2012.05.

24
Crack models
ROSE
RockPhysics andGeomechanics
Course2012
X|z Y}
2012.05.24
Cracks haveastrong impact on rockbehavior
x
x
E
o
c
A
=
A
( )
eff
1
x
x
E E Q
o

c
A
= =
A
The presence of cracks
reduces the stiffness
of the rock
2012.05.24
Cracks haveastrong impact on rockbehavior
x
x
E
o
c
A
=
A
( )
eff
1
x
x
E E Q
o

c
A
= =
A
Increasing stress
Crack closure
Increasing stiffness
2012.05.24
Cracks haveastrong impact on rockbehavior
Increasing stress
Crack closure
Increasing stiffness
0
5
10
15
20
25
30
35
40
45
0 0.005 0.01 0.015 0.02
M
e
a
n

s
t
r
e
s
s

[
M
P
a
]
Volumetricstrain
2012.05.24
Crack or failed grain contact?
2012.05.24
Crack or failed grain contacts?
2012.05.24
Crack or failed grain contacts?
2012.05.24
One large crack
2012.05.24
or several failed grain contacts?
2012.05.24
An assembly of failed grain contacts = a large crack
has a much stronger impact on rock stiffness
than the sum of the individual failed grain contacts
Crack density:
2a
Drainage
parameter
Isotropic distribution of cracks:
Flat cracks: = c/a 0

D 0 for saturated rock

D = 1 for dry rock
2c
Open, flat cracks
n = number of cracks per unit volume
Not consistent with Biot, but
we may use the model to predict
the properties of the dry material,
which gives us the frame moduli.
Non-isotropic distribution of cracks anisotropy
*
1
o k
ij ij ij k
k
C C Q ,
| |
=
|
\ .

k
Q
z
13
Cracks control the velocities:
Large reduction
in velocity
and amplitude
No effect (almost) on velocity
and amplitude of P-wave
Some effect on vertically
polarized S-wave
Non-isotropic distribution of cracks anisotropy
14
Cracks can explain.
*
1
o k
ij ij ij k
k
C C Q ,
| |
=
|
\ .

k
Q
z
No drainage D = 0 (nearly) no P-wave anisotropy
"Saturation eliminates P-wave anisotropy"
Leon Thomsen (1995):
Pore pressure equalization between cracks and pores
Consequence: the drainage parameter will not vanish,
regardless how thin the cracks are
"Saturation does not
eliminate P-wave anisotropy
in porous and permeable rocks"
5/24/2012
General formalismfor displacement discontinuities
(Sayers and Kachanov, 1995):
For open, penny-shapedcracks
This approximation is not valid in general,
hence the open crack model is too simple.
However, we may compensate for this by allowing the
drainage parameter D be an adjustable parameter.
0
N T ijkl
B B |
5/24/2012
Many cracks crack interactions
- the presence of one crack may
affect the influence of another
( ) ( )
*
1
s s
K K Q v , =
Self-consistent models:
Interactions are taken into account by giving the rock around
a crack the properties of the effective medium
( )
* * *
s s
K K K Q v , =
( )
*
*
1
s
s
K
K
Q v ,
=
+

5/24/2012
Alternative, equally valid procedures
( ) ( )
*
1
s s
K K Q v , =
Self-consistency:
( )
* * *
s
K K K Q v , =
( )
*
*
1
s
K
K
Q v ,
=
+

( )
*
1 1 s
s s
Q
K K K
v ,
= +
( )
*
* *
1 1
s
Q
K K K
v ,
= +
( )
( )
* *
1
s
K K Q v , =
*
0 always K > ( )
* *
0 for 1/ K Q , v = =
There are many different self-consistent models,
giving different predictions depending on the initial model.
The Differential Effective Medium (DEM) model resolves this discrepancy
by adding small numbers of cracks in many steps, and recalculating the
effective stiffness for each step (always working in the non-interacting limit)
This gives the same (unique!) solution for all initial models.
But is it more correct because of that?
( )
* * *
dK K Q d v , =
Models accounting for interactions
Many alternatives..
- all are mathematically correct, but they give different predictions.
Which one is correct?
It depends.
If we do not know
how the cracks are positioned
(and we usually don't),
the linear, non-interacting model
may be a good starting point.
The actual position of the cracks
relative to each other
determines which model is the
correct one to use.
Unfortunately, no model comes
with a description of how the
cracks are positioned.
23
Stress effects on the rock framework
Cracks control the velocities:
Large reduction
in velocity
and amplitude
No effect (almost) on velocity
and amplitude of P-wave
Some effect on vertically
polarized S-wave
24
Rule of thumb:
Velocities are mostly affected by changes
in the normal stress in the direction of propagation
(and polarization)
25
Cracks are sensitive
to changes in stress
A compressive principal stress tends to
close a crack that is oriented normal to
the stress
0
500
1000
1500
2000
2500
3000
3500
-0.004 0 0.004 0.008 0.012
Axial strain
Velocity
[m/s]
0
50
100
150
200
250
Stress
[MPa] Axial
P-wave
Radial
P-wave
Axial
S-wave
Axial
stress
Radial stress
Sliding cracks
- induce hysteresis,
permanent deformation,
and difference between
loading and unloading
modulus
27
Cracks are sensitive
to changes in stress
Shear deformations
tend to open up cracks
0
500
1000
1500
2000
2500
3000
3500
-0.004 0 0.004 0.008 0.012
Axial strain
Velocity
[m/s]
0
50
100
150
200
250
Stress
[MPa] Axial
P-wave
Radial
P-wave
Axial
S-wave
Axial
stress
Radial stress
28
Cracks are sensitive
to changes in stress
Shear deformations
tend to open up cracks
0
500
1000
1500
2000
2500
3000
3500
-0.004 0 0.004 0.008 0.012
Axial strain
Velocity
[m/s]
0
50
100
150
200
250
Stress
[MPa] Axial
P-wave
Radial
P-wave
Axial
S-wave
Axial
stress
Radial stress
29
Solid, pores & cracks
Three sets of flat cracks
oriented normal to
the principal stresses
,
x
,
y
,
z
x
o

