Académique Documents
Professionnel Documents
Culture Documents
10.1190/1.3508874
Manuscript received by the Editor 31 January 2010; revised manuscript received 22 June 2010; published online 8 December 2010.
1
Formerly at Stanford University, Stanford, California, U.S.A.; presently at OHM-Rock Solid Images, Houston, Texas, U.S.A. E-mail: a.morcote@ohmrsi
.com.
2
Stanford University, Department of Geophysics, Stanford Rock Physics Laboratory, Stanford, California, U.S.A. E-mail: mavko@stanford.edu.
3
Colorado School of Mines, Center for Rock Abuse, Department of Petroleum, Engineering, Golden, Colorado, U.S.A. E-mail: mprasad@mines.edu.
© 2010 Society of Exploration Geophysicists. All rights reserved.
E227
Downloaded 21 Feb 2012 to 216.198.85.26. Redistribution subject to SEG license or copyright; see Terms of Use at http://segdl.org/
E228 Morcote et al.
coal, three of bituminous coal, two of semianthracite, one of anthra- EXPERIMENTAL RESULTS
cite, and one powder sample from bituminous coal.
Porosity and grain density
EXPERIMENTAL TECHNIQUE Table 1 summarizes porosities and densities obtained in this
study. Coal rock samples have low porosities ranging from 1.5 to
Cylindrical core plugs, 2.5 cm in diameter and 2.0– 4.0 cm in 4.9%. In contrast, the powder sample has a porosity of 20%, which is
length, were taken in two directions: parallel 共90°兲 and perpendicu- very high compared to the rock samples.
lar 共0°兲 to lamination surfaces. The powder from bituminous coal Dry density ranges from 1.11 to 1.67 g / cm. Cannel coal has the
was precompressed at 0.6 MPa. Nine core plugs were dried for lowest dry density whereas humic coals 共bituminous, semianthra-
48 hours in an oven at 60 ° C. cite, and anthracite samples兲 have dry densities that increase as rank
Porosity was measured using a helium porosimeter. This instru- increases. Density and porosity were calculated before pressurizing
ment measures porosity by displacement of pore volume by helium the samples.
at known pressures. Grain and pore volumes were determined by
Boyle’s Law. In this method, the grain volume is calculated from the
volume of helium and the total volume is obtained from the dimen-
Ultrasonic velocities
sions of the plugs. Laboratory data measured in different types/ranks of coal 共Ap-
The P-wave 共VP兲 and S-wave 共VS兲 velocity measurements were pendix A兲 show differences between velocities measured normal
taken using the pulse transmission technique 共Birch, 1960兲. Veloci- and parallel to lamination surfaces. They also show the effect of con-
ties were measured parallel and perpendicular to bedding planes fining pressure in velocities. In addition, the data show a very impor-
with the propagation and vibration directions aligned parallel and tant relationship between coal rank and elastic properties: The veloc-
normal to lamination. In samples with no visible lamination 共powder ities increase as coal rank increases.
and anthracite兲, velocities were measured in only one arbitrary di- Velocities parallel to lamination surfaces are higher than veloci-
rection. ties perpendicular to them and they increase with increasing confin-
The experimental setup consists of a digital Tektronix 共Model ing pressure 共Figures 1 and 2兲. We could not identify lamination sur-
TDS 420A兲 oscilloscope and a Velonex 共Model 345兲 pulse genera- faces on the anthracite sample, and the powder sample does not have
tor. The sample was jacketed with rubber tubing to isolate it from the any internal fabric. These two samples have the highest 共anthracite兲
confining pressure medium. Piezoelectric 共PZT兲 crystals mounted and the lowest 共powder兲 velocities.
on steel endplates were used to generate P- and S-waves. The princi- Pressure dependence of velocities on confining pressure is greater
pal frequency was approximately 1 MHz for P-waves and 0.7 MHz at pressures below 5 MPa. At higher pressures, velocities increase
for S-waves. A couplant was used to bond the endplates to the sam- slightly. The highest variation with pressure is exhibited by the pow-
ple. The experimental configuration allowed simultaneous measure- der sample, with a velocity increase of approximately 35%.
