For circular particles, as I did in this video, was done using a bespoke software I developed called MontCarlGen2D, which is obtained via this link: www.cmvideos.org/shop/ For the grains, that is different, and I don't have any video or resources to help.
@@MichaelOkereke Thanks for the reply.🙂 I want to purchase this one. Just have some questions: 🤔 How Can I use that program in ABAQUS ? Is it automatically connected to ABAQUS to generate random particles ? and when do you release a 3D version? Generating particles in 3 dimension . 🙋♂
There is a documentation that comes with the code that explains how to use the code. The 3D version will be released soon and it will be in 3 dimensions.
The video is very interesting and helpful but what happens when the nanoparticles cannot be detected in the matrix. For example if the nanoparticles melt and combine with the matrix. How would one model that?
Hello @Anastasis Georgiou, thanks for the comments. I think the case you describe will be modelled as I have done here except that you will have to incorporate a constitutive/material behaviour that allows for the nanoparticles to melt and combine with the matrix. I will imagine incorporating a temperature dependent behaviour of the nanoparticles that allows it to melt during the simulation. That is my suggestion.
Nice and informative video. Sir i want to know that how we can perform the tensile strength simulation of the woven nanocomposite? One thing i thought that firstly we will compute homogenized properties in microscale rve( fibre and matrix) and then move towards meso scale rve (microscale homogenized properties are provided to yarn in mesoscale) and compute homogenized properties in mesoscale. After that, move towards macroscale and make a laminated woven composite with each lamina properties as mesoscale rve. After that how can we simulate the tensile strength of that woven composite so that my simulation stress strain curve matches with experimental. Thanks in advance.
Hi, thanks for the query. I think your plan sounds right. It is the whole principle of multiscale modelling of heterogeneous material where you transfer model predictions from the lower lengthscale (microscale) to structural (macroscale) lengthscales. I think your suggestion is the right way to go. With predicting strengths (in tension), you have to extract the homogenized stress-strain plots from your simulation and see if this compares with experiment. It seems to me your experiment will be at macroscale so your predictions need to be at that scale too, if you do not want to work with a micromechanical model. If you however want to go the RVE modelling route, then you need to (a) isolate the right RVE that captures your woven composite best (b) Use adequate/appropriate boundary condition to simulate tensile deformation (see any of the RVE modelling videos on this chanened) and (c) extract a homogenized stress strain. This will be what you ca then compare with your experiments. I hope these suggestions helpd.n
@@MichaelOkereke sir, Thanks for the detailed and descriptive reply to my query. However, If I want the tensile stress-strain plot of the rectangular strip (as per ASTM D3039 standards) made from woven composite laminates, then how can I model the damage so that the plot falls down after some deformation otherwise it will keeps going on increasing. One model as per my opinion that is commonly employed for damage in woven composites is the Hashin damage model, but what are the input parameters to be kept in that model for numerical simulation. Also, if my mesoscale model has a thickness that is less than the rectangular strip sample made as per ASTM standard, then in that case is I have to model as per lamina layer wise and provide each lamina property as the homogenized property coming from mesoscale?
For capturing the right behaviour of the specimen as seen in experimental data, you would of course use some form of damage model. The Hashin is a good one. I will ask that you search through literature to find the right Hashin damage properties for your material. I do not have it. For your mesoscale question, it is also normal to use a lamina-by-lamina approach but since this is a woven composite, it becomes tricky. I think you have to determine what a good representative volume for the problem is and probably look at how this is done in literature. I do not know the specifics to be able to give the right answer.
It is an interesting video and I would like to ask you something more related to the second part of this video but I can't comment there as I didn't find the comment section. Is there is any way to get strain stress curve directly from Abaqus instead of getting displacement and force and turn them to strain stress curve by Excel as you have shown in the second video. Thanks in advance
Definitely John, you can get stress and strain data directly by creating a history output based on a defined nodal set. For example, if you want to extract the complete stress and strain data for your full model, then you create a nodal set of all elements. Select that nodal set in History Output section in ABAQUS and select which ones you want to track the history. This is definitely possible. I may have to make a dedicated video to address this specific requirement, if there is enough like on this comment.
@@MichaelOkereke Do you mean I shall extract the stress and strain for all mesh elements in my model then I shall sum them to get the stress strain curve which represents the full model?
In this video no any small-size effect is considered at nano/micro scale. In other words, this computational model can be employed directly for a continuous structure (beam/bridge) at macro scale. Please provide clear explanations on how the nanoscale effects of the CNTs have been taken into account in the current FEM-based approach
Dear Narendra, thanks for your interest in the channel. I am not fully sure I understand the question but my best guess is that it is about the length scale dicotomy of creating conclusions at microscale and translating that to macroscale to bridges/beams as you suggest. If this is so, then this model is simply a microscale model. To bridge the length scale gap between microscale and macroscale, there are existing frameworks for doing this and this is not covered in this video. It is quite commonplace in FEM studies to develop solutions at representative length scales and extend such conclusions to high lengthscale but there should be good care taking while doing so. I hope this response addresses your query.
