Dear Dr Okereke, Thank you for your wonderful set of videos about the BD composites. I have two equations in regard to this matter. 1- Is it okay to have set of geometry instead of nods when you put constraints and loading on them? 2- When you give the XZ pure shear you do both in the negative direction is there a specific reason for it? Thank you again for your time.
Hello, Thanks for your query. Here are my answers: 1. I prefer to associate the constraint to geometries instead of nodal sets as the later is subject to meshing of the model. If you change the mesh, you will have to re-attach/re-impose the constraints on the new nodes. But when you associate the constraint to a geometry, since the geometry does not change with the simulation, when you re-mesh, it will still inherit the constraint from the previous mesh, since those constraints are attached the the geometry. This is why I use geometries instead of mesh nodal sets. 2. The reason for using the negative shear value is simply aesthetic - nothing technical. The idea is that with the shear pushing away from the viewer, you would see better the shear deformation rather than when it pushes towards the viewer. Hope this makes sense? Thanks.
Hello, I think if you want to do an RVE modelling of a filament wound composite, at RVE level, then you are localizing the behaviour at RVE scale of fibre and matrix. The approach I showed here (especially in video about creating the BD RVE) will come in handy. This is consistent with what is done as reported in the paper here: link.springer.com/article/10.1007/s11029-017-9695-3 You should get some clue from this paper to help you.
Dear Dr. Michael Okereke, I have two questions. Firstly, how do I apply shear to the 2d RVE model? Secondly, I applied XFEM to get the damage of the RVE model, so far I can only get the cracking of the matrix, but I want to get the damage of the fiber as well. I changed the material parameters many times, and the successful result was that the matrix completely cracked without any damage to the fiber. Or there were multiple cracks, cracks appeared at the interface between the fiber and the matrix, but it was impossible to calculate and errors began to be reported. How do I cause damage to the fibers？ Looking forward to your reply and some suggestions.
Thanks for your question. With regards to 2D RVE model shear, here is a video I made in the past that addresses this: th-cam.com/video/oXgN1qK084k/w-d-xo.html I hope you find it helpful. With XFEM issues and fibre damage, I have seen this sort of problem before but that was for a microcapsule for a self-healing concrete. I think in your case you are undertaking a transverse to the fibre direction study in which case the matrix will fracture way before any fracture seen in the fibre. Most fibre damage tend to be in fibre-dominated direction (along the longitudinal direction of the fibre) rather than transverse to it. It is possible your result of the matrix cracking (multiply) over and above the fibre is correct but I do not know exactly your loading condition and what your experimental data say for a definite answer. If you however insist on the fibres to fail in this transverse direction then you must adjust the fibre properties to initiate damage and load to the corresponding loading size to make that happen. It might be unattainable.. but that is a thought to consider.
Hey, Dr.M.O, I'm a Chinese researcher, it is hard to watch your TH-cam video, and, I have some questions about the isotropic calculation of composite and lattice structure, Do we have to calculate the Stiffness matrix of RVE model with PBC ? and then, get the Zener anisotropy index to verify the isotropy of composite or lattice structure, In summary, do we have to apply PBC to RVE model when we calculate Zener anisotropy, You can reference this paper "Isotropic porous structure design methods based on triply periodic minimal surfaces". Looking forward to your reply and some suggestions. if it is possible, can you post a video about the isotropic calculation
Hello, thanks for your comment. I am not too clear on your question about isotropic calculation of composites and lattice structures. Are you interested in homogenized properties of these materials? I do not think you can use PBC to calculate the stiffness properties. PBC is a boundary condition and does not get activated until a probable stiffness matrix for the material is generated. Stiffness matrix is a material property and is independent of boundary conditions. I have not done any work with Zener Anisotropy index so unable to help you in that regard. One thing I know is that PBC is the best boundary condition for heterogeneous media and will give you results where errors due to boundary conditions are most minimal. The challenge with PBCs on RVEs is you have to find a way to obtain the homogenised properties. ~With that, you may be able to them determine the Zener Anisotropy index/ratio. With regards to the paper you identified, I have read through it and it seems a relatively straight forward paper. They have not used a PBC rather they used a Dirichlet BC. See equations (12) and (13) in the paper where they specify how they applied the BCs. They extracted stiffness matrix from their simulation and it was not based on the BC. In ABAQUS, there is a way to extract stiffness matrices directly from the simulation outputs. Look at this Quora post for more insight into this: www.quora.com/How-can-I-get-the-global-stiffness-matrix-from-the-Abaqus-model Hope the above helps?
Dear Dr Okereke, Thank you for your wonderful set of videos about the BD composites.
