RVE modelling of Metal Matrix Composites in ABAQUS

[@anittaa6243]

Great work! Thank you very much for your effort and sharing information, but I cannot understand why young's modulus of the analyzed nanocomposite (21.7 Gpa) is come out to be lower than the matrix's (75 Gpa)? Isn't the purpose of adding nano-particles improving the mechanical properties (including young's modulus) of the matrix? I would be more than glad if you answer my question! :)

[@MichaelOkereke]

Good response again Vincenzo. Well done!

[@MichaelOkereke]

Hello Anitta, the particulate inclusions (even for very small volume fractions) should indeed improve the mechanical properties i.e. stiffness and strength as well. This is the principle of composites generally. Your observation about the lower Young's Modulus is valid and when I was shooting the video, I did wonder about that number but did not get time to review it to assess what it should be. I sense the way I have extracted the stress-strain plot might be a reason for the lower values but this is something I need to review more carefully. Your comment is a validation issue and for such micromechanical studies of particulate composites, the problem is that you have to choose the right micromechanical model (something like Halpin-Tsai type model) to compare numerical predictions with analytical predictions to know exactly if your numerical predictions fit in with analytical-type predictions. As this videos was not about that, I did not spend time to make those comparisons. What I should do in future is to make a video where we can look at the specifics of numerical versus analytical model predictions for particulate composites. For now, the essence of this video was to show how such RVE models can be set up but objectively assessing predictions (quantitatively), that will be subject of a different video. Thanks though for the observations, it really got me thinking. Please do keep coming back to the channel for future videos that can address the subject of your comment!

[@anittaa6243]

@@MichaelOkereke Thank you very much for your kind and detailed response Dr. Okereke! I can't wait to watch more videos regarding this topic and learn more from you! :)

[@allanshadowrine5187]

You are the best material science and modeling teacher I have met in life. Your videos are not scanty and far-fetched, they are precise, educative, and real. Thank you always for making these videos and making FE easy for interested learners like me!

[@allanshadowrine5187]

And also thank you for finding time to reply to students and comments posted here

[@MichaelOkereke]

Hello @Allan Shadowrine, thanks for your kind comments. It is my pleasure to make these videos and share a bit of what I know. I actually enjoy it and I love the engagements on here as it opens me up to diverse ideas and also challenging projects to work on. I am happy you found them helpful and feel honoured. Thank you.

[@MichaelOkereke]

Of course @Allan, it is definitely a responsible thing to respond to comments that people take time to make on this channel. I will keep doing so until I am not able to do so due to the multitudes of comments. We are not there yet so I will keep going and thanks again for taking time to comment. It means a lot to me.

[@eng.husseinmahersaad7056]

Dr. Michael, I appreciate your effort and time in making these videos and educating research fraternity. Thanks

[@MichaelOkereke]

Thanks Eng.

[@SeanRuben]

This was very helpful, thank you Dr. Michael!

[@MichaelOkereke]

Thanks @Sean, am glad it was helpful. Have a nice day.

[@averilfernandes9514]

Dr. Michael, I appreciate your effort and time in making these videos and educating research fraternity. If you get a time kindly share a video on thermal analysis especially to find the thermal conductivity of particulate metal matrix composites. Thanks in advance.

[@MichaelOkereke]

Hello @Averil Fernandes, thanks for your interest in the channel. In deed, that is a great video suggestion and I will look into it.

[@samuelkefyalew5153]

Thank you Dr. Michael Okereke!

[@MichaelOkereke]

You are welcome @Samuel Kefyalew!

[@ducdeubro1]

Thank you very much for great video Dr Michael

[@MichaelOkereke]

You are very welcome @Viet Nguyen. Thanks for your interest in the channel.

[@HamzaDjeloud]

Great wourk Doc;thank you for sharing the information

[@MichaelOkereke]

Glad you found it helpful. Thanks!

[@dsflkvbndflvkndflomvpsdmvlkasd]

Great Work Sir, Thanks for sharing such knowledge. God bless you.

[@MichaelOkereke]

Thanks Abhishek for your interest in the channel.

[@subratkumarmaharana213]

Great work sir

[@MichaelOkereke]

Thanks Subrat for your continued interest.

[@brwasaeed6477]

Amazing ! great tutorial Thanks

[@MichaelOkereke]

Glad you liked it! Thanks.

[@rajanikantrao4929]

Wonderful! Dr.Michael, please make a video on effective electrical conductivity for nanocomposite.

[@MichaelOkereke]

Hello@Rajanikant, thanks for the query. I have actually never worked on electrical conductivity so unable to help now. Maybe in future I could be in a position to make such a video. Thanks and good luck with the research

[@Wani-m2c]

Great Work Sir, can you please do this for microstructure-based RVE modelling. I mean where I can upload my actual two-phase microstructure obtained using SEM.

