HR: 0800h
AN: MR11A-0931    [Abstracts]
TI: Modelling the physical properties of cracked rocks using fracture mechanics and statistical physics
AU: Benson, P
EM: p.benson@ucl.ac.uk
AF: Mineral Ice and Rock Physics laboratory, University College London, London, WC1E 6BT United Kingdom
AU: * Schubnel, A
EM: alexandre.schubnel@utoronto.ca
AF: Lassonde Institute, University of Toronto, 170 College street, Toronto, On M5S 3E3 Canada
AU: Vinciguerra, S
EM: vinciguerra@ov.ingv.it
AF: Osservatorio Vesuviano - INGV, Istituto Nazionale di Geofisica e Vulcanologia, Naples, 80124 Italy
AU: Hazzard, J
EM: hazzard@rocscience.com
AF: Rocscience Inc., 439 University Ave. #780 , Toronto, On M5G 1Y8 Canada
AU: Young, R
EM: paul.young@utoronto.ca
AF: Lassonde Institute, University of Toronto, 170 College street, Toronto, On M5S 3E3 Canada
AU: Meredith, P
EM: p.meredith@ucl.ac.uk
AF: Mineral Ice and Rock Physics laboratory, University College London, London, WC1E 6BT United Kingdom
AB: Cracks play a major role in most rocks submitted to crustal conditions. Mechanically, cracks make the rock much more compliant. They also make it much easier for fluid to flow through any rock body. Relying on Fracture Mechanics and Statistical Physics, we introduce a few key concepts which allow to understand and quantify how cracks do modify both the elastic and transport properties of rocks. The main different schemes which can be used to derive the elastic effective moduli of a rock are presented. It is shown from experimental results that an excellent approximation is the so called non-interactive scheme (Kachanov [1994]). The main consequences of the existence of cracks on the elastic waves is the development of elastic anisotropy due to the anisotropic distribution of crack orientations and the dispersion effect due to microscopic local fluid flow. Experimental data and model fit on very different rock types (basalt, granite and marbles) show both of these behaviors. We perform a simple least square fit inversion of our data in order to recover the common evolution of the crack density and aspect ratio with stress. The agreement between data and predicted velocities is in general very good, with average error between model and data points lower than 0.1km/sec, demonstrating that the inversion was very stable as a direct consequence of the well constrained laboratory data. At a larger scale, macroscopic fluid flow takes place through the crack network above the percolation threshold. Two macroscopic fluid flow regimes can be distinguished: the percolative regime close to the percolation threshold and the connected regime well above it. Using Statistical Physics and permeability models based on Gu\'eguen and Dienes [1989], we also show how to successfully predict the evolution of permeability, again for comparison to well constrained laboratory results. These results clearly demonstrate the importance of understanding the details of specific rock physical properties, and how they change in response to pressure and temperature in interpreting data from field-scale.
UR: http://www.lassondeinstitute.utoronto.ca/young/people/alex2.htm
DE: 5102 Acoustic properties
DE: 5104 Fracture and flow
DE: 5114 Permeability and porosity
DE: 5139 Transport properties
DE: 5144 Wave attenuation
SC: Mineral and Rock Physics [MR]
MN: 2004 AGU Fall Meeting