical stress contrast of a coal seam and barrier is set at 5 MPa, 8 MPa and 10 MPa. It is recognized from Table 3 that the breakout pressure and the treatment pressure increases while the length, the width and the height of the fracture decreases when the in-situ vertical stress contrast between the barrier and the coal seam enlarges. It suggests that a larger in-situ vertical stress contrast produces containment of the fracture geometry and causes higher bottomhole pressure. However, the half fracture length reaches 97.15 m in the coal seam even under a vertical stress contrast of 10 MPa. This suggests that although the degree of fracture propagation is reduced, the fracture geometry is still very effective for the roof and floor horizontal well fracturing technology. 3.4.2. The effect of modulus contrast on fracture propagation The elastic modulus of coal seams are taken as 5 GPa, 10 GPa and 15 GPa respectively. It can be seen from Table 4 that lower layered
Table 6 Anisotropic permeability design of coal seam and its barriers. Principal stress direction
Permeability (103 mm2) Horizontal maximum Horizontal minimum Vertical stress stress direction stress direction direction
Corresponding X direction model direction Coal seam 0.685 Roof and floor 1.37 of coal seam
Y direction
Z direction
0.328 1.37
0.099 0.952
3.4.4. The effect of injection rate on fracture propagation Fracturing fluid rate at 6 m3/min, 8 m3/min and 10 m3/min are simulated. Fig. 13 shows fracture geometry in height and width. A higher injection rate enhances fracture length, height and width. It reflects a positive correlation between injection rate and fracture geometry. Considering the fracture initiation point (35.58 m), and the coal seam position (40e46 m) in the model height direction and the fracture propagation zone, it was found that the fracture could connect a 6 m thick coal seam even under an injection rate of 6 m3/ min. 3.5. The effect of permeability anisotropy on fracture propagation To compare the impact of permeability anisotropy on fracture propagation, a new set of permeability parameters are listed in Table 6 according to the true triaxial permeability test results. Simulation result reveals the following characteristics: (1) Pore pressure superposition degree is larger around a hydraulic fracture after considering permeability anisotropy. The superposition produces mainly along the direction of the fracture width growth, namely the direction of minimal horizontal principal stress. Take this direction from the top of the coal seam as the considered path. Pore pressure at homogeneous simulation model and anisotropic model along the path are mapped in Fig. 14. It is clear from Fig. 14 that pore pressure generally increases in the direction of the fracture width (Y direction) under the influence of permeability anisotropy. (2) Treatment pressure in the anisotropic model is higher than in the homogeneous model. In this paper, the value is 1 MPa. Meanwhile, affected by the permeability anisotropy, half fracture length decreases to 86.27 m, fracture width and height increases to 1.50 cm and 43.59 m respectively. Compared to the fracture geometry in the homogeneous model, overall fracture morphology becomes higher, wider and shorter. Since the permeability anisotropy is common in coal measure strata, the anisotropic model is much closer to the actual formation,
D.Q. Li et al. / Journal of Natural Gas Science and Engineering 21 (2014) 386e396
395
The lower elastic modulus and permeability difference between layers are beneficial in forming a longer fracture in a coal seam. Larger injection rates can also promote the growth of a hydraulic fracture. 5. Pore pressure superposition degree is larger around a hydraulic fracture after considering permeability anisotropy. Pore pressure generally increases in the direction of the fracture width. Compared to the results in the homogeneous model, treatment pressure in the anisotropic model is higher and the overall fracture morphology becomes higher, wider and shorter.
Acknowledgments
Fig. 14. Pore pressure distribution at the top of coal seam along direction Y.
so the shape and parameters of the fracture under this model are closer to the actual fracturing treatment result.
This study was funded by the National Science and Technology Major Projects (Grant No. 2011ZX05042). We thank China United Coalbed Methane Corporation, Ltd. for its permission to publish this paper. We also thank reviewers and editors for their constructive comments and suggestions on improving the manuscript.
4. Conclusions
References
Through true triaxial hydraulic fracturing simulation and threedimensional numerical simulation we get the following theoretical understanding for carrying out fracturing of a horizontal well through interlayer to coal seam:
Anderson, G., 1981. Effects of friction on hydraulic fracture growth near unbonded interfaces in rocks. SPE J. 21 (1), 21e29. Abass, H.H., 1990. Experimental observations of hydraulic fracture propagation through coal blocks. SPE 21289, presented at SPE Eastern Regional Meeting, Columbus, Ohio, 31 Octobere2 November. Blanton, T.L., 1982. An experimental study of interaction between hydraulically induced and pre-existing fractures, SPE 10847, presented at SPE Unconventional Gas Recovery Symposium, Pittsburgh, Pennsylvania, 16e18 May. Daneshy, A.A., 1978. Hydraulic fracture propagation in layered formations. SPE J. 18 (1), 33e41. Defuel, L., Clark, J., 1981. Hydraulic fracture propagation in layered rock: experimental studies of fracture containment. SPE J. 24 (1), 19e32. Fung, R.L., Vilayakumar, S., Cormack, Donald E., 1987. Calculation of vertical fracture containment in layered formations. SPE Form. Eval. 2 (04), 518e522. Guo, T.K., Zhang, S.C., Qu, Z.Q., Zhou, T., Xiao, Y.S., Gao, J., 2014. Experimental study of hydraulic fracturing for shale by stimulated reservoir volume. Fuel 128, 373e380. Hanson, M.E., Anderson, G.D., Schaffer, R.I., Hearst, J.R., Haimson, B., Cleary, M.P., 1978. LLL Gas Stimulation Program. Quarterly Progress Report. Lawrence Livermore National Lab.. JaneMar. Hanson, M.E., Anderson, G.D., Shaffer, R.J., 1980. Theoretical and experimental research on hydraulic fracturing. J. Energy Resour. Technol. 