Keywords Shale gas play Elastic wave velocity Stress Micro-cracks are ubiquitous and present in shale gas sensitivity Micro-cracks Experimental study plays and are the important channels for gas flow and storage. Although these cracks generally make little con- tribution to the porosity (less than 0.1 %), they can rela- 1 Introduction tively influence the rock elastic and transport properties (Walsh 1965; Brace et al. 1968). Thus, it is urgent and Due to its ultralow permeability and porosity (Vernik and significant to develop a useful and easy method to under- Liu 1997; Kwon et al. 2001), hydraulic fracturing is con- stand the micro-cracks in shale gas plays for engineering ducted on shale gas plays to improve their economic pro- applications. Elastic wave velocities are more sensitive to duction. Hydraulic fracturing in shale plays requires the the propagation of micro-cracks than other parameters. formation of a dispersion-type volumetric fracture network Using the velocity evolution with stress in the linear elastic by SRV (stimulated reservoir volume) fracturing, which is phase of shale samples, the stress sensitivity coefficient can quite different from that in conventional reservoirs (Cipolla be developed to quantitatively evaluate the natural cracks. et al. 2008). However, not all of the shale gas reservoirs In the laboratory, the development of micro-cracks of two can achieve a wide range of SRV fracturing by massive types of shale gas play was evaluated (Lujiaping and hydraulic fracturing treatment. Fracability evaluation in Longmaxi group) by the stress sensitivity coefficient. For shale gas plays is important for optimizing the preferred the two shale outcrops, twenty cylindrical shale samples fracturing interval and for forecasting economic benefits. were cored along the bedding-parallel and bedding-normal Studies have shown that performing SRV fracturing on directions. The velocities were obtained along the axial reservoirs depends on the ductile–brittle properties of rock, stress direction under uniaxial stress. For each sample, the the development of natural fractures and anisotropy (Li stress was loaded to 30 MPa, and the velocities were et al. 2013; Jahandideh and Jafarpour 2014). Currently, measured at every MPa stress. fracability evaluation in shale gas plays is almost equiva- lent to brittleness evaluation (Sondergeld et al. 2010; Rickman et al. 2008). However, brittleness is not sufficient 2 Experimental Setup to evaluate fracability because the development of natural fractures or cracks is a dominant factor in fracability. 2.1 Experimental Sample
The shale samples examined in this study come from the
Chongqing area, southwest of China. There are two types & Xiaoqiong Wang of shale outcrop. One is Longmaxi shale formation in the [email protected] lower Silurian developed in the southeast of Chongqing. 1 The other is Lujiaping shale formation in the lower Cam- Unconventional Natural Gas Institute, China University of Petroleum, Beijing, 18 Fuxue Rd., Changping, brian developed in the northeast of Chongqing. The Beijing 102249, China detailed regional geological setting of these two black shale
123 X. Wang et al.
formations could be found in previous published papers noise ratio. Arrival times were obtained by cross-correla- (Zeng et al. 2013; Ma et al. 2015). The two areas of tion, so the relative error of the velocity measurement was Chongqing feature strong tectonic compression. Five estimated to be lower than 1 %. cylindrical specimens each (diameter 25 mm and length 50 mm) were cored from the two shale outcrops along the bedding-parallel and bedding-normal directions. Thus, 20 3 Experimental Results samples were investigated in this study. The samples are denoted as Lujiaping-parallel, Lujiaping-normal, Long- 3.1 Stress Sensitivity Coefficient and Anisotropy maxi-parallel and Longmaxi-normal for convenience. The surfaces of all samples were rectified and polished to As the applied stress increased to 30 MPa, the cracks dis- 10 lm to ensure parallelism and minimum friction during tributed inside the sample gradually closed, thus inducing testing. an increase of the compressional velocities. Figure 2 is the compressional velocity evolution with stress plot for both 2.2 Experimental Method and Testing System Lujiaping and Longmaxi shale samples of different direc- tion. The fitting line was obtained using the least squares Experiments were performed under the uniaxial stress method. With stress increasing, Vp increases almost linearly condition with a digital hydraulic loading system. The due to the closing of natural cracks. The slope of the fitting sample assembly is shown in Fig. 1a. For each sample, the line represents the stress sensitivity coefficients of the stress was loaded to 30 MPa. The compressional wave velocity. The stress sensitivity coefficient is defined as v. velocity (Vp) measurements along the axial stress direction From Fig. 2, the goodness of fit for all the samples is higher were conducted every 2 MPa with a classical ultrasonic than 0.9. The fitting lines for the bedding-normal samples pulse transmission technique. Here, the stress of 30 MPa (Fig. 2a, c) are better than that for the bedding-parallel was choosed from the previous uniaxial compression test as samples (Fig. 2b, d). All stress sensitivity coefficients v of below 30 MPa the shale is linear elastic. A pair of Nano30 the 20 samples are obtained and shown in Fig. 3a. The transducers was used. An electric pulse was generated with average stress sensitivity coefficients of the Lujiaping-par- an Olympus 5077PR pulse generator (Fig. 1b). The allel, Lujiaping-normal, Longmaxi-parallel and Longmaxi- waveforms were recorded using a digital oscilloscope at normal shale samples are 2.18, 3.18, 3.26 and 12.93 m/s/ 10 MHz sampling (Fig. 1c). For each velocity measure- MPa, respectively. The v of the Longmaxi-normal samples ment, 100 waveforms were stacked to improve the signal/ are much larger than those of the other samples.
