- Email: [email protected]
Integrated fracability assessment methodology for unconventional naturally fractured reservoirs: Bridging the gap between geophysics and production
Author’s Accepted Manuscript Integrated Fracability Assessment Methodology for Unconventional Naturally Fractured Reservoirs: Bridging the Gap between Geophysics and Production Zhi Geng, Mian Chen, Yan Jin, Xin Fang, Ning Jing www.elsevier.com/locate/petrol
PII: DOI: Reference:
S0920-4105(16)30252-2 http://dx.doi.org/10.1016/j.petrol.2016.06.034 PETROL3526
To appear in: Journal of Petroleum Science and Engineering Received date: 10 March 2016 Revised date: 13 June 2016 Accepted date: 21 June 2016 Cite this article as: Zhi Geng, Mian Chen, Yan Jin, Xin Fang and Ning Jing, Integrated Fracability Assessment Methodology for Unconventional Naturally Fractured Reservoirs: Bridging the Gap between Geophysics and Production, Journal of Petroleum Science and Engineering, http://dx.doi.org/10.1016/j.petrol.2016.06.034 This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting galley proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
Integrated Fracability Assessment Methodology for Unconventional Naturally Fractured Reservoirs: Bridging the Gap between Geophysics and Production Zhi Genga,b , Mian Chena , Yan Jina,∗, Xin Fanga , Ning Jingc a College
of Petroleum Engineering, China University of Petroleum, Beijing, China de G´ eologie, Ecole Normale Sup´ erieure, CNRS UMR 8538, Paris, France c CNPC International (Chad) Ltd.
b Laboratoire
Abstract Fracability evaluation in unconventional reservoirs is one of the most important initiatives for economic success. Most evaluation methods in geophysics and well logging are based on calculating rock mechanical properties, in which a frequently used term is brittleness. However, oil & gas production, mainly determined by effective natural fracture networks in unconventional low-permeability reservoirs, is the final goal of reservoir stimulation. In this work, a novel integrated methodology was proposed to correlate conventional post-stack seismic data to well production rate. It aims to pre-evaluate regional stimulation performance, facilitating investment decision-making in line with global oil prices and environmental consideration. Structural occurrence, seismic maximum curvature, and geomechanics parameters were combined in a physical model to evaluate underlying development and activation of natural fractures, result in the activation index F1 . The potential of artificial induced fractures were quantified as normalized index F2 by integrating mechanical item based on fracture mechanics and fluid flow. Accordingly, a fracability sweetness index (Fsweet ) model was built to correlate standardized well production rate (Qef f ). A Case study of a naturally fractured igneous reservoir with 8 wells showed agreement between Fsweet and Qef f . A considerable part of highly active fractures with ∗ Corresponding
author Email addresses: [email protected] (Zhi Geng), [email protected] (Yan Jin)
Preprint submitted to Journal of Petroleum Science and Engineering
June 21, 2016
F1 of [0.9, 1] distributed over the area with lower curvature value. By contrast, the majority of fractures with lower F1 of [0.8, 0.9] located in zone with higher curvature value. The weight ratio α = 0.06 indicated that in this case the overwhelming production contribution were from natural fractures, which were evidenced by well core picture. The field data demonstrates the prospect for predicting regional fracability variation, and suggests that it is possible to correlate fracability index with standardized well production rate. The insights contribute to well placement optimization and stimulation treatment design, breaking limitation of engineers0 influence on individual wells, and hence reducing exploitation cost while enhancing ultimate recovery. Keywords: fracability, seismic curvature, fracture density, fracture activation, unconventional reservoir 2015 MSC: 00-01, 99-00
1. Introduction Confronting the worldwide oil price shocks, cost-efficient development in hydrocarbon reservoirs plays a key role in economic success, especially in unconventional reservoirs with low or ultra-low permeability. Fracability evaluation 5
in unconventional naturally fractured reservoirs is one of the most important initiatives. Fracability is generally defined as the capability of reservoirs to be fracture stimulated effectively [1]. Most of the fracability evaluation methods developed from geophysics and laboratory experiments focus on calculating rock mechanical properties, among which a frequently associated term is brittleness.
10
One popular way to predict fracability/brittleness is based on geophysical seismic inversion of rock elastic parameters, such as Young0 s modules, Poisson0 s ratio, Lame0 s constants, etc. Corresponding prediction models are defined by combination of these elastic parameters. For example, Varga et al. (2012) calculated brittleness index using formula defined by Young0 s modules and Poisson0 s
15
ratio, which was estimated by using simultaneous AVO inversion theory [2]. Metzner et al. (2013) found that Mu-Rho (rock0 s shear rigidity and bulk den-
2
sity) attribute, documented by Goodway [3], show the best correlation with their estimated ultimate recovery, indicating good fracability. Similarly, Sharma et al. (2012) [4] and Shereef et al. (2014) [5] used the product of Young0 s modules and 20
bulk density as a brittleness indicator. In addition, Sun et al. (2013) integrated brittleness and fractures which were predicted by inverting limited azimuthal seismic data to estimate preferable area for drilling locations [6]. Rock lithology and mineral content estimated by seismic measurement [7] and well logging [8],[9],[10] are also used as brittleness indicators. Moreover, Guo et al. (2014)
25
evaluated shale reservoir fracability by various experimental tests [11]. Oil & gas production, mainly determined by activated natural fracture networks in unconventional reservoirs with low permeability, is the final goal of reservoir stimulation. However, little attention has been paid to productionoriented regional fracability characterization and prediction. The more under-
30
lying activated natural fractures, the more probably being interlinked by stimulation and consequently enhance ultimate recovery. In this paper we presented a novel integrated fracability assessment methodology applied in unconventional naturally fractured reservoirs, by correlating conventional post-stack seismic data to potential well production capacity us-
35
ing physical models. The insights contribute to well placement optimization and stimulation treatment design, breaking limitation of engineers0 influence on individual wells and hence reducing exploitation cost while enhancing ultimate recovery.
2. Methodology 40
With the similar stimulation treatment in unconventional reservoirs, natural production rate is dominated by two factors: development & activation of natural fractures, and quality of artificial induced fractures. In this paper, the term “artificial induced fractures” means the fractures/cracks to be hydraulic fractured by engineering operation. The curvature of geological structure sur-
45
faces can be used to assess the degree of strain, which is expressed as brittle
3
fracture relevant to natural fractures [12]. Intersection of natural fractures and induced fractures by hydraulic fracturing will lead to successful stimulation [13]. The more connections between fractures, the more promotion to reservoir performance. 50
Taking above into consideration, our fracability assessment is well production oriented. A normalized index of fracability sweetness was defined to evaluate regional fracability distribution, which was correlated to well production (oil & water) rate. As artificial stimulation treatment is inevitably limited by natural properties of rock under in-situ conditions, we gave a new horizon to characterize
55
regional fracability variation, breaking through the stimulation design efforts focused on individual well profiles in depth. Once a target formation is selected to be stimulated, then the regional variation of fracability sweetness can be used to deploy new drilling wells which are more favorable to be fractured. For drilled wells, it aims to pre-evaluate stimulation performance, avoiding investment and
60
environmental risks. The datasets used in our method include: 1. structural attributes (curvatures, azimuth and dip angle) extracted from post-stack seismic data in depth domain; 2. far field stress distribution interpreted from seismic data; 3. production data of drilled wells. For structural attributes and stress distribution
65
data, exhaustive description or interpretation can be found in Roberts (2001) [14], Sena et al. (2011) [15], and Metzner et al. (2013) [16]. It is noteworthy that the datasets above are the most commonly used in petroleum industry, no more specialized techniques or tools are needed. 2.1. Development & Activation of Natural Fractures
70
Structural curvature is a key indicator from which the current, local, stress/strain is illustrated [17]. In this study, seismic structural attributes were used to characterize geometry of geological surfaces. Seismic curvature attributes were introduced in the mid-1990s and shown to be highly correlated with fractures, some of which were measured on outcrops [14],[18],[19]. The detailed introduc-
75
tion and calculation methods of various curvatures were discussed in references 4
[14],[20],[21]. The most important are maximum and minimum curvature, which are also called principle curvatures and used to calculate other curvatures. Previous studies suggested that structural position takes over as the most important controlling factor on fracture formation [22]. The positive maximum curvature 80
indicates anticline structure and more natural fractures, while negative minimum curvature means syncline structure and less fractures or closed fractures. As maximum curvature attribute is highly sensitive to brittle deformation, it is reasonable to suppose that the higher value of maximum curvature, the stronger possibility that fractures develop.
85
Assuming in an ideal situation, underground rock is liner elastic, and brittle failure hasn0 t occurred in suffering tectonic deformation. As illustrated in Figure 1, through one point on geological surface, there is exactly a plane normal to the dip vector which has the same azimuth as maximum curvature . The plane is treated as a fictitious fracture model. Then development & activation
90
of potential natural fractures in the vicinity of the point can be estimated by calculating failure tendencies of the fracture model under the action of far field stress and structural stress indicated by structural curvature. It is a reasonable assumption to analyze mechanics characteristic of the fracture model during elastic deformation until the moment of brittle failure. If the fracture model
95
fails under loading, then relatively active fractures exist with extremely high probability. On the contrary, even though there might be some fractures result from geological processes or environmental influence, they are probably inactive or closed, which is not conducive to fluid flow. The induced stress experienced at the top of the structure can be estimated
100
from [23]: σinduce =
hKE 2
(1)
where σinduce is structural induced stress, h is the thickness of layer, K represents the maximum curvature in three-dimensions [14], E is the Young0 s modulus of rock. Su et al. (2014) assessed the likelihood of natrual fractures activation along
5
Z (Depth)
K min
K max lin
e)
N
Y
(In
φ
X (Crossline) Figure 1: Schematic diagram of fracture model (rectangle with tiny aperture). Kmax is maximum curvature of geological surface (gray color), ϕ is azimuth of Kmax . Kmin is minimum curvature which is orthogonal to Kmax . The strike of fracture model (ϕf ) aligns with the azimuth of Kmin , so is orthogonal to ϕ. The dip angle of fracture model is where DipKmax is the dip angle of surface in the direction of Kmax .