102 (2), 92e98. Hsiao, C., Rabaa, E., Wadood, A., 1987. Fracture toughness testing of rock cores. ARMA 87-0141, presented at the 28th U.S. Symposium on Rock Mechanics (USRMS), Tucson, Arizona, 29 Junee1 July. Li, A.Q., Jiang, H., Chen, C.H., 2004. Practice of hydraulic fracturing and fracturing model selection analysis in China. Nat. Gas Ind. 24 (5), 91e93. Li, L.D., Zhang, S.C., Geng, M., 2010. CBM reservoir hydraulic fracture propagation rule. Nat. Gas Ind. 30 (2), 72e74. Lu, C.H., 2011. Study on Stability and Mechanical Properties of Deep-Seated Coal. PhD thesis. Anhui University of Science and Technology, China. Meng, Z.P., Zhang, J.C., Wang, Rui, 2011. In-situ stress, pore pressure and stressdependent permeability in the Southern Qinshui Basin. Int. J. Rock Mech. Min. Sci. 48 (1), 122e131. Rahim, Z., Holditch, S.A., 1995. The effects of mechanical properties and selection of completion interval upon the created and propped fracture dimensions in layered reservoirs. J. Petrol. Sci. Eng. 13 (1), 29e45. Shan, X.J., Zhang, S.C., Li, A.Q., 2005. Hydraulic fracture propagation rules analysis for coalbed methane well. Nat. Gas Ind. 25 (1), 130e132. Simonson, E.R., Abou-Sayed, A.S., Clifton, R.J., 1978. Containment of massive hydraulic fractures. SPE J. 18 (01), 27e32. Smith, M.B., Bale, A.B., Britt, L.K., Klein H.H., Siebrits E, Dang X., 2001. Layered modulus effects on fracture propagation, proppant placement, and fracture modeling. SPE 71654, presented at SPE Annual Technical Conference and Exhibition, New Orleans, Louisiana, 30 Septembere3 October. Settari, A., 1988. Quantitative analysis of factors influencing vertical and lateral fracture growth. SPE Prod. Eng. 3 (03), 310e322. Thiercelin, M., Jeffrey, R.G., Naceur, K.B., 1989. Influence of fracture toughness on the geometry of hydraulic fractures. SPE Prod. Eng. 4 (04), 435e442. Wright, C.A., Tanigawa, J.J., Mei, S.X., Lei, Z.G., 1995. Enhanced hydraulic fracture technology for a coal seam reservoir in Central China, presented at International Meeting on Petroleum Engineering, Beijing, China, 14e17 November. Warpinski, N.R., Clark, J.A., Schmidt, R.A., Clarence, W.H., 1982. Laboratory investigation on the-effect of in-situ stresses on hydraulic fracture containment. SPE J. 22 (3), 333e340.
1. In-situ stress conditions have a significant impact on the hydraulic fracture morphology. Where differences of vertical stress and horizontal maximum stress are small (1 MPa), the control action of stresses on fracture morphology is weak. The hydraulic fractures can be horizontal, vertical or complex. Where differences of vertical stress and horizontal maximum stress are larger (5 MPa), the control action of stresses on fracture morphology are enhanced, and the hydraulic fractures are generally vertical. 2. The impact of natural fractures on hydraulic fracture propagation are related to the developmental scale of the natural fracture, the angle of approach and the injection pressure. If the opening width of a natural fracture is high (0.05 mm) and the injection pressure of the hydraulic fracture is low (10e15 MPa), the hydraulic fracture tends to divert and propagate along the natural fracture. If the opening width of the natural fracture is low (0.01 mm) and the injection pressure of the hydraulic fracture is high (30e40 MPa), the hydraulic fracture tends to penetrate the natural fracture and propagate along the original direction. When injection pressure in the hydraulic fracture is higher, the impact of the natural fracture on the hydraulic fracture will be weakened. Compared with the development degree of the natural fracture, the impact of the angle of approach between the natural fracture and the hydraulic fracture on fracture propagation is smaller. 3. When developing coalbed methane with interlayer to coal seam fracturing technology of a horizontal well in the roof and floor of a coal seam, coal bearing strata with larger differences (5 MPa) between vertical stress and horizontal stress should be selected. One should try to avoid natural fractures in developed formations, as much as possible. If the construction is implemented in such a formation, it is suggested to increase the operating injection rate so as to ensure the success of the project. 4. Inter-layer geological conditions and fracturing construction conditions have an important influence on hydraulic fracture parameters. Higher vertical stress contrast between layers presents a stronger inhibitory effect on the fracture propagation.
396
D.Q. Li et al. / Journal of Natural Gas Science and Engineering 21 (2014) 386e396
Warpinski, N.R., Defuel, L.W., 1987. Influence of geologic discontinuities on hydraulic fracture propagation (includes associated papers 17011 and 17074). J. Petrol. Tech. 39 (2), 209e220. Yang, J.S., Wang, Y.B., Li, A.Q., 2012. Experimental study on propagation mechanism of complex hydraulic fracture in coal-bed. J. Chin. Coal. Soc. 37 (1), 73e76. Zhao, L.J., Qin, Y., 2010. Research statement of domestic deep coalbed methane. China Coal. Meth. 3, 38e40.
Zhang, S.R., 2011. Outlook of effective development way for deep-Buried coalbed methane. China Coal. Meth. 4, 18e21. Zhou, J., Chen, M., Jin, Y., Zhang, G.Q., 2008. Analysis of fracture propagation behavior and fracture geometry using a tri-axial fracturing system in naturally fractured reservoirs. Int. J. Rock Mech. Min. Sci. 45 (7), 1143e1152. Zhang, G.M., Liu, H., Zhang, J., Wu, H.A., Wang, X.X., 2010. Three-dimensional finite element simulation and parametric study for horizontal well hydraulic fracture. J. Petrol. Sci. Eng. 72, 310e317.
Contents lists available at ScienceDirect
Journal of Natural Gas Science and Engineering journal homepage: www.elsevier.com/locate/jngse
Experimental and numerical simulation study on fracturing through interlayer to coal seam D.Q. Li a, b, *, S.C. Zhang a, S.A. Zhang a a b
MOE Key Laboratory of Petroleum Engineering, China University of Petroleum (Beijing), Fuxue Road No.18, 102249 Beijing, China China United Coalbed Methane Corporation, Ltd., Anwai Street, Jia No. 88, Dongcheng Disrtict, 100011 Beijing, China
a r t i c l e i n f o
a b s t r a c t
Article history: Received 23 July 2014 Received in revised form 26 August 2014 Accepted 27 August 2014 Available online
Taking the technology of horizontal well fracturing through interlayer to coal seam as the simulation object, the true triaxial test system was employed for the first time in a fracturing stimulation experiment of layered combination specimens comprised of natural rock. The effects of in-situ stress, natural fracture and elastic modulus on hydraulic fracture propagation were studied. On this basis, the effects of different geological and engineering factors on fracture propagation in coal seams were studied quantitatively using the three-dimensional hydraulic fracturing numerical simulation model based on the fluidesolid coupling finite element method. The results indicate that the roof and floor in coal-bearing stratum with larger differences (5 MPa) between vertical stress and the maximum horizontal stress is preferential in the implementation of horizontal fracturing through interlayer to coal seam. The influence of a natural fracture on the propagation of a hydraulic fracture is mainly related to the width of the natural fracture, the injection pressure in the hydraulic fracture and the angle of approach. Under high injection pressure, the impact of natural fractures on hydraulic fracture propagation was significantly lessened. Depending on the developmental degree of a natural fracture, the effect of the approaching angle between the natural fracture and the hydraulic fracture will be smaller. The lower stress difference, elastic modulus difference and permeability difference between layers and higher pumping rates are beneficial in forming longer fractures in a coal seam. Permeability anisotropy characteristics of a coal seam and its roof and floor make the fracture geometry higher, wider and shorter. The experimental and numerical simulation study achievements provide a theoretical basis for effective implementation of this new technology in coalbed methane development. © 2014 Elsevier B.V. All rights reserved.
