Vous êtes sur la page 1sur 6

http://www.paper.edu.

cn
Paper 2A 04 — SINOROCK2004 Symposium
Int. J. Rock Mech. Min. Sci. Vol. 41, No. 3, CD-ROM, © 2004 Elsevier Ltd.

NUMERICAL INVESTIGATION OF THE MECHANICAL BEHAVIOR OF ROCK UNDER


CONFINING PRESSURE AND PORE PRESSURE

C.A. Tang 1, 2, T. Xu 1, 2 , T.H. Yang 1 , Z.Z. Liang 1


¹) Centre for Rock Instability and Seiscimity Research, Northeastern University, Shenyang,
110004, China
tca@mail.neu.edu.cn
²) Research Centre for Numerical Tests on Material Failure, Dalian University, Dalian,
116622, China
xukyle@163.com

Abstract: Many of the important problems of rock engineering are concerned with mechanical behaviours of
rock where the internal rock structure sustains pore pressure and confining pressure from the surrounding
rocks. A basic understanding of rock mechanical properties under confining pressure and pore pressure
conditions is of great importance in rock mechanics and rock engineering. In this paper, the newly-developed
pore-pressure incorporated Rock Failure Process Analysis model (RFPA) is briefly outlined at first. Then a
series of numerical tests on rock under different constant confining pressure and pore pressure are conducted
to illustrate how the overall macroscopic responses and mechanical properties of brittle heterogeneous rocks
under different confining pressure and pore pressure were revealed by RFPA code. In addition, through the
modelling of acoustic emission sequences in rock progressive failure, the AE characteristics and the
correlation between AE events and stress-strain curves under different confining pressure and pore pressure
were also investigated. From the numerically simulated results, it can be possible to analyze large-scale
practical rock engineering problems such as mining induced seismicit ies and rock bursts.
Keywords: Numerical simulation, confining pressure, pore pressure, mechanical behaviours, acoustic
emission, RFPA.

1. INTRODUCTION fracturing interaction events also causes the


A basic understanding of rock mechanical complexity of mechanical breakdown of
properties under different stress conditions is of heterogeneous rocks. Numerical models that
great importance in rock mechanics and rock simulate the detailed fracturing sequence are thus
engineering since many of the important problems useful for understanding rock failure mechanisms
of rock engineering are concerned with mechanical under pore pressure and confining pressure. In this
behaviours of rock where the internal rock paper, the further improved Rock Failure Process
structure sustains pore pressure and confining Analysis (RFPA) code by integration of pore-
pressure. Pore pressure greatly affects the pressure was employed to investigate the
probability of rock failure. Meanwhile, rock in mechanical properties of brittle heterogeneous
deep ground also suffers from confining pressure rocks. A series of numerical tests on rock under
from the surrounding rocks. Considerable attention different confining pressure and pore pressure were
has been given to the mechanical behaviours of conducted illustrating how the overall macroscopic
rocks under different confining pressure and pore response of brittle heterogeneous rocks was
pressure by laboratory tests and in-situ presented by pore-pressure incorporated RFPA.
measurements, but the general precise theoretical Moreover, through the modelling of acoustic
formulations and laws of rock mechanical emission sequences, the AE characteristics of rock
properties are still not impossible due to its extreme in failure process were also investigated to gain
complexity. When rock is subjected to stress, some possible insight into some large-scale
multiple cracks can nucleate, propagate, interact practical problems such as mining induced
and coalesce, which induces the change of the pore seismicit ies and rock bursts.
pressure in the rock. This complex multiplicity of

1
中国科技论文在线 http://www.paper.edu.cn

Paper 2A 04 — SINOROCK2004 Symposium


Int. J. Rock Mech. Min. Sci. Vol. 41, No. 3, CD-ROM, © 2004 Elsevier Ltd.