6
π 2
− DipKmax ,
105
wellbores based on effective stress on a fracture by simple projection of the far field stress on its plane according to the normal vector of the plane in principle stress coordinate [13]. For the convenience of simple calculation using geological terms such as azimuth, dip of fractures and far field stress, we used tensor transformation method described by Zoback et al. (2010) [24] to quantify the
110
normal stress Sn and shear stress Sτ on arbitrary fracture planes: [Sn , Sτ ] = f (SH , ϕSH , ϕf , Dipf )
(2)
where SH , ϕSH is the value and azimuth of maximum horizontal stress separately; ϕf , Dipf is the strike/azimuth and dip angle of fracture model respectively. The activation of the fracture model is then estimated by combining Equa115
tion 1, Equation 2, Mohr-Coulomb and tension failure criteria, given by: Sτ , S ≤ τ τ m τm F1 = 1, S > τ or S ≤ σ +P τ
m
n
induce
(3)
p
where τm = µ (Sn − σinduce − Pp ) is shear failure stress of the fracture model, µ is friction coefficient which is normally 0.6 in many cases [24],[25], Pp is pore pressure or fluid pressure. The activation index F1 ∈ [0, 1]. F1 = 0 indicates there are no active fractures, while F1 = 1 shows effective fractures abound. 120
Although the activation index here is not a solid evidence of objective existence of specific activated fracture, it implies the regional variation and distribution of relative activity of dormant natural fractures. 2.2. Potential of Artificial Induced fractures Besides the engineering treatment influence on the shape and quality of
125
artificial fractures, rock mechanical properties are considered as the basic determinant factors contributing to development of induced fractures as well as the fluid flow. Flow rate through a planar fracture integrated with aperture was given by Equation 4 [24], and the critical stress value for cracks extension based
7
130
on Griffith0 s theory of rupture was defined in Equation 5 [26]: " #3 π L 1 − ν 2 (pf − S3 ) Q= ∇p 8η E s πEγ Pc = 2c (1 − ν 2 )
(4)
(5)
where Q is volumetric flow rate in the fracture, η is fluid viscosity, L and c are fracture length, item pf − S3 is net pressure in the fracture, ∇p is pressure gradient, γ is fracture surface energy, ν is Poisson0 s ratio and E is Young0 s
135
modulus. By comparison of above two equations, the shared rock mechanics item 1 − ν 2 /E is positively correlated with Q, and negatively related with Pc . It suggests that the higher value of the item, the better for extension of induced fractures (lower Pc ) as well as the fluid flow result from large fracture aperture (higher Q). Stimulation treatment and geomechanics difference is inevitable whether for
140
individual wells or in regional area. As a consequence, it seems plausible to evaluate the variation of potential for induced fractures based on objective and basic rock mechanical properties. Accordingly, a normalized index of potential for induced fractures is defined as: fi −fi min fi max −fi min 2 i where fi = 1−ν Ei
F2 =
(6)
where fi is the value of item 1 − ν 2 /E for specific point in field area; fi max , 145
fi min is the maximum and minimum value for all the points in the area respectively; F2 ∈ [0, 1] is the corresponding normalized index. The higher value of F2 , the better situation for creation of induced fractures with large aperture and fluid flow. The normalization of index is applicable to characterize the relative variation of rock mechanical properties in different location, so that it enlightens
150
the quality of artificial induced fractures. 2.3. Modeling Fracability Sweetness In unconventional naturally fractured reservoirs, well production rate is generally contributed from natural fracture networks and the extent of artificial 8
fractures. In order to eliminate the impacts of formation pressure and comple155
tion conditions on production rate, the effective individual well production rate is standardized using: Qef f =
Qt HT ∆P
(7)
where Qt is cumulative production, H is the thickness of stimulated reservoir, T is production days, ∆P is production pressure difference. Assuming that normalized individual well production rate is liner combination of contribution 160
from natural fractures and induced fractures, then a correlation with two weight coefficients can be established. Afterwards, the correlation can be defined using index F1 , F2 : Qef f i = c1 × F1 i + c2 × F2 i
(8)
where Qef f i is standardized production rate of the ith well, F1 i is activation index in the vicinity of the ith well, F2 i is the corresponding index for induced 165
fractures, c1 and c2 are weight coefficients respectively which can be calculated by using least square method. With ratio of α = c2 /c1 , the model of fracability sweetness is given: Fsweet = F1 + αF2
(9)
where Fsweet ∈ [0, 2] is index of fracability sweetness, which is the higher the better. The ratio α is convenient to show the relative contribution of effective 170
natural fractures and induced fractures. The index model is then capable of evaluating regional variation of fracability sweetness in a unified pattern.
3. Model Verification with Case Study By evaluating the activation of fracture model and potential of artificial induced fractures in the use of seismic structural attributes and interpretation, 175
the correlation between fracability sweetness index and standardized production rate was established. In order to validate our fracability evaluation approach, a case study of naturally fractured igneous reservoir in western China was shown in the sections bellow.
9
The target zone is Permian system volcanic rock with low formation dip angle 180
(6∼9◦ ), covering an area of 100 km2 . The CMP bin of 3D seismic data is 25×25 m. Wave group characteristics are clear, and the seismic data is in good quality. The effective bandwidth near target zone is 10∼43 Hz, with dominant frequency of 18 Hz. The oil-gas-bearing volcanic rock is mainly composed of basalt and tuff. Mineral components are plagioclase (60∼72%), quartz (9∼15%), calcite
185
(7∼13%), pyroxene (3∼9%), chlorite (4∼9%), and little magnetite (0∼4%). The average thickness of the reservoir is 16 m, with burial depth of 3225∼3395 m. The maximum horizontal stress gradient SH = 2.7 with azimuth of 105◦ , and the minimum horizontal stress gradient Sh = 2.3. The stress state of the area is strike-slip or reverse faulting according to E. M. Anderson0 s (1951) classification
190
scheme [27]. There are 7 vertical wells which were hydraulic fractured with similar treatment, and 1 horizontal well without being stimulated in the field. The high conductive natural fractures, which were observed on some imaging logs, have high dip angle. However, one obvious limitation of imaging tools is that they will especially undersample fractures and faults whose planes are
195
nearly parallel to the wellbore axis [24]. So, in this case the imaging logs could hardly reflect the development and density of fractures nearby wells in the block. As the permeability of the rock matrix is extremely low, the fluid flow in the reservoir can only be achieved through fracture networks, which was confirmed in Figure 2. The evidence points to the fact that in this case, natural fractures
200
contribute dominantly to the production. The distribution of original seismic maximum curvature was shown in Figure 3. Since the inset histogram indicated that the scope of curvature data was mainly 0∼0.2, the host graph was color coded in [0, 0.2]. High value of maximum curvature in red color qualitatively implied dense natural fractures,
205
which concentrated on band zone around wells N32∼N10. While the maximum curvature value around well N11, M3, P8 and N32 were obviously low, which suggested scarce fractures. So was the same situation in the neighboring area around the horizontal section of H1. As described in Section 2.1, the underlying development and activation of 10
Figure 2: Reservoir core from burial depth of 3285 m. Macroscopic oil traces coincide exactly with natural fractures.
11
2000
0.2
1950
0.16
N11 1900
0.12
Inline
M3 P8
1850
N32
P6
H1
0.08
1800
N31 0.04
1750
N10 1700
1550
1500
1450
Crossline
1400
0
Figure 3: The distribution of original seismic maximum curvature. The inset histogram showed data interval of [0, 0.4]. Since the curvature data mainly concentrated in 0∼0.2, the host graph was color coded in [0, 0.2]. Green circles symbolized wells location, the radius of circles was 150 m, which was the estimation of stimulation extent. The lateral length of H1 was 400 m.
12
210
natural fractures was depicted in Figure 4. The azimuth and activation of fracture models were visualized in Figure 4a, in which the back ground (base map) was gray-colored as the same value of maximum curvature in Figure 3. The solid lines in red or blue were color coded by F1 (activation index) in interval of [0.8, 1] for visual friendly and density display of underlying highly
215
active fractures. It should be noted that a considerable part of the relative highly active fractures (red color) as interval [0.9, 1] distributed over the area with lower curvature value (dark gray color). By contrast, the majority of fractures with lower F1 (blue color) as interval [0.8, 0.9] located in zone with higher curvature value (bright gray color). Even though the overall curvature
220
value on upper part of Figure 3 was relatively low, the highly active fractures were common as seen in Figure 4a. It appears that maximum curvature in Figure 3 implied the existence of natural fractures in some degree. However, it0 s obviously insufficient when compared with the method proposed in this paper. The Figure 4b is a different
225
manifestation of F1 distribution in the whole data interval of [0, 1]. For the comparison with following figures, Figure 4b was color coded in [0.4, 1], which means the color of data in [0, 0.4] is the same. Figure 4b shows much more clearly preference to Figure 3 when characterizing development and activation of natural fractures, especially highly active fractures. In addition, Figure 4
230
suggests that both activation and density of fractures around well N11, M3 and P8 was higher than that in Figure 3 (maximum curvature). What0 s more, the corresponding value near well N32 and horizontal section of H1 were still in depress. Figure 5 displays further evidence of our above results. Although it should
235
be cautious to use image logs to reveal development of fractures/faults near wellbore, they provide macro-understanding of relative fracture density. As shown in the figure, the fractures/cracks in image P6 and M3 are obviously much more than that in image N32. This coincides with F1 distribution in Figure 4. However, according to the maximum curvature in Figure 3, the fracture density
240
of well M3 and N32 should be in the same level, which was much lower than 13
2000
0.2
1950
N11
0.16
1900 0.12
Inline
M3 P8
1850
N32
P6 1800
H1
0.08
N31 0.04
1750
N10 1700
1550
1500
1450
Crossline
1400
0
(a) The azimuth and activation of fracture models were shown as solid lines in red or blue, which were color coded by activation index (F1 ) in interval of [0.8, 1], with grid increment of 2 for visual friendly. The back ground (base map) was gray-colored by the same value of maximum curvature in Figure 3, which shared the same inset histogram.