Keywords: Coalbed methane Horizontal well True triaxial Numerical simulation Propagation mechanism
1. Introduction The permeability of coalbed methane reservoirs in China is so low that it must be stimulated before the gas can be produced (Wright et al., 1995; Li et al., 2004, 2010; Shan et al., 2005; Meng et al., 2011). The application of horizontal well technology is limited in coal seams with depths below 1000 m due to substantial formation pressure, in-situ stress and poor borehole stability. In such instances, the initiation and propagation mechanism of hydraulic fractures near a wellbore is more complex. Over-pressure and fracture failure occur frequently in stimulation treatments (Zhao and Qin, 2010; Zhang, 2011; Lu, 2011). Therefore, it is crucial to develop new technologies to overcome the instability problems
* Corresponding author. Yanxiu Building 611, China University of Petroleum (Beijing), Fuxue Road No.18, Changping District, Beijing 102249, China. Tel./fax: þ86 10 89733323. E-mail address: [email protected] (D.Q. Li). http://dx.doi.org/10.1016/j.jngse.2014.08.022 1875-5100/© 2014 Elsevier B.V. All rights reserved.
of deep coal seam wellbores in order to exploit deep coal seam gas efficiently. In this article, we propose a novel stimulation technology: drill horizontal wells in the roof or floor of a coal seam and communicate with the coal seam through hydraulic fracturing. This new technology will not only effectively overcome the adverse effects of the instability of a deep coal seam, but can also eliminate the influence of downhole string on coal mining afterward. The key to this technology is whether hydraulic fracture can penetrate into coal seam from the roof or floor and propagate efficiently in coal seam or not. Several theoretical and experimental studies have been carried out in the past to investigate the effect of formation mechanics characteristics and interfacial properties between layers on the vertical propagation of hydraulic fracturing in layered formation. Simonson et al. (1976) concluded that a formation with higher Young's modulus can halt a fracture from a lower modulus layer based on a linear elastic fracture mechanism. Conversely, a
D.Q. Li et al. / Journal of Natural Gas Science and Engineering 21 (2014) 386e396
hydraulic fracture in a higher modulus formation can penetrate the interface to a lower modulus layer. Hanson et al. (1978, 1980) found in a similar way that the vertical propagation of a fracture may be halted whether the fracture is extended from a lower modulus layer to a higher modulus layer or in the opposite direction. Fung et al. (1987) argued that Young's modulus ratio between the layers (even increases up to 5) have no noticeable effect on the vertical propagation of a fracture using semi-analytical methods. They found that contrasts in elastic modulus between layers are not the main factor influencing the height containment of a fracture. Smith et al. (2001) came up with a similar conclusion. However, he further concluded that Young's modulus could affect the net pressure in a fracture. The net pressure in a layer with higher Young's modulus was larger, which would encourage fracture propagation in a vertical direction. Daneshy (1978) first studied the effect of interfacial property on the vertical propagation of fractures by experimenting on layered rock samples. He argued that fractures could easily penetrate the interface and propagate into other layers if the cementing strength between layers was high enough. With a weak cementing strength, fracture containment is possible within the layer and is associated with slippage at the interface. The conclusion was later confirmed by the experimental work of Anderson (1981), Defuel and Clark (1981) and Abass (1990). According to the study of Defuel and Clark (1981), in-situ stress had control effects on the vertical propagation of hydraulic fractures. With the increase of minimum horizontal principal stress in an adjacent layer, hydraulic fracture was difficult in penetrating into the adjacent layer. Warpinski et al. (1982) and Settari (1988) also agreed that stress contrast between layers was the most important factor affecting the height containment. Meanwhile, modulus contrast and interface cementing strength were not sufficient enough to stop the growth of a fracture. Thiercelin et al. (1989) researched the effect of fracture toughness on the vertical propagation of fractures. His study revealed that the layer with low fracture toughness could promote the propagation of fracture and the fracture might be limited in the layer with higher fracture toughness. However, Hsiao et al. (1987) proposed fracture toughness had a limited effect on the propagation of fracture based on experiments because variation ranges of fracture toughness is rather small. Rahim and Holditch (1995) studied the effect of formation mechanics characteristics (in-situ stress, Young modulus and fracture toughness) and fracturing fluid properties (apparent viscosity and volume of fracturing fluid) comprehensively on the vertical extension of fracture. It turned out that formation mechanics properties had a significant impact on fracture vertical propagation. An adjacent layer with high in-situ stress could limit the fracture in pay formation. For specific reservoirs, vertical propagation of a fracture was mainly affected by apparent viscosity and the volume of the fracturing fluid. However, the research achievements mentioned above were based on the assumption that hydraulic fractures can penetrate into an interlayer interface and an adjacent layer. It was not clear under what conditions such hydraulic fractures can be formed, nor was the effect of natural fracture on the propagation of hydraulic fracture in layered formation considered. What's more, changes of fracture parameters in adjacent layers were not investigated. Therefore, this article will clarify these mechanisms and their rules. The conditions for forming a vertical fracture in the roof or floor of a coal seam is studied through a true triaxial hydraulic fracturing experiment with layered samples made of natural rocks. The effects of natural fracture and elastic modulus between layers on the propagation rules of hydraulic fracture in layered medium were also investigated. In addition, three-dimensional numerical simulation of horizontal well fracturing in the roof and floor of a
387
coal seam was also conducted in order to study the changes of fracture parameters in the coal seam quantitatively. The research results will provide a theoretical basis for the application of this technology in the effective development of coalbed methane. 2. True triaxial experiment of hydraulic fracturing 2.1. Experiment apparatus Experiment apparatus used to conduct hydraulic fracturing simulation is a true triaxial simulation system (Zhou et al., 2008; Guo et al., 2014). This simulation system, as shown in Fig. 1, is composed of a large-sized pressure bearing system, a hydraulic intensifier system, a fluid injection system, a data acquisition system and other auxiliary devices. 2.2. Experiment design The stress parameters and experiment program are designed as shown in Table 1. To study the effects of natural fractures on the propagation of hydraulic fractures, simulated wellbore should be avoided so as not to intersect with natural fractures. Therefore, natural fractures strike are analyzed, parameters of natural fractures are accurately measured, and the location of the borehole section are also measured accurately before drilling a simulated wellbore. Large natural fractures exist in the roof and floor blocks of sandstone 2, limestone 1 and sandstone 4. Natural fracture lengths are 9.60 cm, 21.30 cm and 20.80 cm, respectively with corresponding widths 0.05 mm, 0.01 mm and 0.05 mm, respectively. The angles of approach between natural fractures and vertical hydraulic fractures are 75 , 30 and 15 , respectively. The natural fractures are relatively wide in coal samples M1, M5, M10 and M11. The natural fracture in sample M5 is parallel to the bedding plane with a width and length of 0.07 mm and 19.90 cm, respectively. Two crossed natural fractures exist in sample M1 with widths of 0.09 cm and 0.07 mm, respectively. And the corresponding lengths of natural fractures are 9.41 cm and 16.92 cm, respectively. In sample M10, the width and length of the natural fracture are 0.10 mm and 2.49 cm, respectively. In sample M11, the width and length of the natural fracture are 0.28 mm and 1.28 cm, respectively. In order to study the influence of the difference in modulus between layers on the growth of a hydraulic fracture, a limestone block was designed as a contrast sample. Young's modulus of
Fig. 1. Schematic plot of a true triaxial hydraulic fracturing simulation system. In the figure: 1-roof and floor rock; 2-coal; 3-cement; 4-sealed injection tube; 5-hydraulic cylinder; 6-pressure plate; 7-hydraulic cylinder system; 8-fluid injection pump; 9-air compressor; 10-data acquisition processor.