[@MichaelOkereke]

This is possible. I do not have any video based on SEM-derived virtual domains. Direct-message me on my socials and we can explore what to do about this.

[@mukunthag8760]

Hello Dr. Okereke, is there any scope for applying generative AI models to aid in synthesis of metal matrix composites or just any composite? What do you think about this?

[@MichaelOkereke]

There is certainly a possibility of AI helping with the generation of metallic composites. I made a video for UD composites and AI as seen here: kzbin.info/www/bejne/imaVgJmAf7t2rKM. I believe you can adapt a similar process for metallic composites.

[@kamaleshs4829]

Great Doctor!! Can you please put a video on tensile test of this Particulate metal matrix composite??

[@MichaelOkereke]

Hello @Kamalesh, the tensile test is what I did here. I applied the xTensile load using a slightly different approach but that is what is going on here. Is there anything else you have in mind?

[@vikramroy7471]

Thankyou sir for these videos can also start making videos on crystal Plasticity finite element methods

[@MichaelOkereke]

Noted

[@KesavamoorthiRPHD]

Hey Doc wonderful video and great work. kindly share 3D RVE software DOC that will help to research work

[@MichaelOkereke]

Will do soon

[@shahxaib924]

Great work. Highly appreciate. Dr. Michael Okereke, would you please make some videos on superelastic reinforcement base RVE including their modeling, extracting mechanical properties, and most important their convergence criteria? Thank you in advance.

[@MichaelOkereke]

Hello @Shahzaib, thanks for your interest in the channel and kind comments. I am unsure what you mean by superelastic reinforcement base. What is the main material you are considering here. Is it a reinforced composite or concrete in which the reinforcing medium is a superelastic material? Please explain and you can point me to a paper for me to consider what you mean. Thanks.

[@ishfaqbhat1351]

Very good, can we do it for hollow particles , can you show it please

[@MichaelOkereke]

Dear Ishfaq, thanks for your interest in the channel. In deed, this is a great suggestion. I will consider and see if I can make a video about hollow particles. If you have a reference (publication) that should help me understand the problem better, please share the link and I will start from there.

[@KartheeKSoNTI]

Can you please tell why you have considered reference point?

[@MichaelOkereke]

Hello Kartheek, the use of reference point here was to use it to extract reaction force and displacement measurements for the model. This will mean in calculating normal stresses and strains, history data from the reference point can be stored and used. Without a reference point, you will have to specify a nodal set and then do volume averaging over those nodal sets. This is okay if you have a model with a few elements. If there are a lot, ABAQUS will often complain that you have more that 10,000 variables to store in History and so runs out of memory. The use of a reference point means you are only tracing a nodal point and it will only return values according to the number of time steps used to complete the simulation. Therefore, data storage is manageable but of course you will need to take that data and manipulate it to get your stress-strain plot. I hope the explanation makes sense to you.

[@KimKinal]

Sir, why is periodic boundary condition not needed in this case?

[@KimKinal]

I have seen the comment of your on PBC with 3D. But may I know if the result is the same with or without PBC? Thank you, I hope your reply. Your videos are great valuable resource , sir. I thank you

[@KimKinal]

Also, can I do high-strain rate compression on the same model?

[@MichaelOkereke]

Hello @Kim Kinal, this is a good question. I could use 3D PBC here but that will require a PBC Generator script for 3D which is not currently available for general use. In future, I will make videos where I will do so. In terms of the result, there will be an effect between use and not use of PBC. Usually, you get better convergence of the simulation to the right result with PBC compared to the case without it. I published a paper in the past which discusses this in more detail. You can get the paper from here: doi.org/10.1016/j.commatsci.2012.12.036.

[@MichaelOkereke]

Yes, you can do high strain rate compression with this model. You just have to make the step an ABAQUS Explicit step and allow for am amplitude curve that allows for dynamic loading. There may be some modifications here and there you need to make to damp out dynamic effects due to the high strain rate, but the simple answer is yes, you can.

[@jagadish530]

Thank you so much, dear Dr. Okereke, whatever you share with us is great work. Would you like to make a video for a compression test of the same?

[@MichaelOkereke]

Great suggestion! It should just be a case of reversing the directional of the uniaxial loads. If there is interest for this, then I can consider making such a video.

[@jagadish530]

I am very much grateful, for your valuable response. I am very much interested in this.@@MichaelOkereke

[@ahmedelmisaid8978]

Great work! THANK YOU VERY MUCH ! !!!

[@MichaelOkereke]

Thanks Ahmed... glad its useful to you.

[@meena576]

Very informative video

[@MichaelOkereke]

Glad you liked it

[@davidgonzaleznino5914]

Hi! Great Video! I have a question. Is Periodic Boundary Condition Applied to this simulation? How different would be if PBC is applied vs no-PBC? thanks!