14
2000
1.0
1950
0.9
N11
1900
0.8
Inline
M3 P8
1850
N32
P6 1800
0.7
H1 0.6
N31
0.5
1750
N10 1700
1550
1500
1450
Crossline
1400
0.4
(b) The distribution of F1 in the whole interval of [0, 1], which was displayed in inset histogram. The higher value the better. For the comparison with following figures, F1 was color coded in [0.4, 1], which means data interval of [0, 0.4] shared the same color. Figure 4: Underlying development and activation distribution of natural fractures.
15
that of well P6.
P6
M3
N32
3326
Measured Depth (m)
3297.5 3239
3327
3298.5 3240
0
180
360
0
180
360
0
180
360
Azimuth (°) Figure 5: Reservoir image logs of 3 wells (P6, M3 and N32).
The vertical black strips
(azimuth≈185∼200◦ , azimuth≈10∼20◦ ) in each image indicate well collapse induced by maximum horizontal stress. The fractures/cracks in image P6 and M3 are obviously much more than that in image N32.
The normalized index of potential for induced fractures (F2 ) was shown in Figure 6 The area with warm color was in favor of creating artificial induced fractures and fluid flow from perspective of rock mechanical properties. Owing 245
to the increasing depth of reservoir from east to west, the rock elastic modulus was higher in western region. As a consequence, the figure implies that the western region with lower F2 was not ideal for initiation of artificial fractures with beneficial aperture. Note once again that up to now all attributes around well N32 and horizontal section of H1 hold low level.
250
The final result of fracability sweetness index (Fsweet ) was presented in Figure 7. Since well H1 wasn0 t stimulated, the production data of 7 vertical wells
16
1
2000
1950
0.8
N11
1900 0.6
Inline
M3 P8
1850
N32
P6 1800
H1
0.4
N31 0.2
1750
N10 1700
1550
1500
1450
Crossline
1400
0
Figure 6: Distribution of normalized index of potential for induced fractures (F2 ). The inset histogram showed that F2 distributed in [0, 1]. The warmer color it was, the better situation for induced fractures.
17
in combination with related attributes were used to build fracability sweetness model. The attributes value around wells were root-mean-square of all the grid points in the circles showed in figures. The ratio α = 0.06 indicated the over255
whelming production contribution of natural fractures, which were evidenced by Figure 2. As real fractures in rock with ultra low matrix permeability will not have perfectly smooth surfaces [28], the fractures with finite aperture still can enhance flow to some extent [24]. Hence, the index F2 in this case was further from valuable than it would be in cases that high conductive fractures with
260
large aperture come first. It appeared that overall visual effect of Figure 7 and Figure 4b was similar. One primary cause was the tiny value of α. However, the inset histograms in each figure clearly showed that the data in Figure 7 were more uniformly distributed, which suggested rich recognition of fracability sweetness.
265
The method proposed was verified by comparing data including standardized production rate (Qef f ), maximum curvature (Kmax ), and fracability sweetness index (Fsweet ), as shown in Figure 8. As mentioned above, all attributes around well N32 held low level, so was its fracability sweetness. Since well H1 wasn0 t simulated, it wasn0 t used to build fracability model and its standardized pro-
270
duction rate was extremely low. It has been demonstrated that the maximum curvature alone is easily affected by noise and unstable. In our results, for example, the curvature value of well P6 was abnormally high, while the value of well P8 was excessively low. The whole line of Kmax hardly matched the trend of Qef f . By contrast, the fracability sweetness curve which was obtained by
275
combining F1 and F2 coincided best with the trend of Qef f , especially on well P8 and N10 when compared with the curve of F1 , let alone F2 . In this paper, c1 Fsweet indicated standardized production rate of wells. It could be inferred that after well H1 being stimulated, its productivity must be higher than well N320 s, especially considering the long horizontal section.
280
However, this work represents only a preliminary attempt to establish such a correlation. The actual relationship between natural & artificial induced fractures and well production rate may be more complex. Nevertheless, these results 18
2000
1.0
1950
0.9
N11
1900
0.8
Inline
M3 P8
1850
N32
P6 1800
0.7
H1 0.6
N31
1750
0.5
N10 0
1700
1550
1500
1450
Crossline
1400
0.4
Figure 7: Distribution of fracability sweetness index (Fsweet ). The inset histogram showed that Fsweet located in [0.1, 1.1]. The host graph was color coded in [0.4, 1] for comparison with Figure 4b. The red color ([0.7, 1]) represented superior fracability, while blue color suggested inferior ones.
19
Standardized Production Rate (t/day/m/kpa)
F1 F2
12
Fsweet Kmax Production
0.3
0.9
10 0.2
8
0.9 0.7 0.8
6
0.1 0.5
4 2 0
1.0
0.7
0.0 0.3 0.6 P6
P8 N10 N31 M3 N11 N32 H1
well name
Figure 8: Comparison of standardized well production rate (Qef f , column bars) with maximum curvature (Kmax , red line with stars) and fracability sweetness index (Fsweet , green line with circles).
suggest that the fracability sweetness index (Fsweet ) obtained based on seismic and production data may provide more meaningful information for assessing the 285
profitable area to be exploited than traditional methods.
4. Discussion Our results provide compelling evidence for rationality and applicability of the proposed methodology to characterize regional fracability sweetness in unconventional naturally fractured reservoirs. It suggests that this approach ap290
pears to be effective in assessment of natural fractures activation and potential for induced fractures with aperture. As it0 s so difficult to strictly characterize complex reservoirs only using physical equations, it seems prudent to normalize the key physical parameters so as to ignore uncertain factors when uncover the regional relative variation. Our evaluation contributes to decision making for
295
stimulation in line with global oil prices as well as environmental consideration. As discussed in case study, the potential productivity of well H1 with stimu20
lation can be inferred, at least providing a lower limit. Accordingly, rational investment can be done depending on the current market. 4.1. The reasons for using seismic curvature 300
The prediction or inference of natural fractures in field using seismic data are generally accepted. The seismic data used in this purpose can be divided into two kinds: the pre-stack data and post-stack data. In theory, the methods using azimuth offset data (pre-stack) are detection techniques for fractures occurrence and density. Although they are widely limited by rigorous theoretical
305
preconditions for fracture angle and alignment, the most concern here is their extremely high cost for acquisition, processing and interpretation, let alone most seismic pre-stack data are limited in azimuth in China [6]. By contrast, the seismic curvature calculated by using relative low-cost post-stack data is a useful attribute set for making predictions or inference regarding fracture density [29].
310
Ericsson et al. (1988) showed the relationship between reservoir production and seismic curvature [30]. Lisle (1994) discussed the correlation of Gaussian curvature to fractures measured on an outcrop [18]. Roberts (2001) documented in details regarding the relationship between fractures and surface shape determined by various curvatures, such as maximum curvatures, Gaussian curvatures,
315
total curvatures and mean curvatures etc.[14]. Hart et al. (2002) predicted that large value of strike curvatures were correlated to open fractures [31]. Gao (2013) integrated 3D curvatures for fracture characterization [32]. Most of previous research have used curvatures as indicators to predict natural fracture density.
320
4.2. The features in the methodology As mentioned above, brittleness is generally used as an indicator for fracability. Increasing brittleness means there may be natural fractures present and good for hydraulic stimulation [33]. However, this assumption is not always true because formation with higher brittleness can also be a fracture barrier
325
[34]. The reason is that conventional concept of rock brittleness qualitatively
21
assumes it increases with higher Young0 s modulus and lower Poisson0 s ratio [33]. In the meanwhile, however, the strength and fracture toughness of rock keeps growing with increasing Young0 s modulus, which is unfavorable for initiation and extension of induced fractures. 330
In this study, the principles for extension and fluid flow in aperture of fractures were analyzed basing on fracture mechanics and flow model, which enlightens the nature of rock mechanical properties that in favor of induced fractures. According to the item 1 − ν 2 /E, the ideal situation for induced fractures and fluid flow is lower value for both Young0 s modulus and Poisson0 s ratio. If the
335
Poisson0 s ratio keeps constant, the rock with lower Young0 s modulus is easier to break; and if the Young0 s modulus remains unchanged, the fractures in rock with lower Poisson0 s ratio would retard healing to closure. Besides the view of rock itself, excessive Young0 s modulus and stress in rock also pose new challenges to the selection of costly proppant with better quality, which is indispensable to
340
keep fractures open after reservoirs being hydraulic fractured. Furthermore, fracability evaluation is more than just calculating rock properties [35]. The development and activation of effective natural fractures play key role in exploitation of unconventional reservoirs. Although the natural fracture density can be qualitatively inferred by combing multiple seismic curvatures
345
[14], completion and production behavior doesn0 t necessarily depend heavily on fracture density. The reason behind is that some of the fractures with certain occurrence (strike and dip angle) are probably closed (dormant) in present-day stress field. In other words, these fractures are non-effective. One thing to keep in mind is that activation degree of fractures has a bearing on the flow capacity
350
of the reservoir [17]. Therefore, by analyzing mechanical behavior of geological fracture models, this paper lays emphasis on activation evaluation of fractures under present-day far field stress and structural stress, which is indicated by structural curvature.