388
D.Q. Li et al. / Journal of Natural Gas Science and Engineering 21 (2014) 386e396
Table 1 Hydraulic fracturing simulation experimental program on combination samples. Combination sample no.
Rock combination
M1 þ sandstone 1 þ M2 M3 þ sandstone 2 þ M4 M5 þ limestone 1 þ M6 M7 þ sandstone 3 þ M8 Sandstone 4 þ M9 M11 þ sandstone 5 þ M10 Sandstone 6 þ M12
#1 #2 #3 #4 #5 #6 #7
Triaxial stress MPa
sn
sH
sh
8 8 8 12 12 12 12
7 7 7 7 7 7 7
5 5 5 5 5 1 1
Completion method
Displacement ml/min
Fracturing fluid
Open Open Open Open Open Open Open
25 25 25 25 25 25 25
Clean Clean Clean Clean Clean Clean Clean
hole hole hole hole hole hole hole
water water water water water water water
þ þ þ þ þ þ þ
2%KCl 2%KCl 2%KCl 2%KCl 2%KCl 2%KCl 2%KCl
Fig. 2. Preparation of standard sample. (a) Preparation rock; (b) Drill hole; (c) Fix simulated wellbore; (d) Cement wellbore; (e) Prepare for rock cementation; (f) Complete cementation; (g) Cast by Concrete; (h) Form standard sample.
sandstone and limestone are tested with samples of coring direction perpendicular to the bedding plane. Tested results are averaged 32.08 GPa and 66.18 GPa, respectively. 2.3. Specimen preparation Coal samples are collected from No. 3 coal seam in Duanshi Coal Mine in the southern part of Qinshui Basin belonging to a high-rank anthracite. The limestone and sandstone used as the roof and floor of the coal seam come from an outcrop near the Duanshi Coal Mine. The standard single rock size is 200 mm 90 mm 90 mm. A central 10 mm eyehole parallel to the bedding plane was first drilled in the roof and floor rock to a depth of 60 mm. Steel tube with a length of 140 mm was then
fixed to the hole to simulate the wellbore, leaving a 30 mm open hole section. The non-open hole section was cemented by a specific chemical glue. Then, the roof and floor rock was cemented with a coal sample with cement face parallel to the bedding plane. To obtain the standard 300 mm 300 mm 300 mm cubic samples, the cemented and dried combination sample was then cast with concrete. A detailed sample of the preparation process is shown in Fig. 2. Due to the heterogeneity of rock structure, the injection pressure curve shows fluctuation response during the fracture propagation progress (Guo et al., 2014; Yang et al., 2012). In order to determine whether hydraulic fractures have penetrated the interface into the coal seam, a 20 mm 20 mm rectangular space in the middle of the cement face was reserved (Fig. 3). 2.4. Experiment procedure
200 mm concre
90 mm
coal wellbore
90 mm
roof/floor
300 mm
90 mm
90 mm
coal
300 mm 300 mm Fig. 3. Schematic plot of standard sample with dimensions.
In order to simulate the horizontal well hydraulic fracturing, the direction of wellbore is parallel to the direction of minimum horizontal stress. To prevent mechanical shear failure of the sample caused by the unbalanced loading of triaxial stresses, the threedimensional stresses were first loaded to the value of the minimum horizontal stress at the same time, and then the maximum horizontal stress and vertical stress were increased to the maximum horizontal stress value slowly. Lastly, the vertical stress was again slowly increased to the designed value to finish the three-dimensional stress loading. For the convenience of fracture observation, a red agent was mixed with the fracturing fluid to enhance detection of the hydraulic fracture.
D.Q. Li et al. / Journal of Natural Gas Science and Engineering 21 (2014) 386e396
Fig. 4. Fracture morphology before and after hydraulic fracturing. (a) Views before fracturing. (b) Views after fracturing.
The pump was stopped when the injection pressure remained stable and the fracturing fluid overflowed from the chamber. After stresses were unloaded, the blocks were removed to reveal the fracture geometry. The experiment was completed and all data was saved.