[@MichaelOkereke]

Hello @David, thanks. There is no Boundary Condition on this simulation. If it were to have PBC, you will see a periodic deformation with the shape at the back mirrored at the front for all the three planes (x-, y- and z-planes). In this instance, I have fixed the back using a Dirichlet type Boundary Condition and loaded the front edges. This gives a deformation that may seem periodic in the front but this is entirely due to the presence of the inclusions (particles), causing the deformation to be non-uniform. I hope this makes sense.

[@ghadirahmadijeyed9891]

Dr. Micheal. Please model fiber random RVE in abaqus.

[@MichaelOkereke]

I will take note of that. Thank you. I have videos on randomly distributed unidirectional composite, and another on short fibre composites. Please have a look at them.

[@jishmaalex3283]

Thank you sir for your videos they are really help ful. Can you please show how to apply interaction in abaqus. Not for plane surface . Something like a single spherical particle completely inside a rectangular matrix. How to introduce interaction between these two? Is it possible? Without merging it

[@MichaelOkereke]

Hello @Jishma Alex, thanks for your comment. I believe this is possible. I will think about it and see what I can do about it and make a quick video to show how this can be done.

[@jishmaalex3283]

@@MichaelOkereke thank you sir for your efforts

[@YesufikadFente]

Dear Dr. please provide us 3 different particle type reinforced aluminum metal matrix composites. Regards you did good for us, respect

[@MichaelOkereke]

Thanks. Noted. Do you mean the particle types are 3 different shapes within the same material or three videos with different particle shape per video?

[@YesufikadFente]

Thanks Dr. Michael Okereke. i mean for example, Al203, Graphite, SiC reinforced within AA6061 composites.

[@alexstudio38]

very interesting

[@MichaelOkereke]

Thanks

[@AshwinMS-KIMS]

is it possible to do the same for a ceramic composite to simulate thermal conduction ?

[@MichaelOkereke]

Yes, definitely. All you need to do is to incorporate the thermal properties of the constituents for a thermal simulation (conduction) to be completed.

[@AshwinMS-KIMS]

@@MichaelOkereke could you please do a video on the same

[@lamiakamal1005]

Thank you Very useful

[@MichaelOkereke]

Thanks too

[@hosseinghorbani_menghari2956]

Hello. Thank uou for your great video. I am currently attempting to simulate the compaction powder forming process using the multi-particle finite element method in Abaqus. Using the assembly module, I have assembled a collection of particles and enclosed them within a box. My goal now is to remove the particles that are outside of the box, while retaining the ones that are inside. It's important to note that I do not wish to merge or cut the particles, but rather simply delete the portion that is outside of the box. Can you please provide guidance on how to achieve this?

[@MichaelOkereke]

Hello @Hossein, thanks for the query. I think the challenge to this arises from what you have set as a constraint: "you do not wish to merge or cut the particles". In this case, it becomes difficult. If you merge the particles, then you can easily extrude cut through the particles specifying the boundaries within which you accept (i.e. your current domain) and outside that boundary, anything there will be extrude cut. I however, do understand the necessity of not merging or cutting the particles. If this is a constraint you must live with, I will then suggest that you identify individually the particles that are appearing on the edge. YOu can label them as a set, then individually extrude-cut them one by one so that in the end, they will be the only ones that you extrude cut. It might be easier to write a python script that does this repetitive process once you have a collection of the boundary fibres. Isolate them according to their coordinate positions for example, left edge and right edge particles. This way, you can then deal with them individually. One other thing you should consider is to use a datum plane with say a principal axis to do the extrude cutting. For example, the plane of your extrude cutting with not be the particles (as they are spherical, I hope) but the datum plane. The direction of extrude cutting will be defined by a principle x-, y- or z-axis depending on how you want to extrude cut directionally. You can reach out to me on LinkedIn or Twitter (direct messaging) or the About page here on KZbin for further directions.

[@hosseinghorbani_menghari2956]

@@MichaelOkereke thank you so much for taking your time. I am trying to do individualy in assembly module but I did not managed to do that. My particles are spherical. Could you please explain more how can I do the procedure you explained?

[@xiaolingchen5397]

It is a great job! Thank you very much about this. Could you please make a video about how to deal with the interfaces between the reinforcement and the matrix in composites?

[@MichaelOkereke]

Hello @xiaoling chen, thanks for your interest in the channel. Indeed, this is something in my to do list and hopefully soon, you will get this video. Please keep watching!

[@mukunthag8760]

@@MichaelOkereke is there a video out now on this one?

[@saifulislam-bf3xu]

Hello, Can you make a video for collagen fibril reinforced matrix, specifically for 2D cartilage modelling

[@MichaelOkereke]

Hello @saiful islam, thanks for your interest. Unfortunately, I have not done any work on cartilage modelling before so unable to help with this query, but it does sound interesting. Good luck with the research.

[@kazd17]

Great video! Very helpful! I am interesting in how to incorporate cohesive elements in asphalt concrete to study a fracture problem. Can you, please, advice on this topic? Thank you very much.

[@MichaelOkereke]

Hello Kaoutar, thanks for your interest in the channel. I have not worked on asphalt concrete but this is something that I can research in. I think the incorporation of cohesive elements in many material models is quite common now. I want to do a similar video on UD composites and that you can translate to your problem. Or maybe in future, when I learn a bit more about asphalt concrete, I can consider to make such a video to help you.

[@BishalPudasaini]

Sir, Can we do such analysis for randomly oriented short fiber reinforced polymer in ANSYS? And what about using Periodic Boundary conditions in ANSYS workbench for such analysis?

[@MichaelOkereke]

Thanks Bishal for your interest in this channel. I have actually not started focussing on ANSYS simulations. This is planned for the future. For now, I am building up the knowledge base for ABAQUS and will start creating ANSYS contents in future. On the question of short fibre reinforced polymer, yes, it can be done and I will consider that for a future video but that will be for ABAQUS for now. On PBCs in ANSYS, this again will be for a future video. Thanks for the feedback and do please keep coming back for when ANSYS contents become available.

[@BishalPudasaini]

@@MichaelOkereke Thank you so much for the response. Actually I am doing my thesis related to these contents and your contents are being of great help for me. And the questions I have asked are the areas where I am facing problems. I would be very grateful if I could get some more help from your contents on the areas where I am facing problems.

[@bonthudileep9464]

Sir is there any many method to model porous structure in ANSYS or any possibility like building model in abacus and export to ansys

[@MichaelOkereke]

Hello Bonthu, I am aware that you can export a model from ABAQUS to ANSYS, so this is possible. I will try and make a quick ABAQUS tips and tricks video on how to do so.

[@bonthudileep9464]

@@MichaelOkereke ur so kind hearted and student supportive sir

[@arshks8089]

Superb work sir. Can you make metal matrix hybrid composites with matrix magnesium and reinforcements as titanium diboride and yttrium oxide

[@MichaelOkereke]

Thanks Arsh for your interest in the channel. Definitely, the approach I showed in this video can be applied to the composite you referred to. YOu just have to find the right properties for the constituents and use it in your simulation. Unless, there is a different microstructural arrangement which will require a different virtual domain built for your hybrid composite. You can explain a bit more for us to know what you mean!

[@KartheeKSoNTI]

Sir for pmmc can you please show how to perform compression test in explicit dynamic simulation condition? I really appreciate your work.

[@MichaelOkereke]

Hello Kartheek, the principles shown here will work for an Explicit Dynamic simulation, you just have to specify a different analysis step - in your case, the explicit dynamic case. I do agree that there may be a need for a dedicated video dealing with compression response of PMMCs with explicit damage simulation and maybe damage incorporated. That can be the subject of a future video depending on interest from the viewers.

[@nickg1326]

Very interesting and helpful video! Could you please answer: in case of cell shear, we collect RF2 and U2 data from RP and end up with effective shear modulus and shear strength?

[@MichaelOkereke]

Hello @Nick G, thanks for your interest in the channel. The reason for this is because the RF2 is force parallel to the X-edge and shear stress is this force divided by the area of this x edge. The U2 term is displacement in y-axis and then divided by the Y edge gives you the shear strain. To get shear behaviour in xy plane you need y-axis displacement and forces (u2 and rf2). Similarly for shear on yz plane you will need the z-axis displacement and forces and so on. This is how it works.

[@nickg1326]

@@MichaelOkereke thank you very much for this answer, you helped me a lot.

[@hosseinghorbani_menghari2956]

I am currently attempting to simulate the compaction powder forming process using the multi-particle finite element method in Abaqus. Using the assembly module, I have assembled a collection of particles and enclosed them within a box. My goal now is to remove the particles that are outside of the box, while retaining the ones that are inside. It's important to note that I do not wish to merge or cut the particles, but rather simply delete the portion that is outside of the box. Can you please provide guidance on how to achieve this?

[@MichaelOkereke]

Hello @Hossein, thanks for the query. I think the challenge to this arises from what you have set as a constraint: "you do not wish to merge or cut the particles". In this case, it becomes difficult. If you merge the particles, then you can easily extrude cut through the particles specifying the boundaries within which you accept (i.e. your current domain) and outside that boundary, anything there will be extrude cut. I however, do understand the necessity of not merging or cutting the particles. If this is a constraint you must live with, I will then suggest that you identify individually the particles that are appearing on the edge. YOu can label them as a set, then individually extrude-cut them one by one so that in the end, they will be the only ones that you extrude cut. It might be easier to write a python script that does this repetitive process once you have a collection of the boundary fibres. Isolate them according to their coordinate positions for example, left edge and right edge particles. This way, you can then deal with them individually. One other thing you should consider is to use a datum plane with say a principal axis to do the extrude cutting. For example, the plane of your extrude cutting with not be the particles (as they are spherical, I hope) but the datum plane. The direction of extrude cutting will be defined by a principle x-, y- or z-axis depending on how you want to extrude cut directionally. You can reach out to me on LinkedIn or Twitter (direct messaging) or the About page here on KZbin for further directions.

[@nitinsrivastava6720]

hello sir, thank you for this video I learn many things. can you please tell me how to use metal matrix composite mechanical properties through Abaqus

[@MichaelOkereke]

Not sure what you mean by use? Is it modelling you mean in which case that's what this video is about.

[@amalkrishna.a8968]

Sir, the shared content is great and helpful but it would be really kind if you could show a video on producing voids in different structural shapes like "I" shape.? And please let us know how to generate our own coading for producing voids randomly.?

[@MichaelOkereke]

Hello @Amal, thanks for the comment. I have not seen I-shaped particulates, do you have a reference that I can read to see if I am able to make such a video. With regards to random generation of voids, I have already a video about this. Check it out here: kzbin.info/www/bejne/mKqYmKmnmLpjoJI

[@ZeDe439]

Great job Doc. just a question. Could you explain how this equation/dimensions and calculations will change for a nanocomposite case?

[@MichaelOkereke]

Hello Adedotun, thanks for your interest in the channel. I believe you can still use the same equation. Just make sure you are using the right formula for the inclusion type i.e. if it is not spherical, it could be cylindrical, for example in carbon nanotubes. In this case, you calculation of number of inclusion types must account for the different geometric shape.

[@akshaykumarpakala839]

Great video sir. I have a question. I created an RVE model in Matlab. This RVE model is a cube with Voronoi polygons. I have the faces, cordinates of seeds, vertices, and also information on how the faces are connected with each other in matlab. I would like to export this RVE to Abaqus from MATLAB using python scripting. do you think this is possible?

[@MichaelOkereke]

Hello @Akshay, I have not dealt with this type of problem before but I do think it is possible, if you have all the nodal coordinates and element connectivities. It might help if you can re-format the information into an ABAQUS input keyword style so that it is easy for ABAQUS to read. The problem with your MATLAB model is that it is not in a format that ABAQUS can read. So, some translator of a kind is required. You can study the ABAQUS input keyword file and see if you can re-format the information you have into that format. That means you need to know the element connectivities for all the cells/meshes in the MATLAB model and their nodal coordinates and node numbers. It is not trivial but it is certainly possible, in my opinion. Good luck with the research.

[@akshaykumarpakala839]

@@MichaelOkereke Thank you sir for the reply. So all the information I currently possess is based on geometry rather than the mesh. To be specific, I have the coordinates, faces, vertices and connectivity information. Now all I have to do is understand how to write this in a format so that Abaqus can read it as a geometry input. So, I would like to know if there is any material I can get into so that I can understand how to write this document.

[@MichaelOkereke]

Hello Akshay, I do not have any material. I think you should search for 'from MATLAB to ABAQUS translator', you may be able to get some resources online. Sorry!

[@abhaygore13]

How to do heat transfer analysis of the same.

[@MichaelOkereke]

Hello Abhay, thanks for your interest in the channel. I have not done any heat transfer analysis of this type of composite but will think about it. Kind regards,

[@abhaygore13]

@@MichaelOkereke Thank you for reply. It would be interesting to see particle behavior during heat transfer, applications are for machining by EMD, WEDM, Leser, hybrid and what not.

[@nikhiltg4798]

Is the simulation at a microscopic level.?

[@MichaelOkereke]

Yes it is Nikhil.

[@nikhil.sharma443]

Nicely done. Sir, can you do same simulation for the epoxy-graphene matrix?

[@MichaelOkereke]

Yes I can. Graphene composites modelling takes a slightly different approach because of the differences in lengthscale. This is a frequently requested video and in the coming weeks, I will make a video for graphene-based nanocomposites to address your query Nikhil. Explain what you are looking to do with the graphene-epoxy matrix? Do you have a reference to a paper that you are looking to recreate the study numerically? If you can give me some insights into what you are hoping to do/achieve, then that can help create a video to help you with your work.

[@nikhil.sharma443]

@@MichaelOkereke Sir, kindly share your email. I will share an article for the same and also explain my problem. Thanks.

[@MichaelOkereke]

Hello @Nikhil Sharma, please join my CM Videos Insiders group and you will be able to contact me directly. Here is the link: cmvideos.ck.page/16a99bd0e3

[@nikhil.sharma443]

@@MichaelOkereke Dropped the mail into your mailbox. I hope the data will be helpful for resolving the query.

