Through the µct and what we found there?

(1)

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

(2)

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

• Voids characterisation

Textiles and random fibre assemblies

3D woven glass fibre textile random steel fibre assembly

(3)

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

(4)

Textile deformation during draping

11/04/2019 Dagstuhl

S.V. Lomov [Pazmino et al 2014] 7

Shear of a woven fabric

(5)

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

2

0

1



 



orientation tensor 𝑎_𝑖𝑗= ර 𝑝_𝑖𝑝_𝑗𝜓(𝒑) 𝑑𝒑

(6)

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

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).

(7)

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

(8)

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:

(9)

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

(10)

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

(11)

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.

(12)

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?

(13)

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

(14)

Tortuosity of a woven laminate

carbon fibre twill 2/2

 In plane 0° (Warp):

 In Plane 90° (Weft):

 Through plane:

𝐾

_𝐾𝑇

= φ𝑑

_{𝑐ℎ𝑎𝑛𝑛𝑒𝑙}²

226τ

2 x 2 Twill carbon fabric

D^eff (m²/s)

τ (L/L₀)

d_channel (µm)

K_KT (Darcy)

K_NS (Darcy)

K_Exp (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

• Permeability calculations

4.Conclusions

(15)

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

(16)

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