2.5. Experiment results and analysis 2.5.1. Characteristics of hydraulic fracture propagation in roof or floor (1) Hydraulic fracture morphology under different stresses The results of this experiment revealed that when the difference between vertical stress and maximum horizontal stress was small (1 MPa), hydraulic fractures inside the roof or floor rock were horizontal (Sample #1 in Fig. 4) or vertical (Sample #3 in Fig. 4), but when the difference between vertical stress and maximum horizontal stress was larger (5 MPa), hydraulic fractures inside the roof or floor rock were generally vertical (Sample #5, #6 and #7 in Fig. 4), which indicated that when vertical stress was close to maximum horizontal stress, the control of stress conditions on fracture morphology was weak, and that the hydraulic fractures could be horizontal or vertical. When the difference between vertical stress and maximum horizontal stress was significant, the control of stress conditions on fracture morphology was enhanced, and vertical fractures were generally formed. It can be clearly seen that when developing coalbed methane with fracturing technology through interlayer to coal seam of a horizontal well in the roof or floor of a coal seam, coal bearing strata with a larger difference between vertical stress and horizontal stress should be preferably selected, to ensure a vertical hydraulic fracture can be formed and
389
Natural fracture;
Hydraulic fracture.
that it penetrates the roof or floor of a coal seam to enter the coal seam. (2) Impact of natural fracture on propagation of hydraulic fracture. As shown in Figs. 5e7, obvious natural fractures exist in (Table 1) sandstone 2 in combination sample #2, limestone 1 in combination sample #3 and sandstone 4 in combination sample #5. When preparing samples, the openhole section was avoided so as not to intersect with natural fractures. It was confirmed by the injection pressure curve that the hydraulic fracture was not initiated from a natural fracture. However, the impact of natural fracture on hydraulic fracture propagation in sandstone 2, limestone 1 and sandstone 4 is different. Hydraulic fracture in sandstone 2 was initiated from the rock body, but it extends along the natural fracture surface after intersecting with the natural fracture in the process of propagation (Fig. 5(b)). Hydraulic fracture of limestone 1 also encounters natural fracture, and a natural fracture opens locally (Fig. 6(f)), but the hydraulic fracture finally penetrates the natural fracture and propagates in the original direction. Natural fracture almost has no impact on hydraulic fracture propagation. Hydraulic fracture in sandstone 4 (Fig. 7) fails to intersect with the natural fracture in the process of propagation, thus, natural fracture has no impact on hydraulic fracture propagation. Based on comprehensive analysis, it argues that strike of natural fracture does impact the propagation direction of hydraulic fractures and the degree of impact of natural fracture on hydraulic fracture is connected to the development scale of natural fracture and the injection pressure of hydraulic fracture. If the opening width of a natural fracture is high (0.05 mm) and the injection pressure of hydraulic fracture is low (10e15 MPa), the hydraulic fracture also changes direction and propagates along the
390
D.Q. Li et al. / Journal of Natural Gas Science and Engineering 21 (2014) 386e396
Fig. 5. Contrast plot of combination sample #2 before and after fracturing. (a) View of natural fracture before hydraulic fracturing. (b) Hydraulic fracture after fracturing.
Fig. 6. Effect of natural fracture on hydraulic fracture of limestone 1 in combination sample #3.
natural fracture (such as sandstone 2, Fig. 5), even with a high angle of approach (75 ). If the opening width of a natural fracture is low (0.01 mm) and the injection pressure of a hydraulic fracture is high (30e40 MPa), the hydraulic fracture will penetrate the natural fracture and propagate along the original direction (such as limestone 1, Fig. 6). It also reveals that compared with the development degree of natural fracture, the impact of the angle of approach between natural fracture and hydraulic fracture on fracture propagation is less. Blanton (1982), Warpinski and Defuel (1987) and Zhou et al. (2008) argued that under the condition of a low horizontal stress difference (3.44 MPa), hydraulic fracture will propagate along natural fracture surfaces no matter how great the angle of
Natural fracture;
Hydraulic fracture.
approach. In the experiment above, the horizontal stress difference was only 2 MPa, but the natural fracture was penetrated by the hydraulic fracture. Thus it can be seen that the developmental degree of a natural fracture and the injection pressure both have an important influence on hydraulic fracture propagation as well as insitu stresses. In the process of the experiment, fracturing fluid displacement was set at 25 ml/min, but the injection pressure curve of sample #1 and #3 showed a significant difference, and the injection pressure in the limestone block was obviously higher than that of the sandstone block. In other words, the pressure inside the fracture in high Young's modulus layer was higher and was able to promote the propagation of the fracture, which further confirmed the conclusions proposed by Smith et al. (2001).
D.Q. Li et al. / Journal of Natural Gas Science and Engineering 21 (2014) 386e396
Fig. 7. Hydraulic fracture morphology of sandstone 4 in combination sample #5 and relative position of natural fracture and simulated wellbore. draulic fracture.
For the fracturing technology of a horizontal well through interlayer to coal seam, under the premise that a vertical fracture can be formed, one should avoid formations with a natural fracture with parallel bedding as much as possible. If this technology is implemented in this type of formation, it is suggested that the operating injection rate should be increased so as to ensure the success of this venture. 2.5.2. Response characteristics of hydraulic fracture through interlayer to coal seam In the process of hydraulic fracture in sandstone or limestone through interlayer to coal seam, it propagates from a high modulus
Fig. 8. Hydraulic fracturing injection pressure curve of combination sample #1.
391
Natural fracture;
Hy-
layer to an interface space and then to a low modulus layer. Fluid injection pressure in the fracture shows an obvious fluctuation by decreasing first and increasing later, as shown as the rectangular frame in Fig. 8. Therefore, whether or not hydraulic fracture has entered a coal seam can be judged by pressure fluctuation characteristics of the treatment pressure curve.
2.5.3. Characteristics of hydraulic fracture propagation in coal seam In addition to in-situ stress, hydraulic fracture propagation in coal seam is also affected by the bedding plane and cleat. According to experimental results, the degree of impact of natural fracture in a coal seam on hydraulic fracture propagation was associated with strike and the developmental scale of a natural fracture as well as the injection pressure of hydraulic fracture. When injection pressure was similar (about 10 MPa), hydraulic fracture was more easily affected by natural fracture and propagates along it when encountering natural fracture with a larger opening width. For example, hydraulic fracture inside a coal seam in combination sample #1 is diverted to two split fractures along natural fractures after it enters the coal seam from sandstone 1 and intersects with the natural fracture (Fig. 4); hydraulic fracture in combination sample #6 propagates along the natural fracture at the bottom surface of M10 to form a vertical fracture, and at the same time a secondary fracture extends along the natural fracture at the top surface of M11 (Fig. 4). In comparison, no obvious natural fracture exists in the coal samples of combination samples #5 and #7, and the hydraulic fracture propagates along the main fracture plane to form a single fracture (Fig. 4).
392
D.Q. Li et al. / Journal of Natural Gas Science and Engineering 21 (2014) 386e396
Fig. 9. Local enlarged view of hydraulic fracture at top surface of M5 in combination sample #3.