[@addykhan-tc2fs]

Great Work Sir, Thanks for sharing such knowledge. Can you please explain how to create intersections between composite materials inside RVE at the mesoscale level of normal-strength concrete? i.e., between coarse aggregate and matrix, between sand and matrix etc

[@MichaelOkereke]

Not sure about this... sorry!

[@ashfakursalehinlemon]

Sir, how to get shear modulus from here?

[@MichaelOkereke]

Hello @Ashfakur, here is a video of mine that might help: kzbin.info/www/bejne/ZnbZpZZ8rsx8eKs

[@mamdosein5207]

Dear Dr. Okerke, Thank you for such great and informative videos, can not wait for future works and presentations. I have a question: As I realize from channel videos, for giving periodic boundary conditions (PBCs) we should have periodic mesh in opposite faces, however, in the case of this model the mesh seems to be not periodic... does it mean if we want to apply PBC on a non-periodic mesh we can do it manually as taught in the video? Thank you for your help and time.

[@MichaelOkereke]

Thanks for your kind comments. You are right that when we model with PBCs, we will often have to have periodic / symmetric nodes on opposite and parallel faces/edges. This makes the sort of implementation I have in this channel work. However, for truly heterogeneous media with random distribution of particles (and periodicity of material not guaranteed), it is not always possible to enforce this periodicity of mesh argument. In that instance, a different implementation will apply. I have used the polynomial interpolation approach for this but it is not always that straight forward. I will hope to make some videos on this ideas in future to complement what I do here for periodic meshes.

[@mamdosein5207]

Dear Dr. @@MichaelOkereke Thank you again for your time to answer the comments and your insights. I have another question which is more of a general question: Could you please explain when we should consider PBC for our RVE model and also when we can generally conduct a polynomial interpolation approach, as you mentioned earlier, (does it have any connection to the size of our RVE) With kindest regards Mohammad

[@MichaelOkereke]

Let me attempt to explain: 1. PBCs are best when you are working on heterogeneous materials where the constituents are clearly defined and their behaviours distinct from themselves. So, to appropriately capture their distinct behaviour, a PBC will allow for a stress wave to pass through them and generate the required wavelike response deforming according to their material response (for the constituents). 2. Traditionally, PBCs are implemented easily (sort of) when you can have same nodes on opposite and correspondingly parallel faces/edges. This way, the kinematic constraint required to impose PBCs can be linked between the two colinear nodes taken from the two parallel and opposite edges. For this to happen, you will need same nodes on the two candidate face/edges. This leads to the requirement of periodic or symmetric nodal discretizations (meshes). 3. However, for the second question, if you cannot be sure of enforcing such periodic/symmetric meshes, then a method that guesses (interpolates) the behaviour of a point and a corresponding opposite edge point will have to be made. The guess method that is commonly used for this is the so called polynomial interpolation. It is a type of implementation for applying PBCs where nodal discretizations on candidate face/edges that are parallel and opposite themselves is unsymmetric or arbitrary meshes. So, in summary, PBC and Polynomial interpolation are related with the later being simply an PBC implementation style for arbitrary meshes.

[@qiangchen9503]

could you please elaborate on how the monte carlo works for generating random distributions of particles

[@MichaelOkereke]

Hello Qiang, thanks for your interest in the channel. I am going to make a series of videos on elaborating the Monte Carlo process so please keep watching. I believe it is more beneficial creating videos to address this query that just providing short answers in the comment section. So, keep watching.

[@qiangchen9503]

@@MichaelOkereke thank you for your response and nice videos.

[@brookcui6542]

Thank you Dr. Okereke, the video is really helpful. I have a question: Since this is a uniaxial loading condition, why do you need the roller supports at x, y and z=0 instead of just on the x=0 face?

[@MichaelOkereke]

Hello Brook, you are right about just putting an x axis constraint but I decided to use an RVE that's a quarter of the specimen hence the boundary condition I used. Both are okay but I will usually go with the quarter rve to save computational cost.

[@abhilashhamsa3470]

Great work Dr. Can we do this in ANSYS also??

[@MichaelOkereke]

Yes, definitely Abhilash, I am considering this and will return to it in future.

[@abhilashhamsa3470]

@@MichaelOkereke I was searching internet ocean 🌊. Please model particle metal matrix composites using ANSYS sir.

[@MichaelOkereke]

Yes @Abhilash, this is something in my to-do list. I am now exploring videos that I can make in ANSYS, so please keep watching so you can see the videos when they land. Is ANSYS the software you use?

[@malikademola9781]

Good work sir! I am trying to run similar simulation on ansys Can you help with that sir?

[@MichaelOkereke]

Hello Malik, thanks for your interest in the channel. I am not working on ANSYS models yet but when I do, definitely will consider this.

[@shirilekhaz957]

Great work sir, Can you please make it on Ceramic matrix composites as their behaviour is quiet different.