22
4.3. The limitations 355
The limitations are inevitable in the fracability evaluation method proposed. Seismic curvatures are not immune to noise, whose influence is sample offset (grid spacing) related [12]. The seismic data can also contain false curvature features e.g. velocity pull-ups and push-downs and apparent faulting created by noise contamination [36]. In addition, although the relations between curvature
360
and fracture density are well supported by related geological studies, there are three assumed conditions suggested by Nelson [37]. They are: 1. the rock is brittle and fails largely from fracturing, 2. an increase in curvature implies an increase in strain, and 3. an increase in strain implies an increase in fracture density. It implies that our method is applicable to evaluate structural fractures
365
except that induced by factors such as evolutionary history, environmental conditions, etc. Moreover, it still matters for seismic interpretation accuracy of geomechanics data, such as rock elastic parameters, in-situ stress, etc. Nevertheless, since being calculated using post-stack seismic data, curvature is not only computationally cheaper but also more robust than many fracture
370
density indicators [17]. Besides, by normalization processing of key parameters for regional variation assessment, the negative impact of errors between true value and estimated value on specific point in area can be weakened in some degree.
5. Conclusion 375
In this paper, an integrated fracability assessment methodology was presented by evaluating activation of natural fractures and potential of artificial induced fractures with aperture. The fracability sweetness index was calculated by using available conventional post-stack seismic data and production logs, revealing the correlation between exploration data and well production rate. The
380
insights contribute to drilling locations optimization and stimulation treatment design, breaking limitation of engineers0 influence on individual wells. The case study showed feasibility of the novel method in unconventional naturally frac-
23
tured reservoirs. The productivity of a pre-stimulated well was inferred by using the fracability model, which benefits decision making for stimulation in line with 385
global oil prices and environmental consideration. Our results are encouraging and should be validated in more cases. However, the production mechanisms underlying natural & induced fracture networks remain to be further studied. Future work should be focused on two aspects: 1. time dependent effectiveness of active fractures for fluid flow, 2. influence
390
factors and mechanisms of fluid flow in naturally fractured reservoirs. On the whole, the evaluation of stimulation potential should be promoted throughout the complete reservoir life cycle.
Acknowledgements The authors gratefully acknowledge the financial supports (Major Program 395
No.51490651) from National Natural Science Foundation of China. We also thank the anonymous reviewers for their valuable comments and suggestions on our article.
References [1] K. K. Chong, W. V. Grieser, A. Passman, H. C. Tamayo, N. Modeland, 400
B. E. Burke, et al., A completions guide book to shale-play development: A review of successful approaches toward shale-play stimulation in the last two decades, in: Canadian Unconventional Resources and International Petroleum Conference, Society of Petroleum Engineers, 2010. [2] R. Varga, R. Lotti, A. Pachos, T. Holden, I. Marini, E. S. J. Pendrel, Seis-
405
mic inversion in the Barnett Shale successfully pinpoints sweet spots to optimize well-bore placement and reduce drilling risks, Society of Exploration Geophysicists, 2012, pp. 1–5. [3] B. Goodway, J. Varsek, C. Abaco, Practical applications of p-wave avo for unconventional gas resource plays, Part 2 (2006) 52–65. 24
410
[4] R. K. Sharma, S. Chopra, New attribute for determination of lithology and brittleness, Society of Exploration Geophysicists Las Vegas, 2012, pp. 1–5. [5] M. H. Shereef, O. Al-Khaled, S. K. Singh, M. H. A. Razak, M. Al-Awadhi, et al., Seismic prestack inversion technique to assess the unconventional reservoir potential of the makhul formation: A case study in north kuwait,
415
in: 2014 SEG Annual Meeting, Society of Exploration Geophysicists, 2014. [6] S. Sun, K. Wang, P. Yang, X. Li, J. Sun, B. Liu, K. Jin, et al., Integrated prediction of shale oil reservoir using pre-stack algorithms for brittleness and fracture detection, in: International Petroleum Technology Conference, International Petroleum Technology Conference, 2013.
420
[7] R. Perez, K. Marfurt, et al., Brittleness estimation from seismic measurements in unconventional reservoirs: Application to the barnett shale, in: 2013 SEG Annual Meeting, Society of Exploration Geophysicists, 2013. [8] D. M. Jarvie, R. J. Hill, T. E. Ruble, R. M. Pollastro, Unconventional shale-gas systems: The mississippian barnett shale of north-central texas
425
as one model for thermogenic shale-gas assessment, AAPG bulletin 91 (4) (2007) 475–499. [9] F. P. Wang, J. F. Gale, Screening criteria for shale-gas systems. [10] X. Jin, S. N. Shah, J.-C. Roegiers, B. Zhang, et al., An integrated petrophysics and geomechanics approach for fracability evaluation in shale reser-
430
voirs, SPE Journal 20 (03) (2015) 518–526. [11] T. Guo, S. Zhang, Z. Qu, T. Zhou, Y. Xiao, J. Gao, Experimental study of hydraulic fracturing for shale by stimulated reservoir volume, Fuel 128 (2014) 373–380. [12] T. Wynn, S. Stewart, The role of spectral curvature mapping in character-
435
izing subsurface strain distributions, Geological Society, London, Special Publications 209 (1) (2003) 127–143.
25
[13] K. Su, A comprehensive methodology of evaluation of the fracability of a shale gas play, Unconventional Resources Technology Conference (URTEC), 2014. 440
[14] A. Roberts, Curvature attributes and their application to 3d interpreted horizons, First break 19 (2) (2001) 85–100. [15] A. Sena, G. Castillo, K. Chesser, S. Voisey, J. Estrada, J. Carcuz, E. Carmona, P. Hodgkins, Seismic reservoir characterization in resource shale plays: Stress analysis and sweet spot discrimination, The Leading Edge
445
30 (7) (2011) 758–764. [16] D. Metzner, Case study of 3d seismic inversion and rock property attribute evaluation of the haynesville shale, Unconventional Resources Technology Conference (URTEC), 2013. [17] L. Hunt, S. Chopra, S. Reynolds, S. Hadley, On calibrating curvature data
450
to fracture density: causes, CSEG Recorder 34 (2009) 27–32. [18] R. J. Lisle, Detection of zones of abnormal strains in structures using gaussian curvature analysis, AAPG bulletin 78 (12) (1994) 1811–1819. [19] S. Chopra, K. Marfurt, Seismic curvature attributes for mapping faults/fractures and other stratigraphic features, CSEG Recorder 32 (9)
455
(2007) 37–41. [20] P. Klein, L. Richard, H. James, 3d curvature attributes: a new approach for seismic interpretation, First break 26 (4). [21] R. J. Lisle, J. M. Robinson, The mohr circle for curvature and its application to fold description, Journal of Structural Geology 17 (5) (1995)
460
739–750. [22] H. Watkins, R. W. Butler, C. E. Bond, D. Healy, Influence of structural position on fracture networks in the torridon group, achnashellach fold and thrust belt, nw scotland, Journal of Structural Geology 74 (2015) 64–80. 26
[23] N. J. Price, J. W. Cosgrove, Analysis of geological structures, Cambridge 465
University Press, 1990. [24] M. D. Zoback, Reservoir geomechanics, Cambridge University Press, 2010. [25] J. C. Jaeger, N. G. Cook, R. Zimmerman, Fundamentals of rock mechanics, John Wiley & Sons, 2009. [26] R. Sack, Extension of griffith’s theory of rupture to three dimensions, Pro-
470
ceedings of the Physical Society 58 (6) (1946) 729. [27] E. M. Anderson, The dynamics of faulting and dyke formation with applications to Britain, Hafner Pub. Co., 1951. [28] S. R. Brown, C. H. Scholz, Closure of rock joints, Journal of Geophysical Research: Solid Earth 91 (B5) (1986) 4939–4948.
475
[29] L. Hunt, S. Reynolds, T. Brown, S. Hadley, J. Downton, S. Chopra, Am i really predicting natural fractures in the tight nordegg gas sandstone of west central alberta, Part II: Observations and Conclusions, GeoCanada. [30] J. Ericsson, H. McKean, R. Hooper, Facies and curvature controlled 3d fracture models in a cretaceous carbonate reservoir, arabian gulf, Geological
480
Society, London, Special Publications 147 (1) (1998) 299–312. [31] B. S. Hart, Validating seismic attribute studies: Beyond statistics, The Leading Edge 21 (10) (2002) 1016–1021. [32] D. Gao, Integrating 3d seismic curvature and curvature gradient attributes for fracture characterization: Methodologies and interpretational implica-
485
tions, Geophysics 78 (2) (2013) O21–O31. [33] R. Rickman, M. J. Mullen, J. E. Petre, W. V. Grieser, D. Kundert, et al., 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, Society of Petroleum Engineers, 2008.
27
490
[34] X. Jin, An integrated geomechanics and petrophysics study of hydraulic fracturing in naturally fractured reservoirs, Thesis (2014). [35] M. J. Mullen, M. B. Enderlin, et al., Fracability index-more than rock properties, in: SPE Annual Technical Conference and Exhibition, Society of Petroleum Engineers, 2012.