When the developmental scale and width of a natural fracture are larger, it has greater impact on hydraulic fracture propagation. But when the injection pressure of the hydraulic fracture is higher (average 32 MPa), this impact is weakened. An obvious open natural fracture exists in the back of M5 in combination sample #3, and when the hydraulic fracture intersects with this natural fracture, it dilates the natural fracture to form a T-shaped fracture. But at the same time the hydraulic fracture still propagates along the original direction and finally forms a complex fracture (Fig. 4). The bedding plane and cleat in the coal sample may cause tortuosity of the hydraulic propagation path, as shown in Fig. 9. As a consequence, when coalbed methane is developed with interlayer to coal seam hydraulic fracturing technology, the deep strata should be selected in order to avoid the adverse impact of a natural fracture in the bedding direction.
vp v2 u ¼m 2 vx vy
(2)
In this paper, we consider the solid as a porous media. The fluid flow in porous media obey Darcy's law:
1 vpw grw vw ¼ k$ nw grw vx
3. Three-dimensional numerical simulation of fracture propagation Based on the finite element method coupling seepage, stress and damage, 3-D numerical simulation on the fracturing of a horizontal well through interlayer to coal seam was conducted. In the simulation, coupling of the stress field and the seepage field was fully considered in the process of fracture growth. Formation geologic parameters are obtained from the No.3 coal seam and its roof and floor in a depth of approximately 1000 m at Shizhuang block at the southern part of Qin shui Basin. 3.1. Governing equation The strong form for the solid of the initial boundary value problem is as follows:
sij;j þ bi ¼ 0 in U u ¼ u on Gu s$n ¼ t on Gt s$nc ¼ p on Gcrack
Here s is the stress of solid skeleton, bi is the body force, u is a given displacement on displacement boundary, n is the normal of traction boundary, t is a given surface traction on traction boundary. nc is the normal of crack surface and p is the water pressure action on the crack surface. For an incompressible Newtonian fluid, in order to simplicity, we neglect the gravitational forces, and assume that the flow in the fracture is one dimension flow, then the governing equation of fluid motion can be written as:
(1)
(3)
where vw is the seepage velocity (fluid flow velocity relative to the solid skeleton), k is the vector of permeability coefficient, g is the acceleration of gravity, rw is the density of water, pw is the pore pressure, nw is the ratio between the volume of pore water and the total volume.
3.2. Simulation model The 3-D model is built by finite element software ABAQUS. Fig. 10 is the schematic plot of the model. The dimensions of the model are 300 m, 200 m and 80 m in the direction Y, X, Z respectively. The coal seam and the roof and floor are assumed to be fully bonded. The thickness of the roof, coal seam and floor is 40 m, 6 m and 34 m, respectively. To observe the interaction between adjacent hydraulic fractures, two potential fracture propagation faces with cohesive unit properties were set (Zhang et al., 2010). The fracture faces are perpendicular to the minimum horizontal stress direction, which initiates and extends along the XeZ face, forming
Table 2 Parameters in the model.
Fig. 10. Sketch plot of model.
No.
Items
Unit
Coal seam
Roof/floor
1 2 3 4 5 6 7 8 9 10 11 12
Elastic modulus Poisson's ratio Permeability Average leak off coefficient Vertical stress Maximum horizontal stress Minimum horizontal stress Fluid gravity Saturation Pore pressure Pumping rate Fracture space
GPa
5 0.2 0.685 0.0062 30 26 16 9800 1.0 9.0 8 100
30 0.16 1.37 0.0165 35 30 22 9800 1.0 9.0
103 mm2 m/min0.5 MPa MPa MPa N/m3 MPa m3/min m
D.Q. Li et al. / Journal of Natural Gas Science and Engineering 21 (2014) 386e396
393
Fig. 11. Displacement vector distribution at the end of simulation time.
(1) The pore pressure distribution at both sides of the fracture are asymmetric. Fig. 11 shows the final displacement vector graph of three layers. As depicted, the pore pressure between the two fractures increased as a result of the stress superposition which hinders the fracture propagating, and instead, grows towards the outer part. (2) Hydraulic fracture propagates effectively and penetrates the coal seam. The simulation fracture height in the coal seam (40 me46 m at height in model) is 31.67 m, which proves that the fracture initiating at the roof propagates effectively and grows through the coal seam to the floor. (3) The injection pressure curve presents an obvious response of penetrating layers. For better analysis, injection pressure curves at the injection point, the top point of the coal seam and the top point of the floor in the simulated first 20 s are plotted in Fig. 12. The curves show when the hydraulic fracture is breaking into the coal seam from the roof, the pressure at the injection point increases sharply while the breaking pressure of the coal seam decreases immediately. When the fracture penetrated the coal seam and reached the floor, the pressure at the top point of the coal seam increased while the pressure at the top point of floor decreased. So during the process of interlayer to coal seam fracturing, there will be an energy breakthrough effect when a fracture goes from one layer towards another layer. It demonstrated that the injection pressure will increase in the fracturing layer, and after it was fractured, the injection pressure will gradually decrease to reach a normal propagation pressure.
Fig. 12. Injection pressure curves at the first 20 s.
transverse fractures. The initiating point is located at the upper layer, 4.41 m away from the coal seam. For the sake of improving the accuracy of simulation results, LGR (Local grid refinement) was used to define grids of hydraulic fractures and the coal seam. The parameters used in the model are listed in Table 2.
3.3. Simulation results and analysis According to the simulation results, the following conclusions can be drawn: Table 3 Simulation results under different stresses. Vertical stress difference, MPa
Half fracture length, m
Fracture width, cm
Fracture height, m
Breakout pressure, MPa
Treatment pressure at 3600 s, MPa
5 8 10
108.6 102.86 97.15
1.25 1.24 1.18
31.67 30.77 27.50
24.3 24.9 27.6
21.4 22.2 23.4
Table 4 Simulation results under different elasticity modulus. Elastic modulus difference, GPa
Half fracture length, m
Fracture width, cm
Fracture height, m
Breakout pressure, MPa
Treatment pressure at 3600 s
25 20 15
64.26 70.01 81.42
2.51 2.36 2.05
60.34 53.61 47.83
26.7 27.0 27.2
21.7 22.0 22.4
394
D.Q. Li et al. / Journal of Natural Gas Science and Engineering 21 (2014) 386e396
Table 5 Simulation results under different permeabilities. Ki/Km
Half fracture length, m
Fracture width, cm
Fracture height, m
Breakout pressure, MPa
Treatment pressure at 3600 s, MPa
2 3 4 5
108.6 74.29 51.43 45.72
1.251 1.245 1.213 1.178
31.67 31.67 31.67 27.51
24.3 22.2 20.7 20.7
21.4 20.9 20.6 20.1
modulus contrast increases the breakout pressure in the barrier and the fracture length in the coal seam, but decreases the fracture width and height. Therefore, higher elastic modulus in coal seams and lower modulus contrast between layers are more beneficial in forming a longer fracture in coal seams.