y
o

z
o

A model
Fjr (2006):
30
Cracks are sensitive
to changes in stress
A compressive principal stress tends to
close a crack that is oriented normal to
the stress
We can not close a crack that is
already closed
i i i
d d , , o
( )
n
i i o
T , o

+
Assumption:
31
Local failures induced by compressive stress:
shear
tensile
Simulation by PFC
2D
Cracks are sensitive
to changes in stress
Shear deformations
tend to open up cracks
32
( ) ( )
i i j i k
d d d d d , c c c c
( )
2
i j k
i
e
| c c c
,

Cracks are sensitive

to changes in stress
Shear deformations
tend to open up cracks
Sensitivity to shear strain:
Assumption:
33
Very large shear strains
more turbulent crack development
Changes in crack density more sensitive to magnitude
than to orientation of shear strain
2
i
e
q
,
I

Assumption:
I = maximum shear strain
34
( )
2
2
i j k
n
o
o
i o
i i
i o
T
e
T
| c c c q o
, ,
o
+ I
| |
+
=
|
+
\ .
1
o v
v
| c
|
c

( )
11 11 11 33 11
1
o p
x y z
C C Q Q Q | , , ,
(
= +

11 22 33
o o o
o
C C C H = = =
44 55 66
o o o
o
C C C G = = =
12 13 23
2
o o o
o o
C C C H G = = =
H
o
=80 GPa(fixed value)
Mathematics of the model:
etc.
Assumptions:
11
1

p
C
V

| |
=
|
|
\ .
etc.
Velocities:
v =0.2(fixed value)
35
0
20
40
60
80
100
120
140
160
180
200
0 20 40 60 80
o
3
o
1
A
B
C
Test A
0
500
1000
1500
2000
2500
3000
3500
-0.004 0 0.004 0.008 0.012
Axial strain
Velocity
[m/s]
0
50
100
150
200
250
Stress
[MPa]
Axial
stress
Radial stress
Axial
P-wave
Radial
P-waves
Axial
S-wave
Scaling:
, , ,
o o o
x y z o
G , , ,
36
0
500
1000
1500
2000
2500
3000
3500
-0.004 0 0.004 0.008 0.012
Axial strain
Velocity
[m/s]
0
50
100
150
200
250
Stress
[MPa]
Axial
stress
Radial stress
Axial
P-wave
Radial
P-waves
Axial
S-wave
0
20
40
60
80
100
120
140
160
180
200
0 20 40 60 80
o
3
o
1
A
B
C
Test A
0.1
5
800
n
|
q
=
=
=
2012.05.24

The model based on flat cracks and sperical pores - matches

observations quite well
The match supports the claimthat the stress dependency of
wave velocities may largely be explained in terms of opening and
closure of cracks
2012.05.24
References:
Fjr, E., Holt, R.M., Horsrud, P., Raaen, A.M. and Risnes, R. (2008) "Petroleum Related
Rock Mechanics. 2
nd
Edition". Elsevier, Amsterdam
Fjr, E. (2006) "Modeling the stress dependence of elastic wave velocities in soft rocks".
41
st
USRM Symposium, paper ARMA/USRM 06-1070
Sayers, C.M., Kachanov, M. (1995) "Microcrack-induced elastic wave anisotropy of brittle
rocks", J. Geophys. Res. B, 100, 4149-4156
Thomsen, L. (1995): "Elastic anisotropy due to aligned cracks in porous rock".
Geophysical Prospecting, 43, 805-829