495
[36] J. Hesthammer, Improving seismic data for detailed structural interpretation, The Leading Edge 18 (2) (1999) 226–247. [37] R. Nelson, Geologic analysis of naturally fractured reservoirs, Gulf Professional Publishing, 2001.
28
PII: DOI: Reference:
S0920-4105(16)30252-2 http://dx.doi.org/10.1016/j.petrol.2016.06.034 PETROL3526
To appear in: Journal of Petroleum Science and Engineering Received date: 10 March 2016 Revised date: 13 June 2016 Accepted date: 21 June 2016 Cite this article as: Zhi Geng, Mian Chen, Yan Jin, Xin Fang and Ning Jing, Integrated Fracability Assessment Methodology for Unconventional Naturally Fractured Reservoirs: Bridging the Gap between Geophysics and Production, Journal of Petroleum Science and Engineering, http://dx.doi.org/10.1016/j.petrol.2016.06.034 This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting galley proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
Integrated Fracability Assessment Methodology for Unconventional Naturally Fractured Reservoirs: Bridging the Gap between Geophysics and Production Zhi Genga,b , Mian Chena , Yan Jina,∗, Xin Fanga , Ning Jingc a College
of Petroleum Engineering, China University of Petroleum, Beijing, China de G´ eologie, Ecole Normale Sup´ erieure, CNRS UMR 8538, Paris, France c CNPC International (Chad) Ltd.
b Laboratoire
Abstract Fracability evaluation in unconventional reservoirs is one of the most important initiatives for economic success. Most evaluation methods in geophysics and well logging are based on calculating rock mechanical properties, in which a frequently used term is brittleness. However, oil & gas production, mainly determined by effective natural fracture networks in unconventional low-permeability reservoirs, is the final goal of reservoir stimulation. In this work, a novel integrated methodology was proposed to correlate conventional post-stack seismic data to well production rate. It aims to pre-evaluate regional stimulation performance, facilitating investment decision-making in line with global oil prices and environmental consideration. Structural occurrence, seismic maximum curvature, and geomechanics parameters were combined in a physical model to evaluate underlying development and activation of natural fractures, result in the activation index F1 . The potential of artificial induced fractures were quantified as normalized index F2 by integrating mechanical item based on fracture mechanics and fluid flow. Accordingly, a fracability sweetness index (Fsweet ) model was built to correlate standardized well production rate (Qef f ). A Case study of a naturally fractured igneous reservoir with 8 wells showed agreement between Fsweet and Qef f . A considerable part of highly active fractures with ∗ Corresponding
author Email addresses: [email protected] (Zhi Geng), [email protected] (Yan Jin)
Preprint submitted to Journal of Petroleum Science and Engineering
June 21, 2016
F1 of [0.9, 1] distributed over the area with lower curvature value. By contrast, the majority of fractures with lower F1 of [0.8, 0.9] located in zone with higher curvature value. The weight ratio α = 0.06 indicated that in this case the overwhelming production contribution were from natural fractures, which were evidenced by well core picture. The field data demonstrates the prospect for predicting regional fracability variation, and suggests that it is possible to correlate fracability index with standardized well production rate. The insights contribute to well placement optimization and stimulation treatment design, breaking limitation of engineers0 influence on individual wells, and hence reducing exploitation cost while enhancing ultimate recovery. Keywords: fracability, seismic curvature, fracture density, fracture activation, unconventional reservoir 2015 MSC: 00-01, 99-00
1. Introduction Confronting the worldwide oil price shocks, cost-efficient development in hydrocarbon reservoirs plays a key role in economic success, especially in unconventional reservoirs with low or ultra-low permeability. Fracability evaluation 5
in unconventional naturally fractured reservoirs is one of the most important initiatives. Fracability is generally defined as the capability of reservoirs to be fracture stimulated effectively [1]. Most of the fracability evaluation methods developed from geophysics and laboratory experiments focus on calculating rock mechanical properties, among which a frequently associated term is brittleness.
10
One popular way to predict fracability/brittleness is based on geophysical seismic inversion of rock elastic parameters, such as Young0 s modules, Poisson0 s ratio, Lame0 s constants, etc. Corresponding prediction models are defined by combination of these elastic parameters. For example, Varga et al. (2012) calculated brittleness index using formula defined by Young0 s modules and Poisson0 s
15
ratio, which was estimated by using simultaneous AVO inversion theory [2]. Metzner et al. (2013) found that Mu-Rho (rock0 s shear rigidity and bulk den-
2
sity) attribute, documented by Goodway [3], show the best correlation with their estimated ultimate recovery, indicating good fracability. Similarly, Sharma et al. (2012) [4] and Shereef et al. (2014) [5] used the product of Young0 s modules and 20
bulk density as a brittleness indicator. In addition, Sun et al. (2013) integrated brittleness and fractures which were predicted by inverting limited azimuthal seismic data to estimate preferable area for drilling locations [6]. Rock lithology and mineral content estimated by seismic measurement [7] and well logging [8],[9],[10] are also used as brittleness indicators. Moreover, Guo et al. (2014)
25
evaluated shale reservoir fracability by various experimental tests [11]. Oil & gas production, mainly determined by activated natural fracture networks in unconventional reservoirs with low permeability, is the final goal of reservoir stimulation. However, little attention has been paid to productionoriented regional fracability characterization and prediction. The more under-
30
lying activated natural fractures, the more probably being interlinked by stimulation and consequently enhance ultimate recovery. In this paper we presented a novel integrated fracability assessment methodology applied in unconventional naturally fractured reservoirs, by correlating conventional post-stack seismic data to potential well production capacity us-
35
ing physical models. The insights contribute to well placement optimization and stimulation treatment design, breaking limitation of engineers0 influence on individual wells and hence reducing exploitation cost while enhancing ultimate recovery.
2. Methodology 40
With the similar stimulation treatment in unconventional reservoirs, natural production rate is dominated by two factors: development & activation of natural fractures, and quality of artificial induced fractures. In this paper, the term “artificial induced fractures” means the fractures/cracks to be hydraulic fractured by engineering operation. The curvature of geological structure sur-
45
faces can be used to assess the degree of strain, which is expressed as brittle
3
fracture relevant to natural fractures [12]. Intersection of natural fractures and induced fractures by hydraulic fracturing will lead to successful stimulation [13]. The more connections between fractures, the more promotion to reservoir performance. 50
Taking above into consideration, our fracability assessment is well production oriented. A normalized index of fracability sweetness was defined to evaluate regional fracability distribution, which was correlated to well production (oil & water) rate. As artificial stimulation treatment is inevitably limited by natural properties of rock under in-situ conditions, we gave a new horizon to characterize
55
regional fracability variation, breaking through the stimulation design efforts focused on individual well profiles in depth. Once a target formation is selected to be stimulated, then the regional variation of fracability sweetness can be used to deploy new drilling wells which are more favorable to be fractured. For drilled wells, it aims to pre-evaluate stimulation performance, avoiding investment and
60
environmental risks. The datasets used in our method include: 1. structural attributes (curvatures, azimuth and dip angle) extracted from post-stack seismic data in depth domain; 2. far field stress distribution interpreted from seismic data; 3. production data of drilled wells. For structural attributes and stress distribution
65
data, exhaustive description or interpretation can be found in Roberts (2001) [14], Sena et al. (2011) [15], and Metzner et al. (2013) [16]. It is noteworthy that the datasets above are the most commonly used in petroleum industry, no more specialized techniques or tools are needed. 2.1. Development & Activation of Natural Fractures
70
Structural curvature is a key indicator from which the current, local, stress/strain is illustrated [17]. In this study, seismic structural attributes were used to characterize geometry of geological surfaces. Seismic curvature attributes were introduced in the mid-1990s and shown to be highly correlated with fractures, some of which were measured on outcrops [14],[18],[19]. The detailed introduc-
75
tion and calculation methods of various curvatures were discussed in references 4
[14],[20],[21]. The most important are maximum and minimum curvature, which are also called principle curvatures and used to calculate other curvatures. Previous studies suggested that structural position takes over as the most important controlling factor on fracture formation [22]. The positive maximum curvature 80
indicates anticline structure and more natural fractures, while negative minimum curvature means syncline structure and less fractures or closed fractures. As maximum curvature attribute is highly sensitive to brittle deformation, it is reasonable to suppose that the higher value of maximum curvature, the stronger possibility that fractures develop.
85
Assuming in an ideal situation, underground rock is liner elastic, and brittle failure hasn0 t occurred in suffering tectonic deformation. As illustrated in Figure 1, through one point on geological surface, there is exactly a plane normal to the dip vector which has the same azimuth as maximum curvature . The plane is treated as a fictitious fracture model. Then development & activation
90
of potential natural fractures in the vicinity of the point can be estimated by calculating failure tendencies of the fracture model under the action of far field stress and structural stress indicated by structural curvature. It is a reasonable assumption to analyze mechanics characteristic of the fracture model during elastic deformation until the moment of brittle failure. If the fracture model
95
fails under loading, then relatively active fractures exist with extremely high probability. On the contrary, even though there might be some fractures result from geological processes or environmental influence, they are probably inactive or closed, which is not conducive to fluid flow. The induced stress experienced at the top of the structure can be estimated
100
from [23]: σinduce =
hKE 2
(1)
where σinduce is structural induced stress, h is the thickness of layer, K represents the maximum curvature in three-dimensions [14], E is the Young0 s modulus of rock. Su et al. (2014) assessed the likelihood of natrual fractures activation along
5
Z (Depth)
K min
K max lin
e)
N
Y
(In
φ
X (Crossline) Figure 1: Schematic diagram of fracture model (rectangle with tiny aperture). Kmax is maximum curvature of geological surface (gray color), ϕ is azimuth of Kmax . Kmin is minimum curvature which is orthogonal to Kmax . The strike of fracture model (ϕf ) aligns with the azimuth of Kmin , so is orthogonal to ϕ. The dip angle of fracture model is where DipKmax is the dip angle of surface in the direction of Kmax .