Fig. 13. Relationship curve between fracture width and length under different injection rate.
3.4.3. The effect of permeability contrast on fracture propagation Keeps the vertical permeability to horizontal permeability ratio of the barrier and coal seam consistent, simulates and compares the fracture propagation rules when vertical permeability ratios of barrier to coal (Ki/Km) are 2, 3, 4, and 5. Simulation results in Table 5 show that a larger permeability contrast between the outer barrier and the coal seam results in lower treatment pressure and shorter fracture length in the coal seam. Because of an increase in permeability of the barrier, fracturing fluid loss increases, so the growth ability of the fracture in the coal seam is weakened.
3.4. Parametric study on fracture propagation Based on a standard simulation model established in 3.2, the effects of multiple factors including in-situ stress, Young's modulus, permeability and fluid injection rate on fracture propagation, were quantitatively studied. 3.4.1. The effect of in-situ stress contrast on fracture propagation Maintain stresses on the coal seam constant with parameters in Table 2 and maintains the three-dimensional stress proportions of the roof and floor. The in-situ vertical stress contrast of a coal seam and barrier is set at 5 MPa, 8 MPa and 10 MPa. It is recognized from Table 3 that the breakout pressure and the treatment pressure increases while the length, the width and the height of the fracture decreases when the in-situ vertical stress contrast between the barrier and the coal seam enlarges. It suggests that a larger in-situ vertical stress contrast produces containment of the fracture geometry and causes higher bottomhole pressure. However, the half fracture length reaches 97.15 m in the coal seam even under a vertical stress contrast of 10 MPa. This suggests that although the degree of fracture propagation is reduced, the fracture geometry is still very effective for the roof and floor horizontal well fracturing technology. 3.4.2. The effect of modulus contrast on fracture propagation The elastic modulus of coal seams are taken as 5 GPa, 10 GPa and 15 GPa respectively. It can be seen from Table 4 that lower layered
Table 6 Anisotropic permeability design of coal seam and its barriers. Principal stress direction
Permeability (103 mm2) Horizontal maximum Horizontal minimum Vertical stress stress direction stress direction direction
Corresponding X direction model direction Coal seam 0.685 Roof and floor 1.37 of coal seam
Y direction
Z direction
0.328 1.37
0.099 0.952
3.4.4. The effect of injection rate on fracture propagation Fracturing fluid rate at 6 m3/min, 8 m3/min and 10 m3/min are simulated. Fig. 13 shows fracture geometry in height and width. A higher injection rate enhances fracture length, height and width. It reflects a positive correlation between injection rate and fracture geometry. Considering the fracture initiation point (35.58 m), and the coal seam position (40e46 m) in the model height direction and the fracture propagation zone, it was found that the fracture could connect a 6 m thick coal seam even under an injection rate of 6 m3/ min. 3.5. The effect of permeability anisotropy on fracture propagation To compare the impact of permeability anisotropy on fracture propagation, a new set of permeability parameters are listed in Table 6 according to the true triaxial permeability test results. Simulation result reveals the following characteristics: (1) Pore pressure superposition degree is larger around a hydraulic fracture after considering permeability anisotropy. The superposition produces mainly along the direction of the fracture width growth, namely the direction of minimal horizontal principal stress. Take this direction from the top of the coal seam as the considered path. Pore pressure at homogeneous simulation model and anisotropic model along the path are mapped in Fig. 14. It is clear from Fig. 14 that pore pressure generally increases in the direction of the fracture width (Y direction) under the influence of permeability anisotropy. (2) Treatment pressure in the anisotropic model is higher than in the homogeneous model. In this paper, the value is 1 MPa. Meanwhile, affected by the permeability anisotropy, half fracture length decreases to 86.27 m, fracture width and height increases to 1.50 cm and 43.59 m respectively. Compared to the fracture geometry in the homogeneous model, overall fracture morphology becomes higher, wider and shorter. Since the permeability anisotropy is common in coal measure strata, the anisotropic model is much closer to the actual formation,
D.Q. Li et al. / Journal of Natural Gas Science and Engineering 21 (2014) 386e396
395
The lower elastic modulus and permeability difference between layers are beneficial in forming a longer fracture in a coal seam. Larger injection rates can also promote the growth of a hydraulic fracture. 5. Pore pressure superposition degree is larger around a hydraulic fracture after considering permeability anisotropy. Pore pressure generally increases in the direction of the fracture width. Compared to the results in the homogeneous model, treatment pressure in the anisotropic model is higher and the overall fracture morphology becomes higher, wider and shorter.
Acknowledgments
Fig. 14. Pore pressure distribution at the top of coal seam along direction Y.
so the shape and parameters of the fracture under this model are closer to the actual fracturing treatment result.
This study was funded by the National Science and Technology Major Projects (Grant No. 2011ZX05042). We thank China United Coalbed Methane Corporation, Ltd. for its permission to publish this paper. We also thank reviewers and editors for their constructive comments and suggestions on improving the manuscript.
