�ݺ�ߣ

Avizo EUGM 2012 – 30 mai au 1er juin 2012 - Bordeaux

Studying thermomechanical properties of
thermostructural composites using X-Ray
computed micro-tomography

O.Caty1,
E.Rohmer1, C.Lorrette1,2, G.L.Vignoles1, G.Couégnat1, M.Charron1

- University of Bordeaux -

1 LCTS Laboratoire des Composites ThermoStructuraux

2 CEA centre de Saclay – Commissariat à l’Energie Atomique et aux
Energies Alternatives

Outline
• Motivation

• C/C composites
– 3D images
– Thermomechanical models

• Tubular shaped SiC/SiC
– Image analysis
– Mechanical models

O.CATY et al. - μCT / Thermostructural composites – Avizo2012 2/20
caty@lcts.u-bordeaux1.fr

Motivation
• Material : C or SiC composites
+ High temperature
+ Low CTE coeff. of thermal exp.
+ Mechanical properties
• Applications (highly demanding conditions)
+ Nuclear power plants (fuel confinement – Tokamak plasma-facing components)
+ Aerospace technology (ArianeV boosters, Jet engine divergent, Atmospheric reentry
heat shield parts, Aircraft brakes…)

• Predict thermo mechanical properties
– Materials are highly structured (woven materials)
– Behavior is linked to the 3D geometry
Objective : predict the material behavior


Developed method(s)
• Scan the material
– X-ray μCT
– Visualisation in 3D (Avizo std.)

• Work on images
– Pores and cracks : determine density,
shape and position
– Matrix and fibers : determine orientation,
density, shape and pattern of the weaving
– Mesh the image

• Calculate the thermomechanical
behavior
– Using the mesh
– With elements enriched (density,
orientation… calculated on μCT images).


C/C composites
μCT images :
- Phoenix XRay – Nanotom

X-Z 3D

- Laboratory Micro/nano CT
system
- Resolution : 5 μm

- Woven plies (X-Y directions)
- Stacked and stitched in Z direction
- > Complex geometry : X-Y Y-Z
- Tangled phases (yarn, matrix, pores) hard to label

C/C composites
Solution :
- Use directly the voxel as the structural element (Finite element for calculation)
- Enrich the elements by the direction of fibers and the density (both measured
on the image)

…FE density…

μCT 3D image… … is meshed using
…calculation of
the voxel as FE…
thermal dilation
- Avizo for image visualization (3D image,
density, orientations)
… and orientations enrich the model…
- ImageJ for data translation caty@lcts.u-bordeaux1.fr
- Zebulon or Coda for FE calculation

C/C composites
Measure the density

Relative
Porosity & material density
Porosity
Material
Materials
properties


C/C composites
Measure the fiber directions :
- Image structure tensor
- Image Hessian matrix

 2I 2I 2 I  - Where I is the Grey level of the pixel
 2 
 x xy xz  - x, y and z are the image direction
 2I 2I 2 I  - Derivatives calculated by centred finite
He    difference method – ie :
 yx y 2 yz 
 2 I ( x, y )
 2I 2I 2 I   I ( x  1, y )  2  I ( x, y )  I ( x  1, y )
x 2

 zx 
2 
zy z  - Mask applied


- Calculate eigenvalues and eigenvectors of the Hessian matrix
- The eigenvector associated with the minimum egeinvalue
- > represents the minimum grey level variation
- > material (fibers) direction.

C/C composites
Measurement of the fiber directions :

 

y
x



y X orientation

z
Y orientation

 Z orientation
x

z


C/C composites
Calculation - T=1500°C:

Mesh

Density

- Calculated global thermal dilation : εyy= 0.7%
Directions
- Visualization of the effect on the structure : local variations & stresses

C/C composites
Conclusion
- Complex architecture -> use voxel as finite element
- Enrich elements with measures on images (direction and density)
- Avizo very useful for 3D visualization (grey scale images, combination of
orientation field images)

Next :
tubular shaped SiC/SiC
- More easy to label
- Unstructured meshes based on image segmentation (reconstruction or
direct mesh)


Tubular shaped SiC/SiC

- Laboratory Micro/nano CT
system
- Resolution : 10 μm

- Braided tubular fabrics X-Y 3D - int 3D - ext
- 2 stacks
- Dry and infiltrated materials available
X-Z - Phases and weaving relatively easy to analyze

1st Solution :
- Polar to Cartesian transformation
- Analyze the weaving pattern by locating and marking the path
- Encode the path to reconstruct a finite element model


1st Solution :
- Polar to Cartesian transformation Color = orientation

- Analyze the weaving pattern by locating and marking the path
- Encode the path to reconstruct a finite element model


2nd
solution :
- Mesh the image directly with Avizo (unstructured tetrahedrons)
- Enrich the elements with measurements on images (density & orientations)

…orientations…

μCT 3D image… … is meshed using …calculation of
Avizo… compressive
behavior
… and density enrich the model…

Meshing the image with Avizo :
…A marching cubes
μCT 3D image of …is labeled… …the label is dilated… algorithm is applied…
a dry tube … (To extract the SiC Yarns) (To simulate the matrix)
(To mesh the surface)

…apply an advancing …simplification (nbr of …simplification (max
front algorithm… triangles criterion)… distance criterion)…
(Tetrahedra) (triangles adapted to curvature) (Homogeneous triangles)


Affect local informations :

Determination
of center of
gravity of
tetrahedra Calculation
(compressive BC)
Reading orientations &
density in voxels
(same methods as in CC composites)


Results :

Initial μCT

Calculated
deformation

Measured by
digital image
correlation
εzz
Ecalculated= Emeasured=
O.CATY et al. - μCT / Thermostructural composites – Avizo2012
caty@lcts.u-bordeaux1.fr 180 GPa 177 18/20
GPa

Conclusions
= Less complex architecture -> 2 solutions

- 1st solution : analyze & encode the weaving path to reconstruct a
finite element model
- Weaving hard to analyze, to reconstruct and to calculate. But
accurate method.

- 2nd solution : mesh the image with Avizo & enrich the elements with
measurements on images (density & orientations)
- Thanks to Avizo package, mesh is easy to produce.
- Calculation very efficient and accurate (close to experimental
results)

Conclusions/perspectives
• Calculations based on the real material

• Very accurate methods – halfway methods between
experimental and modeling

• Useful to understand and predict thermostructural
composites behavio

• Importance of the knowledge of micro-scale mechanical
properties

• Avizo® mostly used for 3D observation (density,
geometry, orientations) and meshing.


Acknowledgements
• Industrial partners :
– Safran Herakles (SPS) : A.Dekeyrel,
B.Delperier, B.Reignier, C.Descamps
– CEA Saclay : C.Sauder

• Students :
– T.Agulhon, M.Genin, N.Siefert


Studying thermomechanical properties of thermostructural
composites using X-Ray micro-computed tomography
O.Caty,
E.Rohmer, C.Lorrette, G.L.Vignoles, G.Couégnat, M.Charron

- University of Bordeaux -

Thank you for your attention

Questions ??
- Laboratoire des Composites ThermoStructuraux -


�ݺ�ߣ

Studying thermomechanical properties of thermostructural composites using X-ray microcomputed tomography

Recommended

More Related Content

More from VSG - Visualization Sciences Group (13)

Studying thermomechanical properties of thermostructural composites using X-ray microcomputed tomography