2. PHILOSOPHY OF RFPA The constitutive equation of deformation fields


Numerical simulation is currently the most can be expressed for elastic isotropic materials.
popular method used for modelling deformation σ' ij = λδij εv + 2Gεij
behaviour of rock-like materials. RFPA code, as a
where G is shear modulus and is Lame’s
numerical tool for simulating the geomaterials, has
constant.
been widely used in investigating and contrasting On the basis of the above the equilibrium, the
the simulated results and laboratory findings continuity, and the constitutive equations, the
(Wong & Lin 2001; Tang & Lin 2001; Xu and governing equations for rock deformation
Tang, 2003). On the basis of the RFPA code, considering the gas pressure in rock can be
besides considering the deformation of an elastic
represented as:
material containing an initial random distribution
of micro-features, the interactions between the ( λ + G ) ⋅ u j , ji + Gui , jj + f j + (α ⋅ p) , i = 0
individual elements, pore pressure and stress are For heterogeneity, the material mechanical
also taken into account in the code. The governing parameters (failure strength 0 and elastic modulus
equations for rock deformation in further E0 ) for elements are randomly distributed
developed RFPA code mainly consist of the throughout the specimen by following a Weibull
equilibrium, the continuity, and the constitutive distribution, a detailed information can be referred
equations. to published literature (Tang & Tham, 2000).
According to elastic theory, equilibrium
equation is given
3. NUMERICAL MODEL
σij , j + f j = 0
(1) 1

where σij is the stress tensor,( i , j = 1,2,3 ), MPa.


f j is the body forces per unit volume, MPa.
The generalized effective stress principle based
on Terzaghi’s law was invoked in the stress P 3

2
equilibrium equations:

σij = σ 'ij + α ⋅ p ⋅ δij


(2)
where σij is the solid total stress, σij is the solid Figure 1. Numerical model
'

effective stress, P is the pore pressure, α is a


positive constant equal to 1 when individual grains The mesh for the plane strain numerical sample
are much more incompressible than the grain consists of 200×100 elements with geometry of
100mm×50mm in size (as shown in Fig.1), and all
skeleton, and δij is the Kronecker delta function.
the elements have the same size in scale (square in
Substitution of equation (2) into equation (1) shape). The pore pressure in rock specimen is
leads to: denoted as P, confining pressure and axial pressure
acted on numerical rock specimen are respectively
σ ' ij , j + f j + (α ⋅ p ⋅ δij ) , j = 0
(3) denoted as σ0 and σ1 . The elements are
Thus, the equilibrium equation is expressed characterized by their failure strength, σ0 , Young’s
according to the effective stress principle. modulus, E0 , and Poisson’s ratio, ν. The elements
According to the continuous conditions, the provide resistance against compressive or tensile
geometrical equation can be expressed deformations that are governed by constitutive
equations described above. In order to consider the
1 heterogeneity of rock specimen, a widely used
εij = ( ui , j + u j ,i ) Weibull distribution (Weibull, 1951) was
2 (4)
introduced to describe the material properties of
where εij is strain tensor,( i , j = 1,2,3 ). εv is the elements such as failure strength, Young’s
volumetric strain and u is the displacement of modulus, and Poisson’s ratio at mesoscopic level.
element. The input material mechanical properties
parameters used to simulate numerical model rock

2
中国科技论文在线 http://www.paper.edu.cn
Paper 2A 04 — SINOROCK2004 Symposium
Int. J. Rock Mech. Min. Sci. Vol. 41, No. 3, CD-ROM, © 2004 Elsevier Ltd.

specimen are listed in the Table 1 below. An deformation up to the peak strength. At low
external displacement at a constant rate of confining pressures, the curves show defined peak
0.002mm/step in the axial direction was applied to strength and a gradual strength decrease in the post
the rock specimen and the stress acted on rock failure region until final deformation occurs at
specimen as well as induced deformation in each about constant axial stress, i.e., residual strength.
element were computed in the numerical tests. At higher confining pressures, the rock exhibits
work-hardening and the Young’s modulus of rock
Table 1. Mechanical parameters of numerical is higher than that of rock at lower pressure.
model. Meanwhile, transit ion from brittle to ductile
Mechanical parameters Value deformation in rock with an increase in confining
m 1.5 pressure was also clearly demonstrated in Fig.3.
E0 / GPa 30 70