4. Conclusions
References
Through true triaxial hydraulic fracturing simulation and threedimensional numerical simulation we get the following theoretical understanding for carrying out fracturing of a horizontal well through interlayer to coal seam:
Anderson, G., 1981. Effects of friction on hydraulic fracture growth near unbonded interfaces in rocks. SPE J. 21 (1), 21e29. Abass, H.H., 1990. Experimental observations of hydraulic fracture propagation through coal blocks. SPE 21289, presented at SPE Eastern Regional Meeting, Columbus, Ohio, 31 Octobere2 November. Blanton, T.L., 1982. An experimental study of interaction between hydraulically induced and pre-existing fractures, SPE 10847, presented at SPE Unconventional Gas Recovery Symposium, Pittsburgh, Pennsylvania, 16e18 May. Daneshy, A.A., 1978. Hydraulic fracture propagation in layered formations. SPE J. 18 (1), 33e41. Defuel, L., Clark, J., 1981. Hydraulic fracture propagation in layered rock: experimental studies of fracture containment. SPE J. 24 (1), 19e32. Fung, R.L., Vilayakumar, S., Cormack, Donald E., 1987. Calculation of vertical fracture containment in layered formations. SPE Form. Eval. 2 (04), 518e522. Guo, T.K., Zhang, S.C., Qu, Z.Q., Zhou, T., Xiao, Y.S., Gao, J., 2014. Experimental study of hydraulic fracturing for shale by stimulated reservoir volume. Fuel 128, 373e380. Hanson, M.E., Anderson, G.D., Schaffer, R.I., Hearst, J.R., Haimson, B., Cleary, M.P., 1978. LLL Gas Stimulation Program. Quarterly Progress Report. Lawrence Livermore National Lab.. JaneMar. Hanson, M.E., Anderson, G.D., Shaffer, R.J., 1980. Theoretical and experimental research on hydraulic fracturing. J. Energy Resour. Technol. 102 (2), 92e98. Hsiao, C., Rabaa, E., Wadood, A., 1987. Fracture toughness testing of rock cores. ARMA 87-0141, presented at the 28th U.S. Symposium on Rock Mechanics (USRMS), Tucson, Arizona, 29 Junee1 July. Li, A.Q., Jiang, H., Chen, C.H., 2004. Practice of hydraulic fracturing and fracturing model selection analysis in China. Nat. Gas Ind. 24 (5), 91e93. Li, L.D., Zhang, S.C., Geng, M., 2010. CBM reservoir hydraulic fracture propagation rule. Nat. Gas Ind. 30 (2), 72e74. Lu, C.H., 2011. Study on Stability and Mechanical Properties of Deep-Seated Coal. PhD thesis. Anhui University of Science and Technology, China. Meng, Z.P., Zhang, J.C., Wang, Rui, 2011. In-situ stress, pore pressure and stressdependent permeability in the Southern Qinshui Basin. Int. J. Rock Mech. Min. Sci. 48 (1), 122e131. Rahim, Z., Holditch, S.A., 1995. The effects of mechanical properties and selection of completion interval upon the created and propped fracture dimensions in layered reservoirs. J. Petrol. Sci. Eng. 13 (1), 29e45. Shan, X.J., Zhang, S.C., Li, A.Q., 2005. Hydraulic fracture propagation rules analysis for coalbed methane well. Nat. Gas Ind. 25 (1), 130e132. Simonson, E.R., Abou-Sayed, A.S., Clifton, R.J., 1978. Containment of massive hydraulic fractures. SPE J. 18 (01), 27e32. Smith, M.B., Bale, A.B., Britt, L.K., Klein H.H., Siebrits E, Dang X., 2001. Layered modulus effects on fracture propagation, proppant placement, and fracture modeling. SPE 71654, presented at SPE Annual Technical Conference and Exhibition, New Orleans, Louisiana, 30 Septembere3 October. Settari, A., 1988. Quantitative analysis of factors influencing vertical and lateral fracture growth. SPE Prod. Eng. 3 (03), 310e322. Thiercelin, M., Jeffrey, R.G., Naceur, K.B., 1989. Influence of fracture toughness on the geometry of hydraulic fractures. SPE Prod. Eng. 4 (04), 435e442. Wright, C.A., Tanigawa, J.J., Mei, S.X., Lei, Z.G., 1995. Enhanced hydraulic fracture technology for a coal seam reservoir in Central China, presented at International Meeting on Petroleum Engineering, Beijing, China, 14e17 November. Warpinski, N.R., Clark, J.A., Schmidt, R.A., Clarence, W.H., 1982. Laboratory investigation on the-effect of in-situ stresses on hydraulic fracture containment. SPE J. 22 (3), 333e340.
1. In-situ stress conditions have a significant impact on the hydraulic fracture morphology. Where differences of vertical stress and horizontal maximum stress are small (1 MPa), the control action of stresses on fracture morphology is weak. The hydraulic fractures can be horizontal, vertical or complex. Where differences of vertical stress and horizontal maximum stress are larger (5 MPa), the control action of stresses on fracture morphology are enhanced, and the hydraulic fractures are generally vertical. 2. The impact of natural fractures on hydraulic fracture propagation are related to the developmental scale of the natural fracture, the angle of approach and the injection pressure. If the opening width of a natural fracture is high (0.05 mm) and the injection pressure of the hydraulic fracture is low (10e15 MPa), the hydraulic fracture tends to divert and propagate along the natural fracture. If the opening width of the natural fracture is low (0.01 mm) and the injection pressure of the hydraulic fracture is high (30e40 MPa), the hydraulic fracture tends to penetrate the natural fracture and propagate along the original direction. When injection pressure in the hydraulic fracture is higher, the impact of the natural fracture on the hydraulic fracture will be weakened. Compared with the development degree of the natural fracture, the impact of the angle of approach between the natural fracture and the hydraulic fracture on fracture propagation is smaller. 3. When developing coalbed methane with interlayer to coal seam fracturing technology of a horizontal well in the roof and floor of a coal seam, coal bearing strata with larger differences (5 MPa) between vertical stress and horizontal stress should be selected. One should try to avoid natural fractures in developed formations, as much as possible. If the construction is implemented in such a formation, it is suggested to increase the operating injection rate so as to ensure the success of the project. 4. Inter-layer geological conditions and fracturing construction conditions have an important influence on hydraulic fracture parameters. Higher vertical stress contrast between layers presents a stronger inhibitory effect on the fracture propagation.
396
D.Q. Li et al. / Journal of Natural Gas Science and Engineering 21 (2014) 386e396
Warpinski, N.R., Defuel, L.W., 1987. Influence of geologic discontinuities on hydraulic fracture propagation (includes associated papers 17011 and 17074). J. Petrol. Tech. 39 (2), 209e220. Yang, J.S., Wang, Y.B., Li, A.Q., 2012. Experimental study on propagation mechanism of complex hydraulic fracture in coal-bed. J. Chin. Coal. Soc. 37 (1), 73e76. Zhao, L.J., Qin, Y., 2010. Research statement of domestic deep coalbed methane. China Coal. Meth. 3, 38e40.
Zhang, S.R., 2011. Outlook of effective development way for deep-Buried coalbed methane. China Coal. Meth. 4, 18e21. Zhou, J., Chen, M., Jin, Y., Zhang, G.Q., 2008. Analysis of fracture propagation behavior and fracture geometry using a tri-axial fracturing system in naturally fractured reservoirs. Int. J. Rock Mech. Min. Sci. 45 (7), 1143e1152. Zhang, G.M., Liu, H., Zhang, J., Wu, H.A., Wang, X.X., 2010. Three-dimensional finite element simulation and parametric study for horizontal well hydraulic fracture. J. Petrol. Sci. Eng. 72, 310e317.