Microseismic Monitoring and Geomechanical Modelling of CO2 - bris
Microseismic Monitoring and Geomechanical Modelling of CO2 - bris
Microseismic Monitoring and Geomechanical Modelling of CO2 - bris
Create successful ePaper yourself
Turn your PDF publications into a flip-book with our unique Google optimized e-Paper software.
CHAPTER 3.<br />
INVERTING SHEAR-WAVE SPLITTING MEASUREMENTS FOR FRACTURE PROPERTIES<br />
Fully connected fractures<br />
An alternative approach is to consider the low frequency limit (Hudson et al., 2001). In this limit,<br />
there will be enough time for the pressure gradient to be completely equalised between fractures<br />
<strong>and</strong> pores. In this limit I treat the fractures as being drained. The widely known formulations <strong>of</strong><br />
Gassmann (1951) can then be used to compute the effects <strong>of</strong> fluid saturation on the overall system.<br />
The Gassmann equations have been generalised for anisotropic rocks by Brown <strong>and</strong> Korringa (1975),<br />
who find that the fluid saturated compliance is given by<br />
S sat<br />
ijkl = Sijkl d − (Sd ijαα − Sm ijαα )(Sd klαα − Sm klαα )<br />
( 1<br />
K<br />
) + Φ( 1<br />
m K fl<br />
− 1<br />
K<br />
) m<br />
K d − 1<br />
(3.9)<br />
where Sijkl m <strong>and</strong> Km are the compliance tensor (in uncontracted 3×3×3×3 form) <strong>and</strong> bulk modulus<br />
<strong>of</strong> the minerals making up the rock, Sijkl d <strong>and</strong> Kd are the compliance tensor <strong>and</strong> bulk modulus <strong>of</strong> the<br />
dry rock frame, <strong>and</strong> Φ is the porosity.<br />
Restricted fluid flow<br />
I have now described the two endmembers, high <strong>and</strong> low frequency, that correspond to a fully connected<br />
pore space <strong>and</strong> an isolated pore space. To model between the low <strong>and</strong> high frequency endmembers<br />
then frequency dependence must be factored into the calculations. Hudson et al. (1996) present<br />
an extension to the Hudson (1981) model which can account for flow between fractures <strong>and</strong> equant<br />
porosity. As discussed previously, fluid flow will affect the normal compliance <strong>of</strong> the fracture, so a<br />
correction is made to the Hudson (1981) term for B N , such that K is now<br />
K = K fl (λ r + 2µ r ) 1<br />
πaµ r (λ r + µ r ) 1 + (3(1 − i)J/2c)<br />
(3.10)<br />
where<br />
J 2 = K flΦκ<br />
2ηω . (3.11)<br />
κ is the permeability <strong>of</strong> the rock, η is fluid viscosity, ω is the frequency <strong>of</strong> the incident wave, <strong>and</strong> c is<br />
the average fracture aperture. Note that, because strain parallel to a fracture does not cause a volume<br />
change, the tangential compliance <strong>of</strong> the fracture is not affected by fluid flow, <strong>and</strong> does not need to<br />
be modified.<br />
K can now be considered in terms <strong>of</strong> two parameters, a fluid incompressibility factor P i <strong>and</strong> an<br />
equant porosity factor P ep (Pointer et al., 2000), such that<br />
where<br />
K =<br />
(<br />
Pi<br />
π<br />
λ r + 2µ r ) ( ) −1<br />
3(1 − i)<br />
λ r + µ r 1 +<br />
2 √ (3.12)<br />
P ep<br />
P i = 1 K fl<br />
a µ r ,<br />
P ep = ( c J )2 = 2ωη fl<br />
ΦK fl κ c2 . (3.13)<br />
32