@@MichaelOkereke Sir, can you please guide me how to use in microscale in this software, since I model the nanocomposite with particulate in microns in size. I currently used in mm scale but wanted to do in micron scale.
hi sir i am mechniacal engg and also m tech in material science i want learn more about nano material composite design and fem analysis can tell more about software and tools
Hello @Parashiva, thanks for the comment. I do have a few videos on nanocomposites or composites generally. i want to continue growing the library of videos I have about nanocomposites. In the mean tim, I think with FEM, just any good FEM book will help. If you like the way I see FEM, you may then like to see my textbook about FEM from this link: link.springer.com/book/10.1007/978-3-319-67125-3. Any similar textbook on micromechanics of composites (nanocomposites) will be okay. Good luck with the research.
Hello Prof., i have a RVE (CNT and Epoxy) of dimension of 100*20*20 (nm) . CNT-1000GPa and Epoxy 3.2 GPa But in Abaqus for developing the geometry while sketching if i gave 100e-9, 20e-9, 20e-9, the figure is not displaying. (As it is too small) . So i tried with just 100*20*20 units for sketching. So gere how should i adjust the material property ?
Yes, we can - just need to find properties of the graphene nano-particles, and select an adequate RVE size to make the results representative. If you can give me an idea of what you want modelled, I can consider that.
Dr Okereke; 🙋♂
How to generate a random distribution of particles and grains in a microstructure (min 7:00 ) . Thank you for the great videos . 👌🙏
For circular particles, as I did in this video, was done using a bespoke software I developed called MontCarlGen2D, which is obtained via this link: www.cmvideos.org/shop/
For the grains, that is different, and I don't have any video or resources to help.
@@MichaelOkereke
Thanks for the reply.🙂 I want to purchase this one. Just have some questions: 🤔
How Can I use that program in ABAQUS ? Is it automatically connected to ABAQUS to generate random particles ?
and when do you release a 3D version? Generating particles in 3 dimension .
🙋♂
There is a documentation that comes with the code that explains how to use the code. The 3D version will be released soon and it will be in 3 dimensions.
Dear Dr. Michael Okereke, have you done any modelling of the metal matrix composite using cohesive elements on the interface? thanks
If not, what is the default bond in ABAQUS between the particle and the matrix in this video? thanks
No... I haven't.
For this video, it was perfect bonding between the materials. It should give you the upper bound predictions.
The video is very interesting and helpful but what happens when the nanoparticles cannot be detected in the matrix. For example if the nanoparticles melt and combine with the matrix. How would one model that?
Hello @Anastasis Georgiou, thanks for the comments. I think the case you describe will be modelled as I have done here except that you will have to incorporate a constitutive/material behaviour that allows for the nanoparticles to melt and combine with the matrix. I will imagine incorporating a temperature dependent behaviour of the nanoparticles that allows it to melt during the simulation. That is my suggestion.
thank you Dr Michael , why its not fail or cracked ?
It's because I didn't use a damage model here... only plastic deformation can be seen.
@@MichaelOkereke thanx :)
Nice and informative video. Sir i want to know that how we can perform the tensile strength simulation of the woven nanocomposite? One thing i thought that firstly we will compute homogenized properties in microscale rve( fibre and matrix) and then move towards meso scale rve (microscale homogenized properties are provided to yarn in mesoscale) and compute homogenized properties in mesoscale. After that, move towards macroscale and make a laminated woven composite with each lamina properties as mesoscale rve. After that how can we simulate the tensile strength of that woven composite so that my simulation stress strain curve matches with experimental. Thanks in advance.
Hi, thanks for the query. I think your plan sounds right. It is the whole principle of multiscale modelling of heterogeneous material where you transfer model predictions from the lower lengthscale (microscale) to structural (macroscale) lengthscales. I think your suggestion is the right way to go.
With predicting strengths (in tension), you have to extract the homogenized stress-strain plots from your simulation and see if this compares with experiment. It seems to me your experiment will be at macroscale so your predictions need to be at that scale too, if you do not want to work with a micromechanical model.