6
π 2
− DipKmax ,
105
wellbores based on effective stress on a fracture by simple projection of the far field stress on its plane according to the normal vector of the plane in principle stress coordinate [13]. For the convenience of simple calculation using geological terms such as azimuth, dip of fractures and far field stress, we used tensor transformation method described by Zoback et al. (2010) [24] to quantify the
110
normal stress Sn and shear stress Sτ on arbitrary fracture planes: [Sn , Sτ ] = f (SH , ϕSH , ϕf , Dipf )
(2)
where SH , ϕSH is the value and azimuth of maximum horizontal stress separately; ϕf , Dipf is the strike/azimuth and dip angle of fracture model respectively. The activation of the fracture model is then estimated by combining Equa115
tion 1, Equation 2, Mohr-Coulomb and tension failure criteria, given by: Sτ , S ≤ τ τ m τm F1 = 1, S > τ or S ≤ σ +P τ
m
n
induce
(3)
p
where τm = µ (Sn − σinduce − Pp ) is shear failure stress of the fracture model, µ is friction coefficient which is normally 0.6 in many cases [24],[25], Pp is pore pressure or fluid pressure. The activation index F1 ∈ [0, 1]. F1 = 0 indicates there are no active fractures, while F1 = 1 shows effective fractures abound. 120
Although the activation index here is not a solid evidence of objective existence of specific activated fracture, it implies the regional variation and distribution of relative activity of dormant natural fractures. 2.2. Potential of Artificial Induced fractures Besides the engineering treatment influence on the shape and quality of
125
artificial fractures, rock mechanical properties are considered as the basic determinant factors contributing to development of induced fractures as well as the fluid flow. Flow rate through a planar fracture integrated with aperture was given by Equation 4 [24], and the critical stress value for cracks extension based
7
130
on Griffith0 s theory of rupture was defined in Equation 5 [26]: " #3 π L 1 − ν 2 (pf − S3 ) Q= ∇p 8η E s πEγ Pc = 2c (1 − ν 2 )
(4)
(5)
where Q is volumetric flow rate in the fracture, η is fluid viscosity, L and c are fracture length, item pf − S3 is net pressure in the fracture, ∇p is pressure gradient, γ is fracture surface energy, ν is Poisson0 s ratio and E is Young0 s
135
modulus. By comparison of above two equations, the shared rock mechanics item 1 − ν 2 /E is positively correlated with Q, and negatively related with Pc . It suggests that the higher value of the item, the better for extension of induced fractures (lower Pc ) as well as the fluid flow result from large fracture aperture (higher Q). Stimulation treatment and geomechanics difference is inevitable whether for
140
individual wells or in regional area. As a consequence, it seems plausible to evaluate the variation of potential for induced fractures based on objective and basic rock mechanical properties. Accordingly, a normalized index of potential for induced fractures is defined as: fi −fi min fi max −fi min 2 i where fi = 1−ν Ei
F2 =
(6)
where fi is the value of item 1 − ν 2 /E for specific point in field area; fi max , 145
fi min is the maximum and minimum value for all the points in the area respectively; F2 ∈ [0, 1] is the corresponding normalized index. The higher value of F2 , the better situation for creation of induced fractures with large aperture and fluid flow. The normalization of index is applicable to characterize the relative variation of rock mechanical properties in different location, so that it enlightens
150
the quality of artificial induced fractures. 2.3. Modeling Fracability Sweetness In unconventional naturally fractured reservoirs, well production rate is generally contributed from natural fracture networks and the extent of artificial 8
fractures. In order to eliminate the impacts of formation pressure and comple155
tion conditions on production rate, the effective individual well production rate is standardized using: Qef f =
Qt HT ∆P
(7)
where Qt is cumulative production, H is the thickness of stimulated reservoir, T is production days, ∆P is production pressure difference. Assuming that normalized individual well production rate is liner combination of contribution 160
from natural fractures and induced fractures, then a correlation with two weight coefficients can be established. Afterwards, the correlation can be defined using index F1 , F2 : Qef f i = c1 × F1 i + c2 × F2 i
(8)
where Qef f i is standardized production rate of the ith well, F1 i is activation index in the vicinity of the ith well, F2 i is the corresponding index for induced 165
fractures, c1 and c2 are weight coefficients respectively which can be calculated by using least square method. With ratio of α = c2 /c1 , the model of fracability sweetness is given: Fsweet = F1 + αF2
(9)
where Fsweet ∈ [0, 2] is index of fracability sweetness, which is the higher the better. The ratio α is convenient to show the relative contribution of effective 170
natural fractures and induced fractures. The index model is then capable of evaluating regional variation of fracability sweetness in a unified pattern.
3. Model Verification with Case Study By evaluating the activation of fracture model and potential of artificial induced fractures in the use of seismic structural attributes and interpretation, 175
the correlation between fracability sweetness index and standardized production rate was established. In order to validate our fracability evaluation approach, a case study of naturally fractured igneous reservoir in western China was shown in the sections bellow.
9
The target zone is Permian system volcanic rock with low formation dip angle 180
(6∼9◦ ), covering an area of 100 km2 . The CMP bin of 3D seismic data is 25×25 m. Wave group characteristics are clear, and the seismic data is in good quality. The effective bandwidth near target zone is 10∼43 Hz, with dominant frequency of 18 Hz. The oil-gas-bearing volcanic rock is mainly composed of basalt and tuff. Mineral components are plagioclase (60∼72%), quartz (9∼15%), calcite
185
(7∼13%), pyroxene (3∼9%), chlorite (4∼9%), and little magnetite (0∼4%). The average thickness of the reservoir is 16 m, with burial depth of 3225∼3395 m. The maximum horizontal stress gradient SH = 2.7 with azimuth of 105◦ , and the minimum horizontal stress gradient Sh = 2.3. The stress state of the area is strike-slip or reverse faulting according to E. M. Anderson0 s (1951) classification
190
scheme [27]. There are 7 vertical wells which were hydraulic fractured with similar treatment, and 1 horizontal well without being stimulated in the field. The high conductive natural fractures, which were observed on some imaging logs, have high dip angle. However, one obvious limitation of imaging tools is that they will especially undersample fractures and faults whose planes are
195
nearly parallel to the wellbore axis [24]. So, in this case the imaging logs could hardly reflect the development and density of fractures nearby wells in the block. As the permeability of the rock matrix is extremely low, the fluid flow in the reservoir can only be achieved through fracture networks, which was confirmed in Figure 2. The evidence points to the fact that in this case, natural fractures
200
contribute dominantly to the production. The distribution of original seismic maximum curvature was shown in Figure 3. Since the inset histogram indicated that the scope of curvature data was mainly 0∼0.2, the host graph was color coded in [0, 0.2]. High value of maximum curvature in red color qualitatively implied dense natural fractures,
205
which concentrated on band zone around wells N32∼N10. While the maximum curvature value around well N11, M3, P8 and N32 were obviously low, which suggested scarce fractures. So was the same situation in the neighboring area around the horizontal section of H1. As described in Section 2.1, the underlying development and activation of 10
Figure 2: Reservoir core from burial depth of 3285 m. Macroscopic oil traces coincide exactly with natural fractures.
11
2000
0.2
1950
0.16
N11 1900
0.12
Inline
M3 P8
1850
N32
P6
H1
0.08
1800
N31 0.04
1750
N10 1700
1550
1500
1450
Crossline
1400
0
Figure 3: The distribution of original seismic maximum curvature. The inset histogram showed data interval of [0, 0.4]. Since the curvature data mainly concentrated in 0∼0.2, the host graph was color coded in [0, 0.2]. Green circles symbolized wells location, the radius of circles was 150 m, which was the estimation of stimulation extent. The lateral length of H1 was 400 m.
12
210
natural fractures was depicted in Figure 4. The azimuth and activation of fracture models were visualized in Figure 4a, in which the back ground (base map) was gray-colored as the same value of maximum curvature in Figure 3. The solid lines in red or blue were color coded by F1 (activation index) in interval of [0.8, 1] for visual friendly and density display of underlying highly
215
active fractures. It should be noted that a considerable part of the relative highly active fractures (red color) as interval [0.9, 1] distributed over the area with lower curvature value (dark gray color). By contrast, the majority of fractures with lower F1 (blue color) as interval [0.8, 0.9] located in zone with higher curvature value (bright gray color). Even though the overall curvature
220
value on upper part of Figure 3 was relatively low, the highly active fractures were common as seen in Figure 4a. It appears that maximum curvature in Figure 3 implied the existence of natural fractures in some degree. However, it0 s obviously insufficient when compared with the method proposed in this paper. The Figure 4b is a different
225
manifestation of F1 distribution in the whole data interval of [0, 1]. For the comparison with following figures, Figure 4b was color coded in [0.4, 1], which means the color of data in [0, 0.4] is the same. Figure 4b shows much more clearly preference to Figure 3 when characterizing development and activation of natural fractures, especially highly active fractures. In addition, Figure 4
230
suggests that both activation and density of fractures around well N11, M3 and P8 was higher than that in Figure 3 (maximum curvature). What0 s more, the corresponding value near well N32 and horizontal section of H1 were still in depress. Figure 5 displays further evidence of our above results. Although it should
235
be cautious to use image logs to reveal development of fractures/faults near wellbore, they provide macro-understanding of relative fracture density. As shown in the figure, the fractures/cracks in image P6 and M3 are obviously much more than that in image N32. This coincides with F1 distribution in Figure 4. However, according to the maximum curvature in Figure 3, the fracture density
240
of well M3 and N32 should be in the same level, which was much lower than 13
2000
0.2
1950
N11
0.16
1900 0.12
Inline
M3 P8
1850
N32
P6 1800
H1
0.08
N31 0.04
1750
N10 1700
1550
1500
1450
Crossline
1400
0
(a) The azimuth and activation of fracture models were shown as solid lines in red or blue, which were color coded by activation index (F1 ) in interval of [0.8, 1], with grid increment of 2 for visual friendly. The back ground (base map) was gray-colored by the same value of maximum curvature in Figure 3, which shared the same inset histogram.