ments of VP and VS at various confining pressures. In rock samples, velocities parallel to lamination surfaces in-
The value for VP was picked from the first signal arrival, and VS creased between 4% and 20% as a function of confining pressure and
was picked on the first positive peak of the shear signal. The error in velocities normal to lamination surfaces increased between 3% and
velocity measurement is estimated to be approximately 1% due to 12%. Confining pressure was gradually increased from 0 to 40 MPa
operator error in picking the first arrival. The system delay time was along loading and unloading cycles. During unloading cycles, ve-
measured by taking head-to-head time at 2 MPa. The traveltime cal- locities decreased following almost the same path of the velocities
ibration was confirmed by measuring an aluminum cylinder at dif- during the loading cycle. That was the case for most of the examples
ferent pressures. except the powder. Thus, rock samples showed a small hysteresis
Confining pressure varied up to 40 MPa in increments of 5 MPa whereas the powder sample showed the largest hysteresis.
along loading and unloading cycles. At lower pressures The humic coals used in this study are, in increasing order of rank,
共0 – 2.5 MPa兲, increments were of 0.5 MPa. bituminous, semianthracite, and anthracite. From measured data, we
Sample Inventory
Downloaded 21 Feb 2012 to 216.198.85.26. Redistribution subject to SEG license or copyright; see Terms of Use at http://segdl.org/
Elastic properties of coal E229
observed that VP and VS increase as coal rank increases 共Figure 3兲. 4.0
Therefore, bituminous coal shows lower velocities than anthracite,
which exhibits the highest velocities. On the other hand, cannel coal 3.5
Velocity (km/s)
2.5
Dry bulk moduli 共Kdry兲 and dry shear moduli 共Gdry兲 increase
2.0
with confining pressure and coal rank, as shown in Figure 4 and Fig-
ure 5. These moduli are computed assuming constant density for the
1.5
different confining pressures. In low-porosity rocks such as these
coals, changes in porosity with confining pressure are related to the 1.0
closing of microcracks. Consequently, those changes are small. The
error in density is estimated to be less than 1%. The computed elastic 0.5
moduli show small variations with respect to confining pressure
Cannel Bituminous Semi-anthracite Anthracite
compared to the effect of thermal maturity, even for pressures below 0.0
5 MPa. At higher pressures, changes in moduli are even smaller, Figure 3. Range of velocity variation for different coal ranks. Solid
reaching an asymptote. The larger variations in moduli are associat- boxes show the range of variation in VP and dashed boxes show the
range of variation in VS. Note that velocities on humic coals increase
4.0 as the coal rank increases.
18
3.5 Anthracite
16 Semi-anthracite
Bituminous
3.0 14 Cannel
12
VP (km/s)
2.5
Kdry (GPa)
10
2.0
8
1.5 Anthracite 6
Semi-anthracite
Bituminous 4
1.0
Cannel
Powder 2
0.5
0 5 10 15 20 25 30 35 40
0
Confining Pressure (MPa) 0 5 10 15 20 25 30 35 40
Confining Pressure (MPa)
Figure 1. VP as a function of confining pressure for core plugs paral-
lel 共solid lines兲 and perpendicular 共dashed lines兲 to lamination sur- Figure 4. Calculated dry bulk modulus as a function of confining
faces. VP increases as coal rank increases. pressure. Note that dry bulk modulus increases as the coal rank in-
creases.
2.0 8
Anthracite
1.8
Semi-anthracite
7
Bituminous
1.6 Cannel
6
1.4
1.2 5
VS (km/s)
Gdry (GPa)
1.0
4
0.8
3
0.6 Anthracite
Semi-anthracite
0.4 2
Bituminous
Cannel
0.2 Powder
1
0.0
0 5 10 15 20 25 30 35 40 0
0 5 10 15 20 25 30 35 40
Confining Pressure (MPa) Confining Pressure (MPa)
Figure 2. VS as a function of confining pressure for core plugs paral- Figure 5. Calculated dry shear modulus as a function of confining
lel 共solid lines兲 and perpendicular 共dashed lines兲 to lamination sur- pressure. Note that dry shear modulus increases as the coal rank in-
faces. creases.
