COFIN 2002 Torino 28112002

1
Combining X-ray CT and 3D Digital Image
Correlation for studying localized deformation in
clay rocks under triaxial compression
Cino Viggiani, Steve Hall, Nicolas Lenoir, Pierre
Bésuelle, Jacques Desrues Laboratoire
3S-R (Sols, Solides, Structures - Risques)
Grenoble, France

• background and motivations
• X-ray CT at ESRF
• combining X-ray CT with 3D-DIC
• - results on Callovo-Oxfordian Argillite
• - results on Beaucaire Marl
• partial answers, open questions

2
motivation
1
radioactive waste disposal in deep underground
galleries
URL in Bure (France) , depth 490m
3
motivation
EDZ around radioactive waste underground storage
galleries
DAMAGE around deep excavations in clays IS
LOCALIZED !
4
Opalinus Clay and Boom Clay tested in triaxial
compression
5
X-ray CT and strain localization a long story
3rd IWB (Grenoble - Aussois 1993)
6
X-ray CT and strain localization a long story
6th IWB (Minneapolis - St. Paul 2002)
2D strain fields (FRS and DIC) for plane strain
tests on clay
7
X-ray CT and strain localization a long story
why dont we go 3D for clay as well? non
destructive 3D imaging techniques ? X-ray
tomography
8
micro X-ray CT at ESRF a significant step forward
7th IWB (Chania 2005)
9
micro X-ray CT at ESRF a significant step forward
10
micro X-ray CT at ESRF
basic principle

• recording attenuation profiles through a
  specimen slice, under different angular
  positions
• reconstructing a radiograph of the slice
• repeating to get a complete set of slices over
  the
• specimen
• reconstructing a 3D image of the internal
• structure of the specimen from the spatial
• distribution of the linear attenuation
  coefficient

X-ray characteristics

• X-ray white beam to have a high photon flux
• X-ray energy 50 to 70 keV
• spatial resolution 14 µm (voxel size)
• time for scanning 12 to 15 minutes

11
experimental setup
X-ray micro tomography scans are performed
throughout a triaxial compression test
Frelon camera
loading system
X-ray source
mirror
X-ray imaging system
x-ray imaging system
positioning system

• displacement controlled loading (3 µm/min)
• displacement is stopped during scanning

12
experimental setup
14
13
material studied
1
Callovo-Oxfordian argillite from the Borehole
EST261 (depth 476m) of the Underground Research
Laboratory of Bure (ANDRA)
a few characteristics Clay content 40 Water
content 6 Permeability 10-20 to 10-22 m²
14
High confining pressure test (10 MPa)
X-ray CT images from a test on COX argillite
horizontal slices
stress deviator
axial strain
10 mm
20 mm
after removal of confinement
Results
15
X-ray CT images from a test on COX argillite
vertical slices
stress deviator
axial strain
after removal of confinement
Results
16
X-ray CT provides 3D images
17
X-ray images do not show everything
natural inclusions in yellow open cracks in red
18
can we obtain more info from these images?
micro tomography improved resolution if
density field changes!
what if density does not change ? mode II cracks,
shear bands with no volume changes
19
basic principles of 3D DIC
search in 3D for best correlation displacement
vector (integer - pixel)
two 3D images of specimen at different
loading/deformation levels
sub-pixel refinement
definition of nodes distributed in the first
image definition of the motif, or correlation
window (region about each node) calculation of a
correlation coefficient for each displacement of
this motif, within a region (search window)
around the target node in the second
image definition of the discrete displacement
(integer number of pixels) i.e. that with the
best correlation sub-pixel refinement (because
the displacements are rarely integers of pixels),
which may also involve more complex
transformations than simply rigidbody translation
calculation of strain from the displacements
interpolating correlation coefficient fast,
but can only assess rigid body motion interpolatin
g gray-level slow, but more general
transformations
displacement field with sub-pixel accuracy dx,
dy, dz)
20
3D DIC results for COX argillite
undrained compression, p0 10 MPa
before stress peak
after stress peak
3D DIC using CorrelManu3D 86247 nodes correlation
window size 20 voxels
Lenoir et al. (2007) Strain, International
Journal for Experimental Mechanics, Vol. 43, No.
3, 193205
21
increment before stress peak
Element size 280 µm
Lenoir et al. (2007) Strain, International
Journal for Experimental Mechanics, Vol. 43, No.
3, 193205
22
increment after stress peak
Lenoir et al. (2007) Strain, International
Journal for Experimental Mechanics, Vol. 43, No.
3, 193205
23
3D DIC results for Beaucaire Marl
drained compression, p0 150 kPa
3D DIC using TomoWarp spacing between nodes 10
voxels correlation window size 20 voxels time
for analysis shown (1/4 sample) 5 hours
(vertical slice through 3D image)
24
3D DIC increment before stress peak
25
3D DIC increment at stress peak
26
3D DIC well after stress peak
27
3D DIC upon removal of confining pressure
28
a few conclusions plus one question

• in-situ micro tomography for triaxial testing
  (now also at our lab)
• demonstration of 3D-volumetric DIC for a variety
  of geomaterials
• (clay rock, clay) ? sand
  (following talk)
• clear evidence of strain localization before
  peak from DIC

tremendous possibilities now available with
imaging technology full-field measurement
techniques have opened up new avenues for
research in geomechanics, and will continue to do
so especially when different methods are used in
conjunction, e.g., combination of DIC and X-ray
tomography