σ0 / MPa 200
60
0

φ/ °
2
30 4

µ
50
0.25 8

16
ξ 0.2 40

1 / MPa
P/MPa 0 1 3

¦ Ò
30

σ3 / MPa 0, 2, 4, 8, 16 20

10

4. EFFECT OF CONFINING PRESSURE 0


0 0.05 0.1 0.15 0.2 0.25 0.3

¦ 1Å/%

4.1 Deformation and strength behaviors


Figure 3. Complete stress-strain curves of rock
0MPa 4MPa 8MPa specimens
70

60

50
UCS/ MPa

40

Figure 2. Macroscopic failure patterns of 30

specimens 20

10
Figure 2 is given the numerically simulated
0
macroscopic failure patterns of model specimens -5 0 5 10 15 20
Tension Compression
under different confining pressures and the
correspondingly numerical complete axial stress Confining pressure/ MPa
versus axial strain curves of rock at constant Figure 4. Curve between compressive strength of
confining pressure up to 16 MPa with no pore rock specimens and confining pressure
pressure are presented in Fig.3. As shown in Fig.2,
the angle between the failure plane and the Figure 4 gives the relationship curve between
maximum principal stress direction in uniaxial peak strength of rock specimens and confining
compression is about 30 degrees, and the angle pressure at failure and Figure 5 is numerically
between macroscopic failure plane and the obtained failure envelopes of rock specimens. As
maximum principal stress direction gradually can be seen from Figure 4 and 5, the ultimate
increases with the increase of confining pressure compressive failure strength, i.e., peak strength of
acted on the rock specimens, which agrees well numerical rock specimens gradually increases with
with theoretical predictions. It can be seen from the confining pressure. Even though the linear Mohr-
stress-strain curves in Fig. 3, the rock deforms Coulomb failure criterion with tension cut-off is
linearly and elastically at axial stresses below the adopted in the model, the macroscopic failure
yield strength which is dependent on the confining envelope is concave towards the σ axis. The
pressure. Further compression leads to inelastic numerical results indicate that the macroscopic

3
中国科技论文在线 http://www.paper.edu.cn
Paper 2A 04 — SINOROCK2004 Symposium
Int. J. Rock Mech. Min. Sci. Vol. 41, No. 3, CD-ROM, © 2004 Elsevier Ltd.

70 600 70 600
non-linear phenomena such as rock failure in 60
σ3 =2 MPa 500 60
σ3 =4MPa 500

nature can be described and revealed through some 50 50

AE accounts(N)
400 400

AE accounts(N)
Ò 1/MPa

1/MPa
40 40
simple linear rules at mesoscopic level. In addition, 30
300
30
300

Ò
¦

¦
200 200
it is noticing that the residual strength (or 20

10 100
20
10 100
frictional) of rock, also dependent on the confining 0 0 0 0
0 0 .0 5 0 .1 0 .1 5 0 .2 0 .2 5 0 .3 0 0.05 0.1 0.15 0.2 0.25 0.3
pressure, increases with confining pressure. For Å
¦ 1/ %
¦Å 1/%

rock materials, fracture and friction are 70


60
600 70
60
600

500 500
macroscopic manifestations of the same processes: 50
3 =8MPa
50 σ3 =16MPa

AE accounts(N)

AE accounts(N)
400 400
e.g., grain crushing, crack growth, healing, and

Ò 1 /MPa

Ò 1 /MPa
40 40
300 300
30 30
plastic yielding. When viewed in this way, it is not

¦
200 200
20 20

surprising that the difference between intact 10 100 10 100

0 0 0 0
strength and residual (or frictional) strength should 0 0.05 0.1 0.15 0.2 0.25 0.3 0 0.05 0.1 0.15 0.2 0.25 0.3
¦Å 1/% ¦Å 1/%
vanish with increasing confining pressure. This is
to say, the rock will ideally exhibit a state of plastic Figure 6. Complete stress-strain curves and AE
flow at extremely high confining pressure. characteristic curves of model
specimens
600 1 600 1
Mohr failure envelope 0.9 0.9
500 0.8 500 0.8
ó3=2MPa 0.7 ó3=4MPa 0.7
AE accounts(N)

