X-ray micro-CT based 3D structure Semester project Final - PowerPoint PPT Presentation

Modelling the deformation of an X-ray micro-CT based 3D structure Semester project – Final presentation Olivier Schöpfer – EPFL – GC MA3

Project objective • From an existing sample • to a numerical model • to use with the software “ Akantu ” from the EPFL LSMS lab [1] • in order to perform numerical testing of the sample 14.01.2019 Olivier Schöpfer - EPFL - GC MA3 2

X-ray micro-CT : What is it? • CT : Computed tomography • Before the X-ray CT, the slices had to be taken manually and examined by hand, one by one • Very long • Sample destroyed • Produces images similar to an X-ray, but in 3D • Non destructive • High resolution • Fast + no sample preparation needed • Automated 14.01.2019 Olivier Schöpfer - EPFL - GC MA3 3

X-ray micro-CT : How does it work? • Procedure: [3] [2] 2D 3D 14.01.2019 Olivier Schöpfer - EPFL - GC MA3 4

X-ray micro-CT : How does it work? • The receptor is there to read the intensity of the X-ray in comparison to the intensity that was emitted • The attenuation of the X-ray signal will give an indication on the material properties • Final intensity : 𝐽 = 𝐽 0 ∙ exp −𝜈 ∙ 𝑦 𝜈 ∶ attenuatiuon coefficient 𝑦 ∶ length of the x − ray path through the sample • The attenuation coefficient depends on material properties 14.01.2019 Olivier Schöpfer - EPFL - GC MA3 5

X-ray micro-CT : How does it work? • Need for a good calibration [5] Low energy High energy [4] 14.01.2019 Olivier Schöpfer - EPFL - GC MA3 6

Procedure to get a 3D model 14.01.2019 Olivier Schöpfer - EPFL - GC MA3 7

From a stack of slices to a 3D model • What comes out of the X-ray CT scan • What the 3D model looks like at first 14.01.2019 Olivier Schöpfer - EPFL - GC MA3 8

From a stack of slices to a 3D model • What comes out of the X-ray CT scan • What the 3D model looks like at first 14.01.2019 Olivier Schöpfer - EPFL - GC MA3 9

From a stack of slices to a 3D model • Binarize the slices 14.01.2019 Olivier Schöpfer - EPFL - GC MA3 10

From a stack of slices to a 3D model • Binarize the slices 14.01.2019 Olivier Schöpfer - EPFL - GC MA3 11

From a stack of slices to a 3D model • Remove Small Spots 14.01.2019 Olivier Schöpfer - EPFL - GC MA3 12

From a stack of slices to a 3D model • Remove Small Spots 14.01.2019 Olivier Schöpfer - EPFL - GC MA3 13

From a stack of slices to a 3D model • Possible to extract a subvolume to have lower computation requirements 14.01.2019 Olivier Schöpfer - EPFL - GC MA3 14

From a stack of slices to a 3D model • Generate a surface 14.01.2019 Olivier Schöpfer - EPFL - GC MA3 15

From a stack of slices to a 3D model • Generate a surface 14.01.2019 Olivier Schöpfer - EPFL - GC MA3 16

From a stack of slices to a 3D model • Generate a surface 14.01.2019 Olivier Schöpfer - EPFL - GC MA3 17

From a stack of slices to a 3D model • Generate a surface - Constrained Smoothing 14.01.2019 Olivier Schöpfer - EPFL - GC MA3 18

From a stack of slices to a 3D model • Generate a surface - Constrained Smoothing 14.01.2019 Olivier Schöpfer - EPFL - GC MA3 19

From a stack of slices to a 3D model • Generate a surface - Constrained Smoothing: Problem [6] Smoothing set to 9 Smoothing set to 1 14.01.2019 Olivier Schöpfer - EPFL - GC MA3 20

From a stack of slices to a 3D model • Simplify the modelization 14.01.2019 Olivier Schöpfer - EPFL - GC MA3 21

From a stack of slices to a 3D model • Simplify the modelization 30% 14.01.2019 Olivier Schöpfer - EPFL - GC MA3 22

From a stack of slices to a 3D model • Test the triangles before going to a tetrahedron model 14.01.2019 Olivier Schöpfer - EPFL - GC MA3 23

From a stack of slices to a 3D model • Intersection test 14.01.2019 Olivier Schöpfer - EPFL - GC MA3 24

From a stack of slices to a 3D model • Intersection test 14.01.2019 Olivier Schöpfer - EPFL - GC MA3 25

From a stack of slices to a 3D model • Intersection test 14.01.2019 Olivier Schöpfer - EPFL - GC MA3 26

From a stack of slices to a 3D model • Intersection test 14.01.2019 Olivier Schöpfer - EPFL - GC MA3 27

From a stack of slices to a 3D model • Intersection test 14.01.2019 Olivier Schöpfer - EPFL - GC MA3 28

From a stack of slices to a 3D model • Aspect Ratio test Aspect ratio = 110’053 14.01.2019 Olivier Schöpfer - EPFL - GC MA3 29

From a stack of slices to a 3D model • Aspect Ratio test Aspect ratio = 110’053 Aspect ratio = 50 14.01.2019 Olivier Schöpfer - EPFL - GC MA3 30

From a stack of slices to a 3D model • Aspect Ratio test Aspect ratio = 110’053 Aspect ratio = 15 Aspect ratio = 50 14.01.2019 Olivier Schöpfer - EPFL - GC MA3 31

From a stack of slices to a 3D model • Aspect ratio – autofix : “Prepare generate tetra grid” • Will fix most of the aspect ratio errors • Remaining errors have to be corrected manually 14.01.2019 Olivier Schöpfer - EPFL - GC MA3 32

From a stack of slices to a 3D model • Aspect ratio – autofix : “Prepare generate tetra grid” • Will fix most of the aspect ratio errors • Remaining errors have to be corrected manually 14.01.2019 Olivier Schöpfer - EPFL - GC MA3 33

From a stack of slices to a 3D model • Aspect ratio – autofix : “Prepare generate tetra grid” • Will fix most of the aspect ratio errors • Remaining errors have to be corrected manually 14.01.2019 Olivier Schöpfer - EPFL - GC MA3 34

From a stack of slices to a 3D model • Orientation test 14.01.2019 Olivier Schöpfer - EPFL - GC MA3 35

From a stack of slices to a 3D model • Orientation test 14.01.2019 Olivier Schöpfer - EPFL - GC MA3 36

From a stack of slices to a 3D model • Remesh the surface before creating the tetrahedron model to have a simpler model and better triangles 14.01.2019 Olivier Schöpfer - EPFL - GC MA3 37

From a stack of slices to a 3D model • Remesh the surface before creating the tetrahedron model to have a simpler model and better triangles 14.01.2019 Olivier Schöpfer - EPFL - GC MA3 38

From a stack of slices to a 3D model • Remesh the surface before creating the tetrahedron model to have a simpler model and better triangles 14.01.2019 Olivier Schöpfer - EPFL - GC MA3 39

From a stack of slices to a 3D model • Remesh the surface before creating the tetrahedron model to have a simpler model and better triangles 14.01.2019 Olivier Schöpfer - EPFL - GC MA3 40

From a stack of slices to a 3D model • Generate Tetra Grid 14.01.2019 Olivier Schöpfer - EPFL - GC MA3 41

From a stack of slices to a 3D model • Generate Tetra Grid 14.01.2019 Olivier Schöpfer - EPFL - GC MA3 42

From a stack of slices to a 3D model • Export for Akantu 14.01.2019 Olivier Schöpfer - EPFL - GC MA3 43

From a stack of slices to a 3D model • Size of the model Units Editor Data Parameter Editor 14.01.2019 Olivier Schöpfer - EPFL - GC MA3 44

Traction test 14.01.2019 Olivier Schöpfer - EPFL - GC MA3 45

Bone material • Variable properties depending on : • Species (Human, Cow, Rat, etc …) • Age (Young, Old) • Health (Bone disease, Calcium levels, etc …) • Bone (Femur, Tibia, etc …) • Part of the bone (Cortical, Trabecular) [7] • “Freshness” of the sample and is it conserved wet? 14.01.2019 Olivier Schöpfer - EPFL - GC MA3 46

Bone material – Mechanical properties, example • Typical intervals for human trabecular bone • Young modulus : can vary from 10 to 3000 [MPa] • Ultimate strain : typically between 1.0 and 2.5% • Poisson’s Ratio : between 0.03 and 0.6 [8] • We should probably use a visco elastic constitutive law (Keaveny, Morgan and Yeh, 2004, p. 8.15) 14.01.2019 Olivier Schöpfer - EPFL - GC MA3 47

Bone material – Mechanical properties – Young Modulus • Traction test on a bone sample 14.01.2019 Olivier Schöpfer - EPFL - GC MA3 48

Bone material – Mechanical properties – Young Modulus • Traction test on a bone sample 𝐹 𝑓𝑚𝑏𝑡𝑢𝑗𝑑 = 3060 𝑁𝑄𝑏 14.01.2019 Olivier Schöpfer - EPFL - GC MA3 49

Bone material – Mechanical properties – Young Modulus • How to get the effective modulus 𝑐𝑝𝑜𝑓 𝑜 σ 𝑗 bone + 𝑤𝑝𝑗𝑒 𝑗 𝑐𝑝𝑜𝑓 = = 0.3424 = 34.2% 𝑐𝑝𝑜𝑓 bone + 𝑤𝑝𝑗𝑒 𝑜 𝑢𝑝𝑢 14.01.2019 Olivier Schöpfer - EPFL - GC MA3 50

Simple traction test with Akantu • Input: • E = 1000 [MPa] • 𝜉 = 0.25 • Tests: • 3 Displacement controlled : x, y, z • 3 Force controlled : x, y, z 14.01.2019 Olivier Schöpfer - EPFL - GC MA3 51

Simple traction test with Akantu • Displacement controlled • 𝜗 𝑗 = 1% , i = x, y, z • Young modulus calculation for x 𝜏 𝑦𝑦 • 𝐹 𝑦 = 𝜗 𝑜 𝜏 𝑦𝑦,𝑗 σ 𝑗 • 𝜏 𝑦𝑦 = 𝑜 • n = #elements on the considered surface In Paraview, 𝜏 𝑦𝑦 = 𝑡𝑢𝑠𝑓𝑡𝑡: 0 ; 𝜏 𝑧𝑧 = 𝑡𝑢𝑠𝑓𝑡𝑡: 4 ; 𝜏 𝑨𝑨 = 𝑡𝑢𝑠𝑓𝑡𝑡: 8 14.01.2019 Olivier Schöpfer - EPFL - GC MA3 52