[@MichaelOkereke]

Dear Shiri, thanks for your interest in the channel. Definitely, I will consider making similar videos for Ceramic matrix composites. As you said, they will be different due to the brittle nature of ceramic matrices. Please keep coming back to the channel to see when the said video drops.

[@nss2302]

Thanks for the video, will there be any changes if you were to evaluate young’s modulus and yield strength in y or z direction ?

[@MichaelOkereke]

Hello N SS, certainly since the material studied is a composite, then due to its transverse isotropy, it will have different properties in x, y and z directions. You can find this out by doing a simulation in x- and y-directions, and extracting what the effective properties will be. You will find them to be slightly different as expected for such RVE heterogeneous media. If you however do a modelling on a homogeneous media (say steel, aluminium, without any inclusions, voids or holes), then the properties will be the same in all direction since such homogeneous material show isotropic behaviour.

[@nss2302]

@@MichaelOkereke thank you very much.

[@joseluischanmayen7067]

Hi Michael, I'm currently working on a design for a magnetorheological elastomer isolator, so this can basically apply to that correct? It consists on a silicon matrix with iron particles randomly dispersed

[@MichaelOkereke]

Hello Jose, thanks for your interest in the channel. I believe the system you describe - Magnetorhelogical elastomer isolator - can be modelled using the ideas of this Metal Matrix Composite system. I believe the silicon matrix will be the matrix component and the iron particles will be the reinforcing inclusions. I am not sure what shape the iron particles are - if they are spherical, then definitely the system here will work. I also have another similar video about elliptical shaped inclusions which you can see here: kzbin.info/www/bejne/bquzm2CkZ5h1as0 What is not clear though to me is what ouptut you are looking for. If they objective is mechanical properties (since you have the elastomer in the system), then the approach presented here can be morphed directly into your system. However, such magnetorhelogical system are often looking at some electrical energy exchange, in which case, you will have to change your analysis type from a mechanical investigation to an electromagnetic investigation. ABAQUS does both but here is where the different between my video and your problem may lie. Let me know if I have interpreted the problem correctly.

[@joseluischanmayen7067]

@@MichaelOkereke Thank you for kind answer! The output that I'm looking for is to study the behavior (in terms of stiffness and transmissibility of energy) of the material when a magnetic field is applied. So I'm first trying to create the RVE and I think the most difficult part would be to couple the magnetic and the mechanical physics in abaqus, do you have some tutorial coupling or using 2 different types of physics that can help me with it? Regards from Mexico!

[@MichaelOkereke]

Hello @Jose Luis Chan Mayen, apologies for just getting to this query now. I must have missed it. On coupling of magnetic and mechanical physics in ABAQUS, I do not have any tutorial on that but this is something that one can look into and I believe it can be done. I just have not done it. Good luck with the work.

[@MultiSaded]

Thank you for this video, I have a question, in the case of an elastoplastic material (matrix or inclusions), what are the added properties and are the periodic conditions the same as in the case of an elastic material?

[@MultiSaded]

@@taop_vc For PBC Bondary condition, is the same for the two cases (elastic or elastoplastic material)?

[@MichaelOkereke]

Hello Saad, yes, the boundary conditions will have some slight influence on predictions but essentially you should get the same mechanical behaviour. So, whether it is elastic or elastoplastic, you will get the expected mechanical behaviour. The PBC will also help you recover a more realistic behaviour with error arising from choice of boundary conditions reduced. Bear in mind, this is especially true for micromechanics studies. Well done for your response @Vincenzo Cucumazzo - really helpful.

[@MichaelOkereke]

I agree completely Vincenzo. With the plastic flow curve, typically (for simple cases) what ABAQUS requires is just the yield stress value and plastic strain value (always set at zero as an initial point). If you have a material that shows further post-yield behaviour (softening or hardening), then you need to extract the stress versus plastic strain plots which is what Vincenzo has described as a plastic flow curve. I can make a video in future showing how this is done, if there is enough interest (please like the comment) to this.

[@MultiSaded]

Thank you very much for your help, do you have a matlab file or other code to determine the analytical models (Mori-Tanaka and Self-consistent) for the two-phase case?

[@MichaelOkereke]

Not really Saad.

[@joonabil]

Very interesting video but I wonder why you didn’t use Periodic Boundary Condition (PBC) instead of using the normal boundary conditions, i.e. roller support as you are modelling RVE. Also I would like to know why you used the equation constraint to link the reference point (RP) to the set instead of using coupling constraint. It would be nice if you could let me know what is the difference between coupling constraint and equation constraint. Thanks in advance

[@MichaelOkereke]

Great point @John Nabil! Either PBC or Dirichlet boundary condition will apply anyway. It is harder to apply PBCs for 3D systems because you need to make sure you match the nodes from one face to another hence the meshing should be periodic. It could be done but I decided to do a solution that is easiest for most users of ABAQUS without any dependence on a bespoke code for creating the PBCs for a 3D RVE. I always prefer to use the *Equation constraint as it is better controlled that the coupling constraint. When you use a kinematic coupling, it forces a set of nodes to follow the behaviour of a single point and it is usually a one-to-one correspondence situation. The *Equation allows the user to specify a canonical equation (non-standard deformation behaviour) which you want to impose on the domain. Both do similar things, but *Equation is better controlled and can be used for more advanced studies/control of deformation to the domain.