If you however want to go the RVE modelling route, then you need to (a) isolate the right RVE that captures your woven composite best (b) Use adequate/appropriate boundary condition to simulate tensile deformation (see any of the RVE modelling videos on this chanened) and (c) extract a homogenized stress strain. This will be what you ca then compare with your experiments. I hope these suggestions helpd.n
@@MichaelOkereke sir, Thanks for the detailed and descriptive reply to my query. However, If I want the tensile stress-strain plot of the rectangular strip (as per ASTM D3039 standards) made from woven composite laminates, then how can I model the damage so that the plot falls down after some deformation otherwise it will keeps going on increasing. One model as per my opinion that is commonly employed for damage in woven composites is the Hashin damage model, but what are the input parameters to be kept in that model for numerical simulation.
Also, if my mesoscale model has a thickness that is less than the rectangular strip sample made as per ASTM standard, then in that case is I have to model as per lamina layer wise and provide each lamina property as the homogenized property coming from mesoscale?
For capturing the right behaviour of the specimen as seen in experimental data, you would of course use some form of damage model. The Hashin is a good one. I will ask that you search through literature to find the right Hashin damage properties for your material. I do not have it.
For your mesoscale question, it is also normal to use a lamina-by-lamina approach but since this is a woven composite, it becomes tricky. I think you have to determine what a good representative volume for the problem is and probably look at how this is done in literature. I do not know the specifics to be able to give the right answer.
It is an interesting video and I would like to ask you something more related to the second part of this video but I can't comment there as I didn't find the comment section. Is there is any way to get strain stress curve directly from Abaqus instead of getting displacement and force and turn them to strain stress curve by Excel as you have shown in the second video. Thanks in advance
Definitely John, you can get stress and strain data directly by creating a history output based on a defined nodal set. For example, if you want to extract the complete stress and strain data for your full model, then you create a nodal set of all elements. Select that nodal set in History Output section in ABAQUS and select which ones you want to track the history. This is definitely possible. I may have to make a dedicated video to address this specific requirement, if there is enough like on this comment.
@@MichaelOkereke Do you mean I shall extract the stress and strain for all mesh elements in my model then I shall sum them to get the stress strain curve which represents the full model?
that is very helpful to me
Thanks Xin for your interest in the channel.
In this video no any small-size effect is considered at nano/micro scale. In other words, this computational model can be employed directly for a continuous structure (beam/bridge) at macro scale. Please provide clear explanations on how the nanoscale effects of the CNTs have been taken into account in the current FEM-based approach
Dear Narendra, thanks for your interest in the channel. I am not fully sure I understand the question but my best guess is that it is about the length scale dicotomy of creating conclusions at microscale and translating that to macroscale to bridges/beams as you suggest. If this is so, then this model is simply a microscale model. To bridge the length scale gap between microscale and macroscale, there are existing frameworks for doing this and this is not covered in this video. It is quite commonplace in FEM studies to develop solutions at representative length scales and extend such conclusions to high lengthscale but there should be good care taking while doing so. I hope this response addresses your query.
@@MichaelOkereke thanks
@@MichaelOkereke Sir, can you please guide me how to use in microscale in this software, since I model the nanocomposite with particulate in microns in size. I currently used in mm scale but wanted to do in micron scale.
hi sir i am mechniacal engg and also m tech in material science i want learn more about nano material composite design and fem analysis can tell more about software and tools
Hello @Parashiva, thanks for the comment. I do have a few videos on nanocomposites or composites generally. i want to continue growing the library of videos I have about nanocomposites. In the mean tim, I think with FEM, just any good FEM book will help. If you like the way I see FEM, you may then like to see my textbook about FEM from this link: link.springer.com/book/10.1007/978-3-319-67125-3. Any similar textbook on micromechanics of composites (nanocomposites) will be okay. Good luck with the research.
In this, you have modeled 10 micro meter size RVE as just 10 (and not 0.01) and later while inputting material properties are you adjusting that ?
Hello @Abdul, you are right. I kept the unit as meter but modelled to scale with 10 micrometer.
Hello Prof., i have a RVE (CNT and Epoxy) of dimension of 100*20*20 (nm) .
CNT-1000GPa and Epoxy 3.2 GPa
But in Abaqus for developing the geometry while sketching if i gave 100e-9, 20e-9, 20e-9, the figure is not displaying. (As it is too small) . So i tried with just 100*20*20 units for sketching. So gere how should i adjust the material property ?
Since i am sketching x nm as x. So the elastic properties has to be expressed as N/nm2 . Is it crrct?
So if i am not wrong, 1000Gpa (if i am expressing as N/nano metre2) is equal to 1000e-9
Can we model graphene nanoparticles in carbon fiber composite using abacus
Sir Can we Modeling Of graphene nanoparticles in carbon fiber composite
Yes, we can - just need to find properties of the graphene nano-particles, and select an adequate RVE size to make the results representative. If you can give me an idea of what you want modelled, I can consider that.
@@MichaelOkereke how I can contact you
So that I can share The details with you of graphene properties
Hiii