www.composites-kuleuven.be
Through the µCT and what we found there?
Quantifying images of fibrous materials
Stepan V. Lomov,
Jeroen Soete, Martine Wevers
Department of Materials Engineering, KU Leuven, Belgium
Through the µCT and what we found there?
1. Introduction. Fibrous media:
microstructural features and characteristics
2. Quantification of a µCT image of fibrous materials
3. Applications
• Fibre orientations characterisation
• Finite element models of deformation and damage
• Permeability calculations 4. Conclusions
11/04/2019 Dagstuhl
S.V. Lomov 3
1.Introduction. Fibrous media: microstructural features and characteristics
2. Quantification of a µCT image of fibrous materials 3. Applications
• Fibre orientations characterisation
• Voids characterisation
• Finite element models of deformation and damage
• Permeability calculations 4. Conclusions
Textiles and random fibre assemblies
3D woven glass fibre textile random steel fibre assembly
Textiles: Dual scale
11/04/2019 Dagstuhl
S.V. Lomov 5
bundle &
pore size
~100 µm
fibre & pore size ~10 µm
“An intellect which at a certain moment would know all forces that set nature in motion, and all positions of all items of which nature is composed.
If this intellect were also vast enough to submit these data to analysis, it would embrace in a single formula the movements of the greatest bodies of the universe and those of the tiniest atom.
For such an intellect nothing would be uncertain and the future just like the past would be present before its eyes.”
Laplace
Textiles: Regularity and variability
repetetive cells …
... variable inside
a fabric with very stable geometry ...
... and variable tow contours
Textile deformation during draping
11/04/2019 Dagstuhl
S.V. Lomov [Pazmino et al 2014] 7
Shear of a woven fabric
Textiles: Tow contours and fibrous geometry
11/04/2019 Dagstuhl
S.V. Lomov 9
“blurry” tow boundaries and variable VF within tows
Random fibres: Orientation distribution
dp d
d
0
Orientation distribution function (ODF):
, :
/ 2 / 2
cos / 2 cos cos / 2
, sin
Symmetry:
or , ,
Normalisation:
0 2
0 1
, sin
d d
P d d
d d
P
d d d
p
p p
p p
2
0
1
orientation tensor 𝑎𝑖𝑗= ර 𝑝𝑖𝑝𝑗𝜓(𝒑) 𝑑𝒑
11/04/2019 Dagstuhl
S.V. Lomov 11
1. Introduction. Fibrous media: microstructural features and characteristics
2.Quantification of a µCT image of fibrous materials
3. Applications
• Fibre orientations characterization
• Voids characterisation
• Finite element models of deformation and damage
• Permeability calculations 4. Conclusions
Straumit, I., S.V. Lomov and M. Wevers, Quantification of the internal structure and automatic generation of voxel models of textile composites from X-ray computed tomography data. Composites Part A, 2015. 69: p. 150-158.
µCT image voxel model
~100 mio pixels ~500,000 voxels
• material type (matrix, reinforcement, void);
• fibre orientation vector (for the reinforcement only);
• local fibre volume fraction (for the reinforcement only).
