I'm from Peru, I don't know English, but I thank you for sharing your knowledge. You are a great teacher and I would like to learn more from you, please don't stop uploading videos.
Thanks and glad you found it helpful from Peru. I will continue to do my best and keep suporting a global community of learners embrace and engage in computational modelling.
Appreciate the kind words! It’s nice to know that my attempts at making ABAQUS less of a mystery are paying off-like a well-structured finite element analysis!
Dear Sir, You are doing really amazing job by sharing yours expertise on RUclips. I am doing project in fracture mechanics and this video give me deep insights about the theory and its application in abacus . Love and respect from India.
thank you for this invaluable video and for sharing this knowledge . I have a questions. in the triaxiality parameter, if we had not only stress in one direction but in the different directions even shear stresses, how should we calculate the parameter? you mentioned that the parameter is the trace of the stress tensor. but what if we are in a condition that we should take the shear stresses in account? thank you in advance.
Sir, thank you very much for the brilliant video. Could you tell me please, where I can get this table with the plastic properties (yeld stress, plastic strain) ? I need this for my titanium implant project.
Glad you enjoyed it and felt that way about it. I wanted to make a comprehensive video about ductile damage that tries to show the whole theory and simulation setup for viewers. Glad it struck a cord with you!
@@MichaelOkereke Thank you for your great effort. I was wondering if the method you described can be applied to truss elements like T3D2 or beam elements like B31.
@@MichaelOkereke I am wondering if there is a different approach. This would be significantly important when modelling actual-size reinforced concrete subjected to severe loading such as cyclic load. In this case, the rupture of the reiforcing bar could be identified at critical locations (plastic hinge locations). Anyway, thank you, Dr. Michael Okereke, for your reply
I have a particular interest in modeling superplasticity of metals, maybe use some of your syntactic RVEs to investigate the effect of the voids on the overall behavior of the material!!
Yes, that is fine. I think you can also look at the video about Triaxiality - some of the ideas in that video might help you with the superplasticity argument. Well done!
Thank you for the valuable video! Your work is incredibly helpful, and I truly appreciate it. I just wanted to mention one small detail regarding the density value of steel in the SI(mm) unit system in Abaqus. While you correctly converted the density from kg/m³ to kg/mm³ in this video, Abaqus requires a consistent set of units. When using millimeters (mm) for length, the corresponding units for mass, stress, energy, and density should be tonne, MPa, mJ, and tonne/mm³, respectively. Unfortunately, this is a fixed requirement in Abaqus and cannot be adjusted.
Thank you professor! Your video helps me a lot! But I got confused about which one determines the fracture strain of the simulation? Is Fracture Strain?Or Displacement at Failure? Thanks a lot!
can you compare this failure theory with the traction separation theory? when should we use that one? when do we use Cohesive elements? Thanks for your video.
I am not sure about comparison of the failure theory and the traction separation theory. I know cohesize zone analyis tends to be required where there is interface/contact effect which need to separate as part of a damage event. This is not the case in this example.
Yes, because the mesh is linked ot the characteristic length. You can play around with it and see if it makes any difference but theoretically, yes you need to.
Hi . I want to simulate tensile test for PLA polymer materials. I don't know the damage criteria that the sample will break in the tensile test. Thank you for your guidance
Most of the time it is best to find those in a journal paper. I do not have anyone in mind yet but a quick search on say Google Scholar can help you find papers that deal with PLA and use ductile damage. It might be slightly different from what we showed here (since you are dealing with polymers as against metals), so you have to do some reading around to find what you need.
Yes, with the energy approach, you will do similar things i.e. determine the characteristic length but this time you use energy to make the deductions. I do not have a video about this but it could be a plan for the future.
You would need to include a damage model yo your material model. I can suggest you consider this video: ruclips.net/video/gpSh2KLWnqk/видео.htmlsi=cNq4WBvj7EHVPlAL
I had a thought about it and the simplest way is similar to what I did here but replace the volume of the cubic element by the volume of a tetrahedron shape, but still take the cube root of the volume of the tetrahedron. Take volume of a tetrahedron = a^3/(6*square-root(2)) where a = edge length of a regular tetrahedron. I think you could assume that the elements you are dealing with are regular tetrahedrons where in more cases it would be irregular but deviations from edge length, 'a' can become an error measure which you can factor in in determining the characteristic length.
The simplest way is similar to what I did here but replace the volume of the cubic element by the volume of a tetragonal element, but still take the cube root of that volume. Take volume of a tetragon = a^2c where a = edge length of a regular square base of the tetragon and c is height. I think you could assume that the elements you are dealing with are regular tetragons even though in more cases it would be irregular but deviations from edge length, 'a' and 'c' can become an error measure which you can factor too in determining the characteristic length.
Thank you immensely for your thorough explanation! 🌟 However, I have a question: Why use an explicit solver? What advantages does it bring? If the problem is tackled using Abaqus Standard, what changes? To my knowledge, Abaqus Explicit is typically essential when a problem encounters significant forces due to mass or intricate frictional forces-neither of which seem pivotal in this case. What, then, justifies its use here?
You are right... it could be easily done with ABAQUS Implicit and it would work. However whenever there is significant possibility of nonlinearity, damage, impact or the examples you cited, it is usually recommended to include an ABAQUS Explicit step as its better suited to deal with the convergence of such unstable simulations. I hope it makes sense.
excuse me, Can I ask some question that is not relative with this video ? My model is hybrid composite between KFRP and CFRP under low velocity impact condition but I cannot get force value or plot force and time graph. Can you explain how to I solve this problem.
I think this is possible. YOu need to track a set of node/nodes from which you extract forces or displacements. You have to kinematically linked a reference point to a section of the material. I have a few videos on this channel on RVE modellign where I showed this kinamtically linked *EQUATION approach of using a reference point node to load and extract properties from a model.
Sir, thanks for your sharing. It seems that you explain the initial point of damage，but we dont define the "initial point" in ABAQUS.....So what parameters the "fracture strain" in ABAQUS UI should be defined?
I define the fracture strain as the point of perceived fracture of the specimen. If you look at the video, you can see the argument I used in identifying and specifying it.
hello sir, thank you for your clean explanation. but there is something i want to share with you. based on ABAQUS documentation, "Fracture Strain" means "Equivalent fracture strain at damage initiation". so there is a contraction between what you defined in the video and ABAQUS documentation. in another way, the strain you defined in the video, is more like "rupture strain" based on ABAQUS documentation. have I missed something ?
Thanks for the info. I believe the essence of the video is right, the specific way you or I interpret these boundary values might differ but you ahve to keep iterating your simulation until experimental data matches numerical data. This video is simply a guide for the beginner modeller in this area.
Not really...I was working in an SI unit with mass in a unit of kg. Also, the Pa is N/mm^2 and N = kg.mms^(-2). Therefore, it's better specified in kg. However, if your masses are in tonnes, then indeed you are right.
This is a good question. I have not done this before but should be interesting. What is the matrix and metallic reinforcement you are taling about. What is the hybrid material?
@@MichaelOkereke it may be aluminum matrix and TIB2 + SiC particle reinforcement. Analysis should be on damage initiation, crack propagation and interfacial debonding.
Good question! Lode parameter is definitely essential for describing plasticity of materials. I have not specified it here as the ductile damage model that comes with ABAQUS, and which was reviewed in this video does not have need of this lode parameter. Just like the stress triaxiality, the lode parameter gives and indication of the dominant stress state in the model. Lode parameter, typically represented by mu, can be: mu = 1 (pure shear stress), mu = 0 (uniaxial stress) and mu = -1 (corresponds to a stress state where principal stress, σ2 =σ3 (triaxial compression). It is similar to the stress triaxiality term which gives us an indication of the stress state (dominating) the simulation. If you want to learn a bit more about the lode parameter and the like for a triaxially loaded system, then watch this video of mine: ruclips.net/video/HJ3JzPUhz1U/видео.html).
Sir, in your video you refer the "Fracture Strain" set in ABAQUS as the value when it totally damaged. I think "Fracture Strain" in ABAQUS might be the fracture strain at damage initiation, which you called epsilon zero.
Was going to ask this as well. Because I believe equivalent plastic displacement is then used to get from this value to the point of failure in the damage evolution module.
Hi, good question. If you look at the graph, the initiation aof damage is at the UTS and there D = 0, which suggests that damage has started and this will evolve with changing D until D = 1 (complete damage). Complete damage occurs at the fracture/separation of the specimen. The strain at which this happens is the fracture strain. It could not be at the epsilon_0 (which is equivalent to start of the plastic strain), as this corresponds to the UTS. The specimen cannot fracture at this point rather it will be the final point of yielding. Eventual fracture/separation of the specimen happens at what I have called fracture strain. Within a single element, when damage initiates, we need to know when a failed element will be deleted from the model. This corresponds to the fracture strain through the displacement at failure. Once this is obeyed, the element is then deleted from the model since element deletion is swithced on. Other elements around the failed element might have D values not up to 1 and so not approached the fracture strain. Once they meet this condition, the elements would then fracture and be deleted from the the model. If you do not believe me, I asked ChatGPT to define fracture strain and here is the answer: "In the context of continuum damage mechanics, fracture strain is defined as the strain at which a material undergoes fracture, leading to the complete loss of load-carrying capacity in the material. It is The strain value at which a material experiences fracture due to the accumulation of damage, resulting in a complete loss of its structural integrity." I believe this is consistent with what I have tried to explain above.
@@MichaelOkereke Thank you sir, that's very clear. But in Abaqus manual Ductile Damage Section, the "Fracture Strain" is defined as Equivalent fracture strain at damage initiation, not the end. I tested it in abaqus and plotted the stress-strain curve, and the "Fracture Strain" input value is really just the initial failure strain.
I think the displacement at failure is calculated wrong. It should be the product of the characteristic length and (plastic damage failure strain and plastic damage initiation strain). Per Abaqus keyword documentation: Data lines to specify damage evolution for TYPE=DISPLACEMENT, SOFTENING=LINEAR without the MIXED MODE BEHAVIOR and the RATE DEPENDENT parameters First line Effective total or plastic displacement at failure, measured from the time of damage initiation. (Units of L.)
Greetings sir! Can you share your contact details pls may be email or something, so that I can contact you , I know you have a very good understanding of abaqus, I need some help or guidance on model which I am doing. Pls sir can you help 🙏🥺
Hello, thanks for your interest. Normally, the best way is to subscribe to my CM Videos Insider group here: cmvig.cmvideos.org. You will get a welcome email and then contact me via that. I support people who come through the Insider Group.
I'm from Peru, I don't know English, but I thank you for sharing your knowledge. You are a great teacher and I would like to learn more from you, please don't stop uploading videos.
Thanks and glad you found it helpful from Peru. I will continue to do my best and keep suporting a global community of learners embrace and engage in computational modelling.
all my appreciation for your great effort doctor, really great explaination combining theory and practice in ABAQUS
Appreciate the kind words! It’s nice to know that my attempts at making ABAQUS less of a mystery are paying off-like a well-structured finite element analysis!
Dear Sir, You are doing really amazing job by sharing yours expertise on RUclips. I am doing project in fracture mechanics and this video give me deep insights about the theory and its application in abacus .
Love and respect from India.
Thanks. Glad it helped.
i was looking for this topic from last sic month, finally, i got it all. Thank you very much sir
Aww... glad you found it at last. Thank you for sharing this.
thank you for this invaluable video and for sharing this knowledge . I have a questions. in the triaxiality parameter, if we had not only stress in one direction but in the different directions even shear stresses, how should we calculate the parameter?
you mentioned that the parameter is the trace of the stress tensor. but what if we are in a condition that we should take the shear stresses in account?
thank you in advance.
Sir, thank you very much for the brilliant video.
Could you tell me please, where I can get this table with the plastic properties (yeld stress, plastic strain) ? I need this for my titanium implant project.
Thank you professor, good lesson!
Thanks
excellent and very informative video
Glad you liked it
one of good explanations for the ductile damage so far
Glad you enjoyed it and felt that way about it. I wanted to make a comprehensive video about ductile damage that tries to show the whole theory and simulation setup for viewers. Glad it struck a cord with you!
@@MichaelOkereke Thank you for your great effort. I was wondering if the method you described can be applied to truss elements like T3D2 or beam elements like B31.
I will not think so, as the element type used here are 3d continuum elements so am not sure you can use it for the beam or truss elements.
@@MichaelOkereke I am wondering if there is a different approach. This would be significantly important when modelling actual-size reinforced concrete subjected to severe loading such as cyclic load. In this case, the rupture of the reiforcing bar could be identified at critical locations (plastic hinge locations). Anyway, thank you, Dr. Michael Okereke, for your reply
Excellent video!! Keep them coming Dr. Okereke
Thanks very much. I will keep going. Any suggestions for future videos?
I have a particular interest in modeling superplasticity of metals, maybe use some of your syntactic RVEs to investigate the effect of the voids on the overall behavior of the material!!
Yes, that is fine. I think you can also look at the video about Triaxiality - some of the ideas in that video might help you with the superplasticity argument. Well done!
Thank you for the valuable video! Your work is incredibly helpful, and I truly appreciate it. I just wanted to mention one small detail regarding the density value of steel in the SI(mm) unit system in Abaqus. While you correctly converted the density from kg/m³ to kg/mm³ in this video, Abaqus requires a consistent set of units. When using millimeters (mm) for length, the corresponding units for mass, stress, energy, and density should be tonne, MPa, mJ, and tonne/mm³, respectively. Unfortunately, this is a fixed requirement in Abaqus and cannot be adjusted.
Finally best explanation of this topic, thanks
Glad it was helpful! That is what my intention was and I am glad that vision is achieved. Cheers!
Dear Sir, can you clarify why you choose a tabular amplitude over ramp in loading step?
Thank you so much for the nice and detailed explanation
You are welcome.
@@MichaelOkereke Do you have any material for shear failure of concrete in ABAQUS? Any response would be highly
appreciated.
Can i use the property of ductile damage with concrete or it just use with metal materials?
Sir, very informative video. How to get stress strain curve at different strain rates?
You run the simulations for different strain rates typically different velocities and then follow same steps as in this video for every simulation.
Thank you very much sir
Most welcome
Thank you for your help
Thanks
Thank you professor! Your video helps me a lot! But I got confused about which one determines the fracture strain of the simulation? Is Fracture Strain?Or Displacement at Failure? Thanks a lot!
Hi
I would like to ask you how to extract residual stress from a section in Abaqus from the ODB files? Thank you
You need to read the odd file with a python script and see stress in select elements.
can you compare this failure theory with the traction separation theory? when should we use that one?
when do we use Cohesive elements?
Thanks for your video.
I am not sure about comparison of the failure theory and the traction separation theory. I know cohesize zone analyis tends to be required where there is interface/contact effect which need to separate as part of a damage event. This is not the case in this example.
Thank you for the video. So if we change the mesh size do we also need to change the displacement at failure parameter?
Yes, because the mesh is linked ot the characteristic length. You can play around with it and see if it makes any difference but theoretically, yes you need to.
Hi .
I want to simulate tensile test for PLA polymer materials. I don't know the damage criteria that the sample will break in the tensile test. Thank you for your guidance
Most of the time it is best to find those in a journal paper. I do not have anyone in mind yet but a quick search on say Google Scholar can help you find papers that deal with PLA and use ductile damage. It might be slightly different from what we showed here (since you are dealing with polymers as against metals), so you have to do some reading around to find what you need.
Great video, could you explain something about the energy approach of the fracture evolution, fracture energy.
Yes, with the energy approach, you will do similