Great content! Thanks Dr. Michael Okereke.

@MichaelOkereke

Жыл бұрын

My pleasure @Jose Pablo!

Very thanks Dr. Michael Okereke.

@MichaelOkereke

Жыл бұрын

You are very welcome @Valerii

Dear Dr.Okereke, thank you for your informative tutorials. I'm wondering if there is a semi/automatic method to impose PBC on 3D RVEs which are made from two parts. I'm working on a 3D RVE of a UD composite and I have to model fibers and matrix in two different parts to be able to define cohesive surfaces between them. Moreover, my 3D RVE has fibers on edge. Do you have any idea how should I define PBC in this case ? Thanks a lot.

@MichaelOkereke

Жыл бұрын

This is definitely not a straightforward way of imposing PBCs on 3D RVEs. I published about this a while back so you can read about it in this publication: doi.org/10.1016/j.compstruct.2016.10.114 You should get some pointers in that publication. I do not have a publicly available code for doing such but am working on it. For the situation with cohesive surfaces, yes, you have to allow all components to be independent (unattached) so you can specify them independently. It can prove difficult but it is doing. I do not have a video about this but can explore this in future.

Thanks for this insightful video. I'm wondering weither I can use this PBCGen2D software for a hexagonal shape marix unit cell ?

@MichaelOkereke

Жыл бұрын

Hello @Syamil, I will say no. The PBCGen2D software works on 2D Quadrilateral RVEs so it will unfortunately not work in this case.

Thanks sir, your videos are very helpful. I was wondering how to do the same procedure on a lattice RVE in which nodes must be paired considering opposite surfaces. It's a huge effort doing that manually expecially when each face has around 200 nodes. What do you think? Thanks a lot

@MichaelOkereke

Жыл бұрын

Hello @Martina, I am glad that the video is helpful. For 2D RVE lattices, I did a video in the past using a the automatic way of implementing PBCs (see here kzread.info/dash/bejne/Zo140K2ol73Ffso.html). For surfaces, as you describe here, it is a different implementation. This manual or semi-automatic approach will not be suitable as there are too many nodes to be able to use them. You will have to use an automatic approach where a script searches through the surfaces and note node numbers, nodes coordinate positions and their collinearity with spatially complementary opposite nodal sets. I plan to bring some theoretical and numerical highlight to these things in future.

@martinascordio1648

Жыл бұрын

@@MichaelOkereke thank you very much Dr. Michael Okereke

Sir...iam using ansys workbench for my research work..how I can apply PBC in ansys workbench or APDL

@MichaelOkereke

Жыл бұрын

Hello @Jiyad, I am not sure I know how this is done in ANSYS. It is something I am still investigating. Thanks for the question.

Very interesting video. Are you aware of the Micromechanics Plugin for Abaqus/CAE that DS provides? It was released at the end of 2016 and automates the imposition of PBCs (and a few other types of constraints) as well as defining loads and performing homogenization on RVE models. Is there a reason that you only constrain the normal component of displacement across the periodic faces? If the tangential component(s) of displacement aren't constrained, you're not actually going to get a periodic displacement field - the tangential displacements will vary between faces in a non-periodic manner. I also think that this will lead to a zero shear condition along the cut faces of the RVE as the tangential components of stress will necessarily go to zero at the boundary (although the zero shear in the derivative fields may not be immediately apparent due to low mesh refinement) and in general the strain and stress field will be non-periodic. One nice way to check the correctness of PBCs is to create a model with multiple tessellations and compare the results when you model one cell and two cells. The stress/strain solution should be identical at a given point.