Structure tensor and anisotropy
Degree of anisotropy
11/04/2019 Dagstuhl
S.V. Lomov 13
Estimation of precision based on anisotropy
-1.5 -1 -0.5 0 0.5 1 1.5 2
-1 -0.5 0 0.5 1 1.5
Log10(scatter, degrees)
Anisotropy
wr=6 wr=8 wr=10 wr=12 wr=14
Material
Fibre diameter,
μm
Voxel size,
μm Anisotropy Scatter, degrees
Carbon 7 2.34 0.61 1.1
Flax 15…20 1.48 0.95 0.13
Steel 25 4.4 0.91 0.19
Glass 12 5.5 0.89 0.23
Recommended voxel size parameters
Segmentation
The variables, extracted from the image, reflect physical properties of the material:
• grey value – material attenuation
• structural anisotropy – material microstructure type
Methods of segmentation:
• Unsupervised (k-means)
• Supervised (Gaussian mixture model)
Matrix
Yarns
Voids
11/04/2019 Dagstuhl
S.V. Lomov 15
Supervised segmentation
Uses a training set: selected regions in the image with the class known a priori
Components defined as Gaussian distributions:
Classification criterion:
VoxTex software
Nanotom (3D volume)
ABAQUS
FlowTex
TexComp Finite element analysis
Permeability
Homogenization (Mori- Tanaka)
ParaView
Visualization SkyScan,
Tomohawk (image stack)
Root (CERN) Histograms, statistics
Simcenter
11/04/2019 Dagstuhl
S.V. Lomov 17
1. Introduction. Fibrous media: microstructural features and characteristics 2. Quantification of a µCT image of fibrous materials
3.Applications
• Fibre orientations characterization
• Voids characterisation
• Finite element models of deformation and damage
• Permeability calculations
4. Conclusions
Dry Tape Laying precision
(-45/0/45/-20/0/20/90/0)so Produced at NLR, Coriolis machine Hexcel HiTape UD126 - AS7/V800E/QD2/6.35mm
orientations output from the mid-plane of a ply
11/04/2019 Dagstuhl
S.V. Lomov Nguyen, N., M. Mehdikhani, I. Straumit, L. Gorbatikh, L. Lessard, and S.V. Lomov, Composites Part A, 2018. 104: 14-23 19
Randomly oriented chopped carbon strands
formation of similarly oriented layers
local morphology:
intact strands;
splitting;
micro-voids
Random short fibre composites
11/04/2019 Dagstuhl
S.V. Lomov 21
1.4 μm/pix 3.2 μm/pix
8 μm/pix 16 μm/pix
Glass fibre/PP composite, VF ~20%, d = 15 µm Ellipsometry @ 1.4 µm/pix vs structure tensor @ 16 µm/pix
VoxTex, different image resolutions
3.2 μm/pix 8 μm/pix 16 μm/pix
Karamov R., S.V. Lomov et al EUROMECH-602 Composites manufacturing, Lyon, 2019
Characterisation of voids
• carbon/epoxy laminates
• void content intentionally increased (low pressure in autoclave)
• fibre diameter 7 µm
• µCT: HECTOR @ Ghent University
• pixel size 6.55 µm [±45]2s, [±67.5]2s, [67.5/22.5]2s, [0/90]2s, and [0/90]4s.
VoxTex finite element voxel models
Geometry corrector Geometric modeller
Meshing Assign material
properties
Boundary conditions FE solver, postprocessor Homogenisation
Damage analysis
VoxTex creates a voxel model;
after segmentation: identification of materials per voxel
HEX elements = voxels
fibre direction identification (structure tensor) and Chamis UD homogenisation
periodic boundary conditions NB: “weak” periodicity
no need for the geometrical corrections
Abaqus Simcenter
11/04/2019 Dagstuhl
S.V. Lomov 23
Impregnated fibre bundle test
Quasi-UD flax:
22.9±0.5 GPa -> flax 53.2 GPa UD flax:
26.6±2.3 GPa -> flax 62.4 GPa VF = 40%
How to obtain correct flax fibre modulus from quasi- UD tests?