14
2000
1.0
1950
0.9
N11
1900
0.8
Inline
M3 P8
1850
N32
P6 1800
0.7
H1 0.6
N31
0.5
1750
N10 1700
1550
1500
1450
Crossline
1400
0.4
(b) The distribution of F1 in the whole interval of [0, 1], which was displayed in inset histogram. The higher value the better. For the comparison with following figures, F1 was color coded in [0.4, 1], which means data interval of [0, 0.4] shared the same color. Figure 4: Underlying development and activation distribution of natural fractures.
15
that of well P6.
P6
M3
N32
3326
Measured Depth (m)
3297.5 3239
3327
3298.5 3240
0
180
360
0
180
360
0
180
360
Azimuth (°) Figure 5: Reservoir image logs of 3 wells (P6, M3 and N32).
The vertical black strips
(azimuth≈185∼200◦ , azimuth≈10∼20◦ ) in each image indicate well collapse induced by maximum horizontal stress. The fractures/cracks in image P6 and M3 are obviously much more than that in image N32.
The normalized index of potential for induced fractures (F2 ) was shown in Figure 6 The area with warm color was in favor of creating artificial induced fractures and fluid flow from perspective of rock mechanical properties. Owing 245
to the increasing depth of reservoir from east to west, the rock elastic modulus was higher in western region. As a consequence, the figure implies that the western region with lower F2 was not ideal for initiation of artificial fractures with beneficial aperture. Note once again that up to now all attributes around well N32 and horizontal section of H1 hold low level.
250
The final result of fracability sweetness index (Fsweet ) was presented in Figure 7. Since well H1 wasn0 t stimulated, the production data of 7 vertical wells
16
1
2000
1950
0.8
N11
1900 0.6
Inline
M3 P8
1850
N32
P6 1800
H1
0.4
N31 0.2
1750
N10 1700
1550
1500
1450
Crossline
1400
0
Figure 6: Distribution of normalized index of potential for induced fractures (F2 ). The inset histogram showed that F2 distributed in [0, 1]. The warmer color it was, the better situation for induced fractures.
17
in combination with related attributes were used to build fracability sweetness model. The attributes value around wells were root-mean-square of all the grid points in the circles showed in figures. The ratio α = 0.06 indicated the over255
whelming production contribution of natural fractures, which were evidenced by Figure 2. As real fractures in rock with ultra low matrix permeability will not have perfectly smooth surfaces [28], the fractures with finite aperture still can enhance flow to some extent [24]. Hence, the index F2 in this case was further from valuable than it would be in cases that high conductive fractures with
260
large aperture come first. It appeared that overall visual effect of Figure 7 and Figure 4b was similar. One primary cause was the tiny value of α. However, the inset histograms in each figure clearly showed that the data in Figure 7 were more uniformly distributed, which suggested rich recognition of fracability sweetness.
265
The method proposed was verified by comparing data including standardized production rate (Qef f ), maximum curvature (Kmax ), and fracability sweetness index (Fsweet ), as shown in Figure 8. As mentioned above, all attributes around well N32 held low level, so was its fracability sweetness. Since well H1 wasn0 t simulated, it wasn0 t used to build fracability model and its standardized pro-
270
duction rate was extremely low. It has been demonstrated that the maximum curvature alone is easily affected by noise and unstable. In our results, for example, the curvature value of well P6 was abnormally high, while the value of well P8 was excessively low. The whole line of Kmax hardly matched the trend of Qef f . By contrast, the fracability sweetness curve which was obtained by
275
combining F1 and F2 coincided best with the trend of Qef f , especially on well P8 and N10 when compared with the curve of F1 , let alone F2 . In this paper, c1 Fsweet indicated standardized production rate of wells. It could be inferred that after well H1 being stimulated, its productivity must be higher than well N320 s, especially considering the long horizontal section.
280
However, this work represents only a preliminary attempt to establish such a correlation. The actual relationship between natural & artificial induced fractures and well production rate may be more complex. Nevertheless, these results 18
2000
1.0
1950
0.9
N11
1900
0.8
Inline
M3 P8
1850
N32
P6 1800
0.7
H1 0.6
N31
1750
0.5
N10 0
1700
1550
1500
1450
Crossline
1400
0.4
Figure 7: Distribution of fracability sweetness index (Fsweet ). The inset histogram showed that Fsweet located in [0.1, 1.1]. The host graph was color coded in [0.4, 1] for comparison with Figure 4b. The red color ([0.7, 1]) represented superior fracability, while blue color suggested inferior ones.
19
Standardized Production Rate (t/day/m/kpa)
F1 F2
12
Fsweet Kmax Production
0.3
0.9
10 0.2
8
0.9 0.7 0.8
6
0.1 0.5
4 2 0
1.0
0.7
0.0 0.3 0.6 P6
P8 N10 N31 M3 N11 N32 H1
well name
Figure 8: Comparison of standardized well production rate (Qef f , column bars) with maximum curvature (Kmax , red line with stars) and fracability sweetness index (Fsweet , green line with circles).
suggest that the fracability sweetness index (Fsweet ) obtained based on seismic and production data may provide more meaningful information for assessing the 285
profitable area to be exploited than traditional methods.
4. Discussion Our results provide compelling evidence for rationality and applicability of the proposed methodology to characterize regional fracability sweetness in unconventional naturally fractured reservoirs. It suggests that this approach ap290
pears to be effective in assessment of natural fractures activation and potential for induced fractures with aperture. As it0 s so difficult to strictly characterize complex reservoirs only using physical equations, it seems prudent to normalize the key physical parameters so as to ignore uncertain factors when uncover the regional relative variation. Our evaluation contributes to decision making for
295
stimulation in line with global oil prices as well as environmental consideration. As discussed in case study, the potential productivity of well H1 with stimu20
lation can be inferred, at least providing a lower limit. Accordingly, rational investment can be done depending on the current market. 4.1. The reasons for using seismic curvature 300
The prediction or inference of natural fractures in field using seismic data are generally accepted. The seismic data used in this purpose can be divided into two kinds: the pre-stack data and post-stack data. In theory, the methods using azimuth offset data (pre-stack) are detection techniques for fractures occurrence and density. Although they are widely limited by rigorous theoretical
305
preconditions for fracture angle and alignment, the most concern here is their extremely high cost for acquisition, processing and interpretation, let alone most seismic pre-stack data are limited in azimuth in China [6]. By contrast, the seismic curvature calculated by using relative low-cost post-stack data is a useful attribute set for making predictions or inference regarding fracture density [29].
310
Ericsson et al. (1988) showed the relationship between reservoir production and seismic curvature [30]. Lisle (1994) discussed the correlation of Gaussian curvature to fractures measured on an outcrop [18]. Roberts (2001) documented in details regarding the relationship between fractures and surface shape determined by various curvatures, such as maximum curvatures, Gaussian curvatures,
315
total curvatures and mean curvatures etc.[14]. Hart et al. (2002) predicted that large value of strike curvatures were correlated to open fractures [31]. Gao (2013) integrated 3D curvatures for fracture characterization [32]. Most of previous research have used curvatures as indicators to predict natural fracture density.
320
4.2. The features in the methodology As mentioned above, brittleness is generally used as an indicator for fracability. Increasing brittleness means there may be natural fractures present and good for hydraulic stimulation [33]. However, this assumption is not always true because formation with higher brittleness can also be a fracture barrier
325
[34]. The reason is that conventional concept of rock brittleness qualitatively
21
assumes it increases with higher Young0 s modulus and lower Poisson0 s ratio [33]. In the meanwhile, however, the strength and fracture toughness of rock keeps growing with increasing Young0 s modulus, which is unfavorable for initiation and extension of induced fractures. 330
In this study, the principles for extension and fluid flow in aperture of fractures were analyzed basing on fracture mechanics and flow model, which enlightens the nature of rock mechanical properties that in favor of induced fractures. According to the item 1 − ν 2 /E, the ideal situation for induced fractures and fluid flow is lower value for both Young0 s modulus and Poisson0 s ratio. If the
335
Poisson0 s ratio keeps constant, the rock with lower Young0 s modulus is easier to break; and if the Young0 s modulus remains unchanged, the fractures in rock with lower Poisson0 s ratio would retard healing to closure. Besides the view of rock itself, excessive Young0 s modulus and stress in rock also pose new challenges to the selection of costly proppant with better quality, which is indispensable to
340
keep fractures open after reservoirs being hydraulic fractured. Furthermore, fracability evaluation is more than just calculating rock properties [35]. The development and activation of effective natural fractures play key role in exploitation of unconventional reservoirs. Although the natural fracture density can be qualitatively inferred by combing multiple seismic curvatures
345
[14], completion and production behavior doesn0 t necessarily depend heavily on fracture density. The reason behind is that some of the fractures with certain occurrence (strike and dip angle) are probably closed (dormant) in present-day stress field. In other words, these fractures are non-effective. One thing to keep in mind is that activation degree of fractures has a bearing on the flow capacity
350
of the reservoir [17]. Therefore, by analyzing mechanical behavior of geological fracture models, this paper lays emphasis on activation evaluation of fractures under present-day far field stress and structural stress, which is indicated by structural curvature.