Fig. 1 a Sample assembly
under uniaxial stress. b Block diagram of the ultrasonic wave system. c Measured waveform
123 Evaluation of the Micro-cracks in Shale from the Stress Sensitivity of Ultrasonic Velocities
Fig. 2 Vp versus stress plot for both Lujiaping and Longmaxi shale samples along different directions. a Lujiaping-normal, b Lujiaping-parallel, c Longmaxi-normal, d Longmaxi-parallel
Fig. 3 a Plot of the stress sensitivity coefficients of all the tested shale samples, b plot of the crack density of all the tested shale samples
For the same outcrop, the v of different directions j v1 v2 j
(bedding-normal or bedding-parallel) are obtained. We vanisotrophy ¼ ð1Þ averageðv1 ; v2 Þ roughly estimated the anisotropy of the stress sensitivity with the following equations where v1 and v2 are the stress sensitivity coefficients along the bedding-normal and bedding-parallel directions,
123 X. Wang et al.
respectively. According to Eq. (1), the Lujiaping shale out- With Eq. (2), the average anisotropy (both e and c) of crop has a stress sensitivity anisotropy of approximately Longmaxi shale is approximately 20 %, much more ani- 37 %, whereas that of the Longmaxi shale outcrop is 119 %. sotropic than that of Lujiaping shale (approximately 4 %) (Table 1). 3.2 Elastic Velocity at Room Pressure and Elastic Anisotropy 3.3 Crack Density Characterization
Before performing the uniaxial loading test, the bulk den- Comparing the stress sensitivity coefficient to crack density sity q of the core samples was determined. Then, the is possible using cracking models. According to the non- compressional and shear velocities of each sample along interactive approximation (NIA), an elastic solid contain- the longitudinal (Vp(L), Vs(L)) and diametrical (Vp(D), ing many cracks can be treated as the sum of individual Vs(D)) directions at room pressure were measured using sources of extra strains due to each singular crack. Thus, the classical pulse transmission technique. With the for planar cracks of circular shape, the extra compliance velocity measurements, two independent isotropic elastic due to cracks can be calculated (Kachanov 1994). parameters (Poisson ratio m, Young’s modulus E) can be According to NIA theory, the crack density qc can easily obtained. Table 1 listed all the basic physical and be estimated from bulk modulus K of the sample for iso- mechanical parameters of the samples. According to tropic sample as follows Rickman et al. (2008), the brittleness index BI of all the K0 16 ð1 m20 Þ samples was calculated with the elastic parameters. The BI ¼1þ q ð3Þ K 9 ð1 2m0 Þ c was between 30 and 50 %. From the velocities measurements at room pressure, we where K0 and t0 are the bulk modulus and Poisson ratio of can make a simple assumption that P wave anisotropy e and the solid matrix, respectively. The detailed information has S wave anisotropy c can be obtained as follows been given in Schubnel and Guéguen (2003). According to Thomsen (1986), since the parameters of Vp ðLÞ Vp ðDÞ e¼ 100 elastic anisotropy of the samples are lower than 100 % min Vp ðLÞ; Vp ðDÞ ð2Þ (Sect. 3.2), the samples are actually weakly anisotropic jVs ðLÞ Vs ðDÞj because of the presence of bedding. The cracking models c¼ 100 minfVs ðLÞ; Vs ðDÞg for anisotropic samples have more assumptions and more
Table 1 Basic physico-mechanical parameters of the shale samples
Sample group ID Vp(L) (km/s) Vs(L) (km/s) Vp(D) (km/s) Vs(D) (km/s) e (%) c (%) q (g/cm3) m E (GPa) BI (%)
123 Evaluation of the Micro-cracks in Shale from the Stress Sensitivity of Ultrasonic Velocities
complexity, as it needs at least five or nine velocities. Thus,
although we are aware that the presence of micro-cracks in a rock contributes to increase its seismic anisotropy (e.g., Pappalardo et al. 2016), in this paper we treat all the weakly anisotropic samples as isotropic material. This is for simplicity to invert the approximated initial crack density from the velocities measurements at room pressure according to Eq. (3). The crack densities of the 20 samples are shown in Fig. 3b. The average crack densities of the Lujiaping-par- allel, Lujiaping-normal, Longmaxi-parallel and Longmaxi- normal shale samples are 0.07, 0.08, 0.09 and 0.23, respectively. The crack density of the Longmaxi-normal sample is much larger than that of the other samples.