Damage in 3D woven glass/epoxy composites
11/04/2019 Dagstuhl
S.V. Lomov Liu, Y., I. Straumit, D. Vasiukov, S.V. Lomov, and S. Panier, Composite Structures, 2017. 179: 568-579 25
Permeability calculations
0.0E+00 5.0E-05 1.0E-04 1.5E-04 2.0E-04 2.5E-04 3.0E-04 3.5E-04 4.0E-04
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16
Permeability, mm2
Unit cell No
Ky Kx
Results for 16 unit cells, resolution of the image 6 μm
EXPERIMENT
Liquid resin flow lines in a unit cell model FlowTex
K
carbon NCF (Saertex) [+45°/-45°]
tricot stitching, 540 g/m2 measured permeability:
VF = 50.8% -> K = 0.5·10-4mm2 laminate RTM’ed,
thickness 4.0 mm, VF = 45.5%
estimated permeability:
VF = 45.5% -> K = (1…2)·10-4mm2
Tortuosity of a woven laminate
carbon fibre twill 2/2
In plane 0° (Warp):
In Plane 90° (Weft):
Through plane:
𝐾
𝐾𝑇= φ𝑑
𝑐ℎ𝑎𝑛𝑛𝑒𝑙2226τ
2 x 2 Twill carbon fabric
Deff (m²/s)
τ (L/L0)
dchannel (µm)
KKT (Darcy)
KNS (Darcy)
KExp (Darcy)
In plane (0°) 0.0452 1.91 300 34.7 30.7 82.1
In plane (90°) 0.0755 1.47 450 101 110.8 131.7
Through plane 0.0024 8.35 100 0.9 0.4 /
11/04/2019 Dagstuhl
S.V. Lomov Soete, J., S.V. Lomov, L. Gorbatikh, and M. Wevers, FPCM-14) Lulea, Sweden. 2018. 27
1. Introduction. Fibrous media: microstructural features and characteristics 2. Quantification of a µCT image of fibrous materials
3. Applications
• Fibre orientations characterisation
• Finite element models of deformation and damage
• Permeability calculations
4.Conclusions
Conclusions
1. Micro-computed X-Ray tomography is a state-of-the-art tool for detailed investigation of internal structure of fibrous materials.
2. Lab-scale µCT allows quantification of the fibrous structure and a seamless transfer of data to mechanical modelling software, allowing calculation of:
• homogenised stiffness
• damage development and strength
• permeability
• effect of defects
11/04/2019 Dagstuhl S.V. Lomov
VoxTex licences:
industry 4 (one trial) university 6
29
Proposal for a benchmark exercise
µCT based identification of reinforcement architecture of fibre reinforced polymers
1. Textile composites 2. Laminates
3. Random fibre composites
Output parameters 1. Segmentation:
• Meso-level: yarns/matrix/voids
• Micro-level: fibres 2. Local identification:
• Fibre volume fraction
• Fibre orientation 3. Voids
• overall content
• size, morphology Organisation:
1. Starting point:
• Same specimens OR
• Same images for everybody
• Data in a public depository 2. Core partners: two-three
3. Open for all; no rules/norms in the first edition
4. Time scope: 12 ... 18 months 5. Result:
• a joint paper
• participants can publish their own results
6. Final aim: best practice in µCT imaging
Pixel size:
• (sub) micron
• (sub) fibre diameter
• X fibre diameters Image size:
5*5*5 mm Hardware:
• table-top
• synchrotron
Acknowledgments
11/04/2019 Dagstuhl S.V. Lomov
The work leading to this publication has been partially funded by: the ICON project “M3Strength” (n° 130546), which fits in the MacroModelMat (M3) research program funded by SIM (Strategic Initiative Materials in Flanders) and VLAIO (Flanders), ICON project “T4” in NANOFORCE program of SIM/VLAIO, FWO-Vlaanderen (G.0354.10, G.0898.14, SBO S007219N) and VLAIO (Innovation Mandate HBC.2017.0189), KU Leuven Research Council C2 project ‘Novel methods and tools for in situ characterization of materials and tissues, and their microstructural changes during functional testing using X-ray computed tomography’, QUICOM FP7, CANAL FP7 and JOIN’EM H2020 (Grant Agreement No. 677660) projects of EU, Toyota Motor Europe NV/SA (Belgium), the Vietnamese Ministry of Education and Training, China Scholarship Council, Japan Society for the Promotion of Science Research Fellow Grant 15J09248. S.V. Lomov is Toray Professor (Toray Chair for Composite Materials, KU Leuven). The X-ray computed tomography images have been made on the X-ray computed tomography facilities at the KU Leuven, financed by the Hercules Foundation (project AKUL 09/001).
31