22
4.3. The limitations 355
The limitations are inevitable in the fracability evaluation method proposed. Seismic curvatures are not immune to noise, whose influence is sample offset (grid spacing) related [12]. The seismic data can also contain false curvature features e.g. velocity pull-ups and push-downs and apparent faulting created by noise contamination [36]. In addition, although the relations between curvature
360
and fracture density are well supported by related geological studies, there are three assumed conditions suggested by Nelson [37]. They are: 1. the rock is brittle and fails largely from fracturing, 2. an increase in curvature implies an increase in strain, and 3. an increase in strain implies an increase in fracture density. It implies that our method is applicable to evaluate structural fractures
365
except that induced by factors such as evolutionary history, environmental conditions, etc. Moreover, it still matters for seismic interpretation accuracy of geomechanics data, such as rock elastic parameters, in-situ stress, etc. Nevertheless, since being calculated using post-stack seismic data, curvature is not only computationally cheaper but also more robust than many fracture
370
density indicators [17]. Besides, by normalization processing of key parameters for regional variation assessment, the negative impact of errors between true value and estimated value on specific point in area can be weakened in some degree.
5. Conclusion 375
In this paper, an integrated fracability assessment methodology was presented by evaluating activation of natural fractures and potential of artificial induced fractures with aperture. The fracability sweetness index was calculated by using available conventional post-stack seismic data and production logs, revealing the correlation between exploration data and well production rate. The
380
insights contribute to drilling locations optimization and stimulation treatment design, breaking limitation of engineers0 influence on individual wells. The case study showed feasibility of the novel method in unconventional naturally frac-
23
tured reservoirs. The productivity of a pre-stimulated well was inferred by using the fracability model, which benefits decision making for stimulation in line with 385
global oil prices and environmental consideration. Our results are encouraging and should be validated in more cases. However, the production mechanisms underlying natural & induced fracture networks remain to be further studied. Future work should be focused on two aspects: 1. time dependent effectiveness of active fractures for fluid flow, 2. influence
390
factors and mechanisms of fluid flow in naturally fractured reservoirs. On the whole, the evaluation of stimulation potential should be promoted throughout the complete reservoir life cycle.
Acknowledgements The authors gratefully acknowledge the financial supports (Major Program 395
No.51490651) from National Natural Science Foundation of China. We also thank the anonymous reviewers for their valuable comments and suggestions on our article.
References [1] K. K. Chong, W. V. Grieser, A. Passman, H. C. Tamayo, N. Modeland, 400
B. E. Burke, et al., A completions guide book to shale-play development: A review of successful approaches toward shale-play stimulation in the last two decades, in: Canadian Unconventional Resources and International Petroleum Conference, Society of Petroleum Engineers, 2010. [2] R. Varga, R. Lotti, A. Pachos, T. Holden, I. Marini, E. S. J. Pendrel, Seis-
405
mic inversion in the Barnett Shale successfully pinpoints sweet spots to optimize well-bore placement and reduce drilling risks, Society of Exploration Geophysicists, 2012, pp. 1–5. [3] B. Goodway, J. Varsek, C. Abaco, Practical applications of p-wave avo for unconventional gas resource plays, Part 2 (2006) 52–65. 24
410
[4] R. K. Sharma, S. Chopra, New attribute for determination of lithology and brittleness, Society of Exploration Geophysicists Las Vegas, 2012, pp. 1–5. [5] M. H. Shereef, O. Al-Khaled, S. K. Singh, M. H. A. Razak, M. Al-Awadhi, et al., Seismic prestack inversion technique to assess the unconventional reservoir potential of the makhul formation: A case study in north kuwait,
415
in: 2014 SEG Annual Meeting, Society of Exploration Geophysicists, 2014. [6] S. Sun, K. Wang, P. Yang, X. Li, J. Sun, B. Liu, K. Jin, et al., Integrated prediction of shale oil reservoir using pre-stack algorithms for brittleness and fracture detection, in: International Petroleum Technology Conference, International Petroleum Technology Conference, 2013.
420
[7] R. Perez, K. Marfurt, et al., Brittleness estimation from seismic measurements in unconventional reservoirs: Application to the barnett shale, in: 2013 SEG Annual Meeting, Society of Exploration Geophysicists, 2013. [8] D. M. Jarvie, R. J. Hill, T. E. Ruble, R. M. Pollastro, Unconventional shale-gas systems: The mississippian barnett shale of north-central texas
425
as one model for thermogenic shale-gas assessment, AAPG bulletin 91 (4) (2007) 475–499. [9] F. P. Wang, J. F. Gale, Screening criteria for shale-gas systems. [10] X. Jin, S. N. Shah, J.-C. Roegiers, B. Zhang, et al., An integrated petrophysics and geomechanics approach for fracability evaluation in shale reser-
430
voirs, SPE Journal 20 (03) (2015) 518–526. [11] T. Guo, S. Zhang, Z. Qu, T. Zhou, Y. Xiao, J. Gao, Experimental study of hydraulic fracturing for shale by stimulated reservoir volume, Fuel 128 (2014) 373–380. [12] T. Wynn, S. Stewart, The role of spectral curvature mapping in character-
435
izing subsurface strain distributions, Geological Society, London, Special Publications 209 (1) (2003) 127–143.
25
[13] K. Su, A comprehensive methodology of evaluation of the fracability of a shale gas play, Unconventional Resources Technology Conference (URTEC), 2014. 440
[14] A. Roberts, Curvature attributes and their application to 3d interpreted horizons, First break 19 (2) (2001) 85–100. [15] A. Sena, G. Castillo, K. Chesser, S. Voisey, J. Estrada, J. Carcuz, E. Carmona, P. Hodgkins, Seismic reservoir characterization in resource shale plays: Stress analysis and sweet spot discrimination, The Leading Edge
445
30 (7) (2011) 758–764. [16] D. Metzner, Case study of 3d seismic inversion and rock property attribute evaluation of the haynesville shale, Unconventional Resources Technology Conference (URTEC), 2013. [17] L. Hunt, S. Chopra, S. Reynolds, S. Hadley, On calibrating curvature data
450
to fracture density: causes, CSEG Recorder 34 (2009) 27–32. [18] R. J. Lisle, Detection of zones of abnormal strains in structures using gaussian curvature analysis, AAPG bulletin 78 (12) (1994) 1811–1819. [19] S. Chopra, K. Marfurt, Seismic curvature attributes for mapping faults/fractures and other stratigraphic features, CSEG Recorder 32 (9)
455
(2007) 37–41. [20] P. Klein, L. Richard, H. James, 3d curvature attributes: a new approach for seismic interpretation, First break 26 (4). [21] R. J. Lisle, J. M. Robinson, The mohr circle for curvature and its application to fold description, Journal of Structural Geology 17 (5) (1995)
460
739–750. [22] H. Watkins, R. W. Butler, C. E. Bond, D. Healy, Influence of structural position on fracture networks in the torridon group, achnashellach fold and thrust belt, nw scotland, Journal of Structural Geology 74 (2015) 64–80. 26
[23] N. J. Price, J. W. Cosgrove, Analysis of geological structures, Cambridge 465
University Press, 1990. [24] M. D. Zoback, Reservoir geomechanics, Cambridge University Press, 2010. [25] J. C. Jaeger, N. G. Cook, R. Zimmerman, Fundamentals of rock mechanics, John Wiley & Sons, 2009. [26] R. Sack, Extension of griffith’s theory of rupture to three dimensions, Pro-
470
ceedings of the Physical Society 58 (6) (1946) 729. [27] E. M. Anderson, The dynamics of faulting and dyke formation with applications to Britain, Hafner Pub. Co., 1951. [28] S. R. Brown, C. H. Scholz, Closure of rock joints, Journal of Geophysical Research: Solid Earth 91 (B5) (1986) 4939–4948.
475
[29] L. Hunt, S. Reynolds, T. Brown, S. Hadley, J. Downton, S. Chopra, Am i really predicting natural fractures in the tight nordegg gas sandstone of west central alberta, Part II: Observations and Conclusions, GeoCanada. [30] J. Ericsson, H. McKean, R. Hooper, Facies and curvature controlled 3d fracture models in a cretaceous carbonate reservoir, arabian gulf, Geological
480
Society, London, Special Publications 147 (1) (1998) 299–312. [31] B. S. Hart, Validating seismic attribute studies: Beyond statistics, The Leading Edge 21 (10) (2002) 1016–1021. [32] D. Gao, Integrating 3d seismic curvature and curvature gradient attributes for fracture characterization: Methodologies and interpretational implica-
485
tions, Geophysics 78 (2) (2013) O21–O31. [33] R. Rickman, M. J. Mullen, J. E. Petre, W. V. Grieser, D. Kundert, et al., 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, Society of Petroleum Engineers, 2008.
27
490
[34] X. Jin, An integrated geomechanics and petrophysics study of hydraulic fracturing in naturally fractured reservoirs, Thesis (2014). [35] M. J. Mullen, M. B. Enderlin, et al., Fracability index-more than rock properties, in: SPE Annual Technical Conference and Exhibition, Society of Petroleum Engineers, 2012.
495
[36] J. Hesthammer, Improving seismic data for detailed structural interpretation, The Leading Edge 18 (2) (1999) 226–247. [37] R. Nelson, Geologic analysis of naturally fractured reservoirs, Gulf Professional Publishing, 2001.
28