400 400

Normalized AEE
AE accounts(N)
Normalized AEE
0.6 0.6
300 0.5 300 0.5
0.4 0.4
200 0.3 200 0.3
0.2 0.2
100 100
0.1 0.1
0 0 0 0
0 9 18 2 7 3 6 4 5 5 4 6 3 7 2 8 1 9 0 9 9 1 0 8 1 1 7 1 2 6 0 9 18 2 7 3 6 4 5 5 4 6 3 7 2 8 1 9 0 9 9 1 0 8 1 1 7 1 2 6

Loading step Loading step


600 1 600 1
0.9 0.9
500 500
0.8 0.8
ó3=8MPa 0.7 0.7
400 400 ó3=16MPa

Normalized AEE

Normalized AEE
AE accounts(N)
AE accounts(N)

0.6 0.6

Figure 5. Simulated failure envelope of model 300 0.5


0.4
300 0.5
0.4

rock specimens
200 0.3 200 0.3
0.2 0.2
100 100
0.1 0.1
0 0 0 0
0 9 1 8 2 7 3 6 4 5 5 4 6 3 7 2 8 1 9 0 9 9 1 0 81 1 7 1 2 6 0 9 1 8 2 7 3 6 4 5 5 4 6 3 7 2 8 1 9 0 9 9 1 0 8 1 1 7 1 26

Loading step Loading step

4.2 AE characteristics
As we know, the monitoring of acoustic Figure 7. AE curves and normalized AE energy
emission (AE) or seismic events has proven to be curves of model specimens under
one of the powerful tools available in analyzing different confining pressure
damage or brittle fracture during rock deformation.
There is generally a good correlation between AE Fig.6 shows the complete stress-strain curves
rate and inelastic strain rate so that the AE rate can and corresponding AE characteristic curves of
be used to quantify damage accumulation occurring model specimens at different constant confining
in the rock sample. Locker (1991), Cox and pressure. Fig.7 shows the AE and corresponding
Meredith (1993) have analyzed catalogues of AE normalized AE energy curves of model specimens
events recorded during compression tests in rock in under different confining pressure. A comparison
terms of the information they give about the among the curves in Fig.6 shows a good
accumulated state of damage in a material. And relationship between the simulated stress-strain
combine this measured damage state with a model curves and the modelled curves of event rate. It can
for the weakening behaviour of cracked solids, be seen from Fig. 6 and 7, in general, a sharp
which shows that reasonable predictions of the increase of AE event rate in AE characteristic
mechanical behaviour are possible. Based on this curves corresponds to an abrupt stress drop in
background knowledge, by recording the counts of complete stress-strain curve and the maximum rate
failed elements, the seismicities associated with the of AE events appears in the post-peak range. It
progressive failure can be simulated in RFPA that indicates that the initiation and propagation of
allows elements to fail when overstressed. In RFPA mesoscopic main-fractur ing which precedes the
code, a single AE event represents a micro-crack final stage macroscopic fracture development has
forming event to indirectly assess the damage occurred in the rock. The results show that the
evolution (Tang 1997, 1998). maximum AE event or main shock emitting from
rock can be regarded as the precursor of macro-

4
中国科技论文在线 http://www.paper.edu.cn

Paper 2A 04 — SINOROCK2004 Symposium


Int. J. Rock Mech. Min. Sci. Vol. 41, No. 3, CD-ROM, © 2004 Elsevier Ltd.

fracture of rock, which is essential for the location pore pressures. As can be seen from Fig. 9, at the
of earthquake source, the search for earthquake same constant confining pressure, the rock
precursors and earthquake prediction research. In specimens with pore pressure have lower peak
addition, the main-fracturing will occur later when strength than those with no pore pressure.
rock specimens are subjected to the same strain Generally, the pore pressure decreases the peak
value with the increase of the constant confining strength of rock at failure and increases the
pressure. The confining pressure enhances the brittleness of rock. Moreover, the nonlinear
ultimate compressive strength and defers the relationship curves between the maximum
occurrence of main-fracturing at failure, which is compressive strength and confining pressure at
termed as the typical confining pressure effect in different constant pore pressure are basically
rock failure. parallel.