4 Discussion
4.1 The Anisotropy and Mineralogy Fig. 4 Scanning electron micrograph of the Longmaxi shale sample
In Sects. 3.1 and 3.2, the elastic anisotropy of Longmaxi
shale and Lujiaping shale is 20 and 4 %, respectively. 4.2 The Preferred Orientation of Stress Sensitivity However, the stress sensitivity anisotropy for these two and Crack Density types of shale is approximately 119 and 37 %, respectively. The elastic anisotropy and stress sensitivity anisotropy In Sects. 3.1 and 3.3, we found that although the bedding- have a positive correlation. However, the stress sensitivity parallel and bedding-normal samples of Longmaxi shale magnifies the anisotropy, which is more sensitive to the were cored from the same block, the stress sensitivity and microscale anisotropy than that of elastic velocities. crack density along the bedding-normal direction are much The mineralogy of the samples constrained by powder larger than those along the bedding-parallel direction X-ray diffraction analysis shows that for Lujiaping shales, because the compressional velocities are more sensitive to the quartz ? feldspar ? pyrite (QFP), carbonate and clay the crack aperture direction than to the crack radius contents do not vary significantly, with average values of direction (Vernik and Milovac 2011; Sarout and Guéguen approximately 65, 20 and 15 %, respectively. However, the 2008). This is also in good agreement with Fig. 2 that the Longmaxi group shales represent a wide range of miner- stress sensitivity coefficients along the bedding-normal alogy. The QFP, carbonate and clay contents vary from 40 direction have better goodness of fit than that along the to 80, 0 to 30 and 13 to 50 %, respectively. In previous bedding-parallel direction. The SEM analysis in Fig. 4 studies, organic shales are generally anisotropic due to the again proves that the stress sensitivity and crack density of occurrence of clay minerals and kerogen (Sarout and bedding-normal samples are much larger than those of Guéguen 2008; Vernik and Liu 1997; Vernik and Milovac bedding-parallel samples. 2011). The clay particles are made of clay platelets. They are aligned with strongly anisotropy. The alignments create 4.3 Stress Sensitivity and Crack Density more or less large contact zones in the plane of sedimen- tation (bedding plane). These contact zones (micro-cracks) Figure 5 is a plot of the stress sensitivity coefficient versus are much more compliant in the direction perpendicular to the crack density of the 20 shale samples. The stress sen- the bedding plane. Thus, the SEM analysis of the bedding- sitivity and crack density have a positive correlation, and parallel Longmaxi shale is also provided in Fig. 4, which the goodness of fit is as high as 0.97, which proves that the shows several long and thin micro-cracks inside the sample stress sensitivity is efficient for evaluating the micro-crack with a preferred orientation sub-parallel to the bedding. development in shale samples. Obtaining the stress sensi- Our results of the mineralogy and SEM analysis showed tivity under uniaxial stress is much easier than calculating that the clay contents of Longmaxi shale are greater than the crack density in the laboratory because the model has those of Lujiaping shale and also vary significantly, which some assumptions, which may influence the result is one reason for the elastic anisotropy of Longmaxi shale. resolution.