@MichaelOkereke

11 ай бұрын

Hello @RossMcLendon, Thanks for your detailed comment and pointing out a few things. On *Micromechanics Plugin for ABAQUS CAE*, I actually did not know about it. I only researched into it to find your article about it here: www.linkedin.com/pulse/micromechanics-plugin-abaqus-ross-mclendon/ I know ANSYS has a similar module as this but not ABAQUS. It will be interesting to have a play with it and more importantly understand the theoritical framework that underpins its development. I believe the ABAQUS license I hold with my organisation might not cover that plugin. On *Constraining the periodic faces of rve*, I am not sure you understand completely the framework I am working on. It is not constrained only in the normal direction, rather it is for both normal and transverse in otherwords x and y-axis. The principle here is to isolate first identify the corner nodes (the vertices of the quadrilateral) and use those to use as anchor points to apply the required loading and boundary conditions on the model. The nodes away from vertices are then implemented with a *EQUATION driven periodic boundary conditions formulation that constraints periodic nodes (directly opposite semi-colinear nodes) in both x- (normal) and y- (transverse) directions. It is actually a well established approach that has been published widely in the community and peer reviewed. A lot of authors have adopted this approach and as a student at Oxford University many years ago, I first read about it by an author called Kouznetsova and implemented this as part of my PhD work. I have since published about it extensively and here is a paper where I laid down the principles in detail doi.org/10.1016/j.commatsci.2012.12.036. I also gave detailed description about this in a 2018 Postgraduate textbook I authored and published by Springer which you can read here: link.springer.com/book/10.1007/978-3-319-67125-3 You may like to look at Chapter 8 about Boundary Conditions. On *Correctness of PBCs*, I have already explored this in the said publications above and shown it to be true. The journal paper has a parametric study where I explored different mesh descritations beyond just unit cells and it showed that the PBC implementation is right. Thanks once again for your detailed comments. I will love to catch up separately to continue this discussion. Cheers!

@RossMcLendon

11 ай бұрын

@@MichaelOkereke Okay, watching again I think you are defining them correctly - I was mistakenly thinking that you were describing the equation for N_A and N_B as involving only DOF 1 and N_C and N_D as involving only DOF 2. On review I see that you're describing two equations for N_C and N_D, one for DOF 1 and another for DOF 2. Assuming you also define similar equations for N_A and N_B (with N1 and N2), then what you're describing should be periodic. The Micromechanics Plugin uses a somewhat different approach to define PBCs, but the two results will be in essence equivalent apart from how loads are defined (with the plugin, you directly specify strain components rather than specifying relative displacements between periodic faces).

@MichaelOkereke

11 ай бұрын

Certainly @RossMcLendon, I agree. There is definitely a kinematical constraint equation involving also N1 and N2 to make the model truly periodic. This is why I divided them into internal nodes (NA, NB, NC and ND) and the corner nodes (N1, N2, N3 and N4). By intelligent combination of displacements at corner nodes kinematically linked to the internal nodes, we can then define a canonical equation (which I discussed in the video) which is then subsequently implemented using *EQUATION to enforce the right periodicity argument. I am aware of your strain-tensor based approach for imposing PBCs. I had a PhD student who used that approach as part of his PBC implementation for a textile composite study. In fact, in that method, he was able to deal with non-periodic meshes and you can read up about this here: doi.org/10.1016/j.compstruct.2016.10.114. I intend to make a video on that approach to complement what I have presented here. I first got exposed to this displacement-based approach and love the intuitive way it works, something users who are already comfortable with Dirichlet BC-type approach, can easily understand. Your approach will certainly allow for applying combined loading easily based the strain tensor. I know there are other benefits. I will write you outside this forum to follow up on these interesting conversations. I look forward to playing with the Micromechanics plugin and see how useful and complementary it is to what I present on this channel. Thanks once again for taking time to comment on this.

sir i have a doubt, I am modelling a composite laminate of 8 layers along with the frame on top and bottom. I am using VUMAt code and applying a predefined velocity in initial step on the ball and applying pressure on top frame in step one. I need all my boundary conditions and pressure on the frame has to be applied before the ball hits plate. so how can i do it. Please help me in this case

@MichaelOkereke

Жыл бұрын

Hello @Manoj, this is too specific a request and I don't usually honour this sort of requests. It seems a project you are working on and I'd leave you to figure it out. I tend to make general videos to give ideas to viewers rather than specifically take a project and offer solutions here. As an academic I want my students to do their projects themselves. Good luck with the work 👍.

@manojkumarneesu2773

Жыл бұрын

@@MichaelOkereke oke sir i respect your words....thanks 🙂