5.2 AE characteristics
5. EFFECT OF PORE PRESSURE
80 450 80 450
70 400 70 400
=2MPa =4MPa
5.1 Deformation and strength
3 3
60 350 60 350
300 300

AE accounts
50 50

AE accounts
behaviors
1/MPa

1/MPa
250 250
40 40
200

¦ Ò
200

¦ Ò
30 30
150 150
20 20
100 100
10 50 10
P=2MPa P=4MPa P=8MPa P=16MPa 50
0 0 0 0
0 0.05 0.1 0.15 0.2 0.25 0.3 0.35 0.4 0 0.1 0.2 0.3 0.4
¦ 1Å/%
¦ Å1/%
80 450 80 450

70 400 70 400
3 =8MPa 3 =16MPa
350 350
60 60
300 300

AE accounts
50

AE accounts
50
1/MPa

250

¦ 1Ò/MPa
250
40 40
200
¦ Ò

200
30 30
150 150
20 100 20 100

Figure 8. Macroscopic failure patterns under 10


0
50
0
10

0
50

given pore pressure 1 MPa 0 0.05 0.1 0.15 0.2


¦ 1Å/%
0.25 0.3 0.35 0.4 0 0.05 0.1 0.15 0.2

¦ Å
1
/%
0.25 0.3 0.35 0.4

Figure 10.Complete stress-strain curves and AE


70 characteristic curves of model
60 specimens with pore pressure 1 MPa
50
1 /MPa

40 450 0.45 450 1


P=0MPa 400 0.4 400 0.9
3 =2MPa 3 =4MPa
30 350 0.35 350 0.8
0.7
P=1MPa
Normalized AEE

Normalized AEE
300 0.3 300
AE accounts

20
AE accounts

0.6
250 0.25 250
0.5
200 0.2 200
10 P=3MPa 0.4
150 0.15 150 0.3
100 0.1 100
0 0.2
50 0.05 50 0.1
0 5 10 15 20 0 0 0 0
0 0.04 0.08 0.12 0.16 0 . 2 0.24 0.28 0.32 0.36 0 . 4 0 0.04 0.08 0.12 0.16 0 . 2 0.24 0.28 0.32 0.36 0 . 4

3 /MPa
1(%) 1(%)

450
450 0.45
1 450 0.45
1
400
400
3 =8MPa
0.9
0.4 400 3 =16MPa 0.9
0.4

Figure 9. Relationship between compressive 350


350 0.8
0.35
0.7
350 0.8
0.35
0.7
AEE

300 0.3
Normalized AEE

300
AEE
Normalized AEE

300 0.3

strength and confining pressure for


accounts
AE accounts

0.6
AE accounts
accounts

250
250 0.25 0.6
250
Normalized

0.25
Normalized

0.5 0.5
200
200 0.2
given poro pressure 0.4 200 0.2
0.4
AE

150
150 0.15
AE

0.3 150 0.15


0.3
100
100 0.1
0.2 100 0.1
0.2
50 0.05
0.1 50
50 0.05
0.1
0 0
Several series of numerical tests were carried 0 0.040.08
0 0.04 0.080.11
0.12 0.16
0.15 0.18 00.21
. 2 0.24
0.25 0.28 0.320.35
0.28 0.32 0.360.38
0.4
00
00 0.09
0.040.11 0.14
0.08 0.170.16
0.12 0 . 20.23
0 . 20.25
0.240.28 0.310.32
0.28 0.34 0.36
0.37 0.39
0.4
0

1(%)
out to investigate the effect of pore effect on the 1(%)
1(%)

mechanical properties of rock. The numerically Figure 11.AE and normalized AE energy curves of
simulated macroscopic failure patterns of rock with model specimens with given pore
1MPa pore pressure at different constant confining pressure (1MPa)
pressures are presented in Figure 8. Compared with
the macroscopic failure modes of rock in Fig. 2, Fig.10 shows the complete stress-strain curves
much splitting occurred in the rocks with pore and corresponding AE characteristic curves of
pressure which exhibit remarkably brittle. Fig.9 model specimens with 1MPa pore pressure at
gives the relationship curves between peak strength different constant confining pressure and Fig.11
of rock and applied confining pressure for given gives the AE events and corresponding. As stated