123 X. Wang et al.
References
Brace WF, Walsh JB, Frangos WT (1968) Permeability of granite
under high pressure. J Geophys Res 73:2225–2236 Cipolla CL, Warpinski NR, Mayerhofer M, Lolon EP, Vincent MC (2008) The relationship between fracture complexity, reser- voir properties, and fracture-treatment design. In: SPE 115769, presented at the SPE annual technical conference and exhibition. Denver, Colorado, USA; September 21–24, 2008 Jahandideh A, Jafarpour B (2014) Optimization of hydraulic fractur- ing design under spatially variable shale fracability. SPE: Society of Petroleum Engineers, April 17, 2014 Kachanov M (1994) Elastic solids with many cracks and related problems. Adv Appl Mech 30:259–445 Kwon O, Kronenberg A, Gangi A (2001) Permeability of Wilcox shale and its effective pressure law. J Geophys Res Fig. 5 Plot of the stress sensitivity parameter versus the crack density 106(B9):19339–19353 of 20 shale samples Li Q, Chen M, Zhou Y, Jin Y, Wang FP, Zhang R (2013) Rock Mechanical Properties of Shale Gas Reservoir and their Influ- ences on Hydraulic Fracture. In: IPTC: International petroleum 5 Conclusions technology conference, March 26, 2013 Ma Y, Zhong N, Li D, Pan Z, Cheng L, Liu K (2015) Organic matter/clay mineral intergranular pores in the lower Cambrian 1. Ultrasonic wave velocities are highly sensitive to crack Lujiaping shale in the north-eastern part of the upper Yangtze closure and propagation. Thus, a new and simple area, China: a possible microscopic mechanism for gas preser- method was established to characterize the micro- vation. Int J Coal Geol 137:38–54 Pappalardo G, Punturo R, Mineo S, Ortolano G, Castelli F (2016) crack development in shale samples from the stress Engineering geological and petrographic characterization of sensitivity of ultrasonic velocities in this study. migmatites belonging to the Calabria-Peloritani Orogen (South- 2. The experimental results of two types of shale outcrops ern Italy). Rock Mech Rock Eng 49:1143–1160 show that the stress sensitivity coefficients of the Rickman R, Mullen MJ, Petre JE, Grieser WV, Kundert D (2008) A practical use of shale petrophysics for stimulation design velocity have a strong and positive correlation with the optimization: all shale plays are not clones of the Barnett shale. initial crack density, which proved that the stress SPE: Society of Petroleum Engineers, January, 1, 2008 sensitivity was efficient for evaluating the micro-cracks Sarout J, Guéguen Y (2008) Anisotropy of elastic wave velocities in in shale samples. This method can be easily conducted deformed shales: Part1-Experimental results. Geophysics 73(5):D75–D89. doi:10.1190/1.2952744 in the laboratory and is useful for engineering work. Schubnel A, Guéguen Y (2003) Dispersion and anisotropy of elastic 3. The stress sensitivity coefficients of velocity are orienta- waves in cracked rocks. J Geophys Res 108:2101–2116 tion preferred; thus, they quantitatively provide the micro- Sondergeld CH, Newsham KE, Comisky JT, Rice MC, Rai CS (2010) crack developments. The stress sensitivity is more sensi- Petrophysical considerations in evaluating and producing shale gas resources. SPE: Society of Petroleum Engineers, January 1, 2010 tive to the microscale anisotropy than elastic velocities. Thomsen L (1986) Weak elastic anisotropy: applied seismic 4. The experimental results for two types of shale anisotropy: theory, background, and field studies. Geophys Repr outcrops show that the stress sensitivity coefficients Ser 20:34–46 of Longmaxi shales normal to bedding are much larger Vernik L, Liu X (1997) Velocity anisotropy in shales: a petrophysical study. Geophysics 62(2):521–532 than those of the other samples due to the preferred- Vernik L, Milovac J (2011) Rock physics of organic shales. Lead oriented micro-cracks, which may be induced by the Edge 30:318–323 clay mineral contents. Walsh JB (1965) The effect of cracks on the uniaxial compression of rock. J Geophys Res 70:399–411 Zeng W, Zhang J, Ding W, Zhao S, Zhang Y, Liu Z, Jiu K (2013) Acknowledgments This work was supported by the National Natural Fracture development in paleozoic shale of Chongqing area Science Foundation of China (Grant No. 41304141), the Joint Funds (South China). Part one: fracture characteristics and comparative of the National Natural Science Foundation of China (Grant No. analysis of main controlling factors. J Asian Earth Sci U1562215) and the National Basic Research Program of China (Grant 75:251–266 No. 2015CB250903).
Compliance with Ethical Standards
Conflict of interest The authors declare that they have no conflict of