[@joonabil]

@@MichaelOkereke Thanks a lot for your helpful explanations. Is that means the normal boundary conditions, i.e. roller or fixed supports called Dirichlet boundary condition. Also what do you mean by a bespoke code

[@horming_official]

Thanks you so much Dr. Okereke for these great videos. There is a question, I am intrested in a the same structure (soft matrix and rigide particles) but with a higher volume fraction (above 40%) and large deformation (NLGEOM). Abaqus cannot converge after serveral increments at a small strain (around 1-2%), but I really want to go further. This may be due to excessive element distortion. Did you have a similar problem?

[@MichaelOkereke]

Dear, thanks for your interest in the model. I did not have the problem you described because my matrix is not as 'soft' as yours. When the matrix mismatch between matrix and rigid particles is so high, then I have see this sort of problem you describe occur. It is quite commonplace. You are doing the right thing by switching on NLGEON. I suggest also that you may try smaller meshes, explore doing a Dynamic explicit study (to help with the time steps) and just stick at it. It might also not be possible to achieve your large deformation due to the excessive distortion but I suggest you keep trying all possible things until you get it working. If you sign up to my CM Videos Insider Group, you can contact me directly to see if there is more help I can offer you.

[@samadhandeshmukh8730]

Hello Sir. Really nice videos. Really appreciable!!! I have some requests/queries. 1) Please let us know about the way in which you found positions of those centers of particles (Monte Carlo implementation video you discussed about). Please put a video regarding the same, if possible. 2) If I take Vf value as 5% or 10% (Let’s say for TiB2 particles), then no of particles will increase for size of 60*60*60. Please let us know about the way in which we can put those particles in matrix in faster way. We can put 5,10 particles manually in quick time. But I guess it will take time for a greater number of particles. Any code available? 3) Any reason why Young's Modulus value came less in that final graph you showed? Any calculation errors? Thanking you in advance...

[@MichaelOkereke]

Hello Samadhan, Thanks for the interest in the channel. I will attempt to provide some answers. 1.) I am committed to making a series of video on the Monte Carlo implementations leading up to releasing a software to automate the process for Abaqus. I will say, keep coming back or subscribe so you get notified when they become available. 2.) You are right about the time constraints for multiple particles system. A script of some kind will bring automation but you also need the Monte Carlo scheme in place to help with the randomization. So again a video about this is coming soon. You can start with the script I provided in the description section and modify it for your system but some prior experience with Python language is needed. 3.) On the Young'sModulus being wrong, yes, there was some errors which I addressed in the pinned comments under the video. I will try and make a dedicated video on calculating the Modulus correctly next time. Thanks.

[@samadhandeshmukh8730]

@@MichaelOkereke Thank you so much Sir for your prompt response. I am looking forward to check video on Monte Carlo implementations.. I will check the script provided. Thanks again..

[@meena576]

Would you like to make a video RVE including for the multiscale? For example micro-scale to a meter or nanoscale to millimetre scale.

[@MichaelOkereke]

It is difficult to make a simulation across such length scales. It will not be computationally efficient. This is why you choose and RVE which should be representative of the bulk at higher scales and use same for simulation to determine properties of the bulk.

[@ayushjaiswal242]

Sir, I created 41 of those spheres but when merging Abaqus shows "merge operation was unsuccessful", any help?

[@MichaelOkereke]

Hello @Ayush Jaiswal, I am not sure what you are doing wrong. The merging operation is quite a straight forward process. I was wondering if you selected all the spheres and the matrix (as one set) and merge them. Please join my CM Vidoes Insiders group so that you can share more specifically with me what you are doing. Here is the link: cmvideos.ck.page/16a99bd0e3

[@ayushjaiswal242]

@@MichaelOkereke Is there any way I can send you my script (with the excel sheet containing the center of spheres) to help me with this easily? I would be a great help.

[@ayushjaiswal242]

@@MichaelOkereke If the merge operation is not successful, can I replace that with some 'interaction' between the cube and sphere?? If so, what type of interaction should I go with?

[@subhadeepbakshi8810]

Hey Doc great work !! I need your help on RVE modeling of dual phase TRIP steel This is very important for my job Please help me

[@MichaelOkereke]

Hello Subhadeep, thanks for your interest in the chnnel. Unfortunately, I do not have any experience modellingTRIP steel so at the moment uhnable to help you.