5
中国科技论文在线 http://www.paper.edu.cn
Paper 2A 04 — SINOROCK2004 Symposium
Int. J. Rock Mech. Min. Sci. Vol. 41, No. 3, CD-ROM, © 2004 Elsevier Ltd.

above, the simulated stress-strain curves, the AE 8. REFERENCES


event rate curves and the normalized AE energy
curves have a good correlation. In contrast to Fig.6 Cox, S.J.D., Meredith, P.G. 1993. Microcrack
formation and material softening in rock
and 7, it can be seen that the main-fracturing will
measured by monitoring acoustic emissions. Int
occur earlier when rock specimens are subjected to
the same confining pressure with increasing pore J Rock Mech Min Sci Geomech Abstr 30(1):
pressure. The pore pressure reduces the ultimate pp.11-24.
compressive strength and accelerates the Lockner, D.A., Byerlee, J.D., Kuksenko, V., et al.
occurrence of main-fracturing at failure, which can 1991. Quasi-static fault growth and shear
be termed as the pore pressure effect in rock fracture energy in granite. Nature 350(7) : pp.
failure. The confining pressure effect and the pore 39-42.
pressure effect are greatly indicative for the
Tang, C.A. 1997. Numerical simulation of
earthquake prediction and hydraulic fracturing
progressive rock failure and associated
research.
seismicity. Int. J. Rock Mech. Min. Sci. 34:
pp.249-262.
6. CONCLUSSIONS Tang, C.A., Kaiser, P.K. 1998. Numerical
The problems of deformation and failure in Simulation of Cumulative Damage and Seismic
heterogeneous rock at confining pressure and pore Energy Release in Unstable Failure of Brittle
pressure are of great importance in associated rock Rock-Part I. Fundamentals. Int. J. Rock Mech.
mechanics and engineering areas. It is important to & Min. Sci. 35(2): pp.113-121.
identify the main failure mechanisms associated
Tang, C.A., Tham, L.G., Liu, H.Y., et al. 2000.
with AE characteristics in compression. This
Numerical Tests on Micro-Macro Relationship
identification is crucial for a better understanding
of Rock Failure under Uniaxial Compression,
and interpretation of the experimental results, and
Part I: Effect of heterogeneity. Int. J. Rock
consequently, improves our concepts on rock
Mech. Min. Sci. 37: pp.555-569.
mechanical properties or analysis of rock
engineering structures. Tang, C.A., Lin , P., Wong, R.H.C., et al. 2001.
It is noted that the model predications of Analysis of crack coalescence in rock-like
mechanical behaviour of rocks using pore pressure materials containing three flaws—Part II:
incorporated RFPA in this paper capture most of Numerical approach. Int. J. Rock Mech. Min.
the experimental observed phenomena, especially Sci. 38: pp.925-939.
the confining pressure effect and pore pressure XU, T., Tang, C.A., Zhang, Z., et al. 2003.
effect of rock specimens in failure process.
Theoretical, experimental and numerical studies
Although the simulations are not a quantitative on deformation and failure of brittle rock in
approach and many conclusions presented here uniaxial compression. Journal of Northeastern
may have already been obtained by laboratory University 24(1): pp.87-90.
tests, the significance of mimicking these
phenomena by numerical simulation is obvious. At Weibull, W. 1951. A statistic al distribution
least, and the most important, the successful function of wide applicability. Journal of
reproducing of the experimentally observed failure applied mechanics, pp. 293-297.
phenomena with a numerical method implies that Wong, R.H.C., Lin , P., Tang C.A., et al. 2001.
our understanding to the mechanisms of rock Analysis of crack coalescence in rock-like
failure has reached a more reasonable level, which materials containing three flaws—Part
in turn will help us to make further progresses in I experimental approach Int. J. Rock Mech.
the field of rock mechanics and rock engineering. Min. Sci 38: pp.909-924.

7. ACKNOWLEDGEMENTS
The present research in this paper was carried
out with the jointly financial support of the
National Natural Science Foundation of China (No.
50134040, 50204003 and 50174013).

Vous aimerez peut-être aussi