Downloaded 21 Feb 2012 to 216.198.85.26. Redistribution subject to SEG license or copyright; see Terms of Use at http://segdl.org/
E230 Morcote et al.
ed with coal rank. Therefore, bituminous and cannel coals have low- Castagna et al. 共1993兲, Yu et al. 共1993兲, and Yao and Han 共2008兲.
er moduli than semianthracite and anthracite, the latter of which has This compilation of coal-velocity data shows an empirical linear re-
the highest rank and moduli. lationship that can be approximated by
VP and VS anisotropy
VS (km/s)
0.0
where VP 共90°兲 is VP parallel to lamination surfaces and VP 共0°兲 is VP
0 0.5 1.0 1.5 2 2.5 3 3.5 4 perpendicular to lamination surfaces. Similarly, S-wave anisotropy
VP (km/s) 共 ␥ 兲 can be obtained from
Figure 6. VP versus VS for dry samples measured in this study. The VS共90°兲 ⳮ VS共0°兲
␥⳱ , 共5兲
line shows the best linear fit for this data set. VS共0°兲
2.0
1.8
VS = 0.4811VP + 0.00382
1.6 2.3
1.4 2.2
1.2
VS (km/s)
2.1
1.0
2.0
VP/VS
0.8
Anthracite
0.6 Semi-anthracite 1.9
Bituminous
0.4 Cannel 1.8
Powder Anthracite
0.2 Castagna et al. (1993) Others Semi-anthracite
1.7 Bituminous
Cannel
1 1.5 2 2.5 3 3.5
1.6
VP (km/s) 0 5 10 15 20 25 30 35 40
Figure 7. VP versus VS for different coals. The solid line corresponds Confining Pressure (MPa)
to the best linear fit to all data, which includes dry and wet samples. Figure 8. VP / VS ratio as a function of confining pressure for different
For comparison purposes, the graph also shows the fit obtained for types of coal. Note that the velocity ratio decreases as the coal rank
some bituminous coal samples by Castagna et al. 共1993兲. increases.
Downloaded 21 Feb 2012 to 216.198.85.26. Redistribution subject to SEG license or copyright; see Terms of Use at http://segdl.org/
Elastic properties of coal E231
where VS 共90°兲 is VS parallel to lamination surfaces and VS 共0°兲 is VS effective confining pressures. Below 5 MPa, the VP-VS relationship
perpendicular to lamination surfaces. shows variation as a function of pressure. Therefore, equation 1
Figure 9 shows P-wave and S-wave anisotropy as a function of should be used with caution at low confining pressures. In addition,
confining pressure. As explained earlier, the dependence of veloci- we calculated another empirical VP-VS relationship 共equation 2兲 that
ties on confining pressure is greater at pressures below 5 MPa. How- fits most of the data from this study 共dry samples兲 and data from the
ever, at high confining pressures, bituminous coal and semianthra- aforementioned papers 共wet and dry samples兲.
cite show high anisotropy. Note that bituminous coal shows increas- Semianthracite and bituminous coal exhibit high P-wave and
ing anisotropy at low confining pressures. S-wave anisotropy that do not depend on cracks. When cracks are
randomly oriented, the increase of velocity under increasing confin-
DISCUSSION ing pressure should remain independent of direction. However,
rocks with a nonrandom orientation of cracks will exhibit velocity
In this study, pressure dependence of acoustic velocities is greater anisotropy 共Nur and Simmons, 1969兲, which should decrease as the
at pressures below 5 MPa. At higher confining pressures, velocities cracks close under increasing confining pressure. The trends ob-
vary slightly. This finding is in agreement with the data reported by served in Figure 1 and Figure 2 suggest that the cleats are closed
Yu et al. 共1991兲 although in a later work Yu et al. 共1993兲 report the de- when the confining pressure reaches a value of 5 MPa; however, the
pendence of velocities on confining pressure up to 10 MPa. The samples present significant anisotropy at pressures up to 40 MPa.
presence of pore space and microcracks tend to elastically soften the The simplest explanation for pressure dependence is the presence
rock, producing a decrease in VP and VS 共King, 1966兲. The gradual of compliant cracks or grain-to-grain contacts. Bedding anisotropy
increase in velocities at low confining pressures is caused by the clo- is enhanced by having horizontal cracks parallel to the bedding.
sure of microcracks. As the confining pressure increases, the rock Closing these cracks with pressure reduces the anisotropy. On the
frame gets stiffer, causing an increase in elastic properties. other hand, vertical open cracks tend to counteract the effect of bed-
The dependence of elastic properties on pressure shows an elastic ding. In theory, enough vertical cracks can give negative and ␥ .
and reversible behavior in the samples, except in the powder. In oth- Consequently, horizontal velocities can be smaller than vertical ve-
er words, by reducing the confining pressure 共unloading cycle兲, the locities. Sealing the vertical cleats via pressure causes the anisotropy
velocities decrease following almost the same path of the velocities to become more and more dominated by the fabric and can cause a
during the loading cycle, causing little hysteresis. On the other hand, shift from negative to positive and ␥ .
the powder exhibits larger hysteresis due to compaction. The anisotropy presented by the samples at high pressures seems
We obtained a linear empirical VP-VS relationship for dry samples to be related to fine-scale lamination 共Figure 10兲 and the preferred
共equation 1兲, which covers a wide range of coal ranks, porosities, and orientation of the carbon-based minerals. As the coal rank increases,
the carbon-based minerals get more oriented along lamination and
the pore space decreases 共Van Krevelen, 1993兲.
0.15
CONCLUSION
0.10
Coal thermal maturity has a significant influence on dynamic elas-
tic properties of coal. Bituminous coal has lower velocities than
Anisotropy
0.05
semianthracite and anthracite; the latter has the highest rank and ve-
locities. Dry bulk and dry shear moduli increase with increasing coal
0.00 rank whereas the VP-VS ratio decreases with increasing coal rank.
Coal velocities also depend on confining pressure. The depen-
ε−Semi-anthracite
γ−Semi-anthracite dence of velocities on confining pressure is greater at lower pres-
−0.05
ε−Bituminous sures up to 5 MPa and is due to the presence of microcracks; above
γ−Bituminous
ε−Cannel this pressure, changes in velocities are minimal.
−0.10 γ−Cannel The VP-VS relationship is empirical and is approximately linear
0 5 10 15 20 25 30 35 40 over a wide range of velocities, coal ranks, and effective pressures.
However, it should be used with caution at confining pressures lower
Confining Pressure (MPa)
than 5 MPa.
Figure 9. P-wave anisotropy 共solid lines兲 and S-wave anisotropy Semianthracite and bituminous coal exhibit high P-wave and
共dashed lines兲 as a function of confining pressure for different coal S-wave anisotropy at high pressures that does not depend on the
ranks.
presence of cleats 共cracks兲. The anisotropy at high pressures might
be due to fine lamination and preferred orientation of the carbon
minerals.
ACKNOWLEDGMENTS
This work was financially supported by the Stanford Rock Phys-
ics Laboratory and DOE 共grant DE-FC26-01BC15354兲. Special
Figure 10. Sample of cannel coal 共left兲 and bituminous coal 共right兲 thanks to Juan-Mauricio Florez, Kevin Wolf, Kyle Spikes, and Jorge
showing lamination. This bedding induces the intrinsic anisotropy Mejia. We also thank our Geophysics reviewers and the associated
observed in the samples. Bright bands 共BB兲 and dull bands 共DB兲. editor for their comments to improve the manuscript.
Downloaded 21 Feb 2012 to 216.198.85.26. Redistribution subject to SEG license or copyright; see Terms of Use at http://segdl.org/
E232 Morcote et al.
APPENDIX A
Downloaded 21 Feb 2012 to 216.198.85.26. Redistribution subject to SEG license or copyright; see Terms of Use at http://segdl.org/
Elastic properties of coal E233
Downloaded 21 Feb 2012 to 216.198.85.26. Redistribution subject to SEG license or copyright; see Terms of Use at http://segdl.org/
E234 Morcote et al.
Downloaded 21 Feb 2012 to 216.198.85.26. Redistribution subject to SEG license or copyright; see Terms of Use at http://segdl.org/