I have two equations in regard to this matter.
1- Is it okay to have set of geometry instead of nods when you put constraints and loading on them?
2- When you give the XZ pure shear you do both in the negative direction is there a specific reason for it?
Thank you again for your time.
Hello,
Thanks for your query. Here are my answers:
1. I prefer to associate the constraint to geometries instead of nodal sets as the later is subject to meshing of the model. If you change the mesh, you will have to re-attach/re-impose the constraints on the new nodes. But when you associate the constraint to a geometry, since the geometry does not change with the simulation, when you re-mesh, it will still inherit the constraint from the previous mesh, since those constraints are attached the the geometry. This is why I use geometries instead of mesh nodal sets.
2. The reason for using the negative shear value is simply aesthetic - nothing technical. The idea is that with the shear pushing away from the viewer, you would see better the shear deformation rather than when it pushes towards the viewer. Hope this makes sense?
Thanks.
Would you mind sharing a video about making RVE model of filament wound composite?Tnx
Hello, I think if you want to do an RVE modelling of a filament wound composite, at RVE level, then you are localizing the behaviour at RVE scale of fibre and matrix. The approach I showed here (especially in video about creating the BD RVE) will come in handy. This is consistent with what is done as reported in the paper here: link.springer.com/article/10.1007/s11029-017-9695-3
You should get some clue from this paper to help you.
Dear Dr. Michael Okereke, I have two questions. Firstly, how do I apply shear to the 2d RVE model? Secondly, I applied XFEM to get the damage of the RVE model, so far I can only get the cracking of the matrix, but I want to get the damage of the fiber as well. I changed the material parameters many times, and the successful result was that the matrix completely cracked without any damage to the fiber. Or there were multiple cracks, cracks appeared at the interface between the fiber and the matrix, but it was impossible to calculate and errors began to be reported. How do I cause damage to the fibers？ Looking forward to your reply and some suggestions.
Thanks for your question.
With regards to 2D RVE model shear, here is a video I made in the past that addresses this: th-cam.com/video/oXgN1qK084k/w-d-xo.html I hope you find it helpful.
With XFEM issues and fibre damage, I have seen this sort of problem before but that was for a microcapsule for a self-healing concrete. I think in your case you are undertaking a transverse to the fibre direction study in which case the matrix will fracture way before any fracture seen in the fibre. Most fibre damage tend to be in fibre-dominated direction (along the longitudinal direction of the fibre) rather than transverse to it.
It is possible your result of the matrix cracking (multiply) over and above the fibre is correct but I do not know exactly your loading condition and what your experimental data say for a definite answer.
If you however insist on the fibres to fail in this transverse direction then you must adjust the fibre properties to initiate damage and load to the corresponding loading size to make that happen. It might be unattainable.. but that is a thought to consider.
Hey, Dr.M.O, I'm a Chinese researcher, it is hard to watch your TH-cam video, and, I have some questions about the isotropic calculation of composite and lattice structure, Do we have to calculate the Stiffness matrix of RVE model with PBC ? and then, get the Zener anisotropy index to verify the isotropy of composite or lattice structure, In summary, do we have to apply PBC to RVE model when we calculate Zener anisotropy, You can reference this paper "Isotropic porous structure design methods based on triply periodic minimal surfaces". Looking forward to your reply and some suggestions. if it is possible, can you post a video about the isotropic calculation
Hello, thanks for your comment. I am not too clear on your question about isotropic calculation of composites and lattice structures. Are you interested in homogenized properties of these materials? I do not think you can use PBC to calculate the stiffness properties. PBC is a boundary condition and does not get activated until a probable stiffness matrix for the material is generated. Stiffness matrix is a material property and is independent of boundary conditions.
I have not done any work with Zener Anisotropy index so unable to help you in that regard. One thing I know is that PBC is the best boundary condition for heterogeneous media and will give you results where errors due to boundary conditions are most minimal. The challenge with PBCs on RVEs is you have to find a way to obtain the homogenised properties. ~With that, you may be able to them determine the Zener Anisotropy index/ratio.
With regards to the paper you identified, I have read through it and it seems a relatively straight forward paper. They have not used a PBC rather they used a Dirichlet BC. See equations (12) and (13) in the paper where they specify how they applied the BCs.
They extracted stiffness matrix from their simulation and it was not based on the BC. In ABAQUS, there is a way to extract stiffness matrices directly from the simulation outputs. Look at this Quora post for more insight into this: www.quora.com/How-can-I-get-the-global-stiffness-matrix-from-the-Abaqus-model
Hope the above helps?
@@MichaelOkereke Thank you for your answer. I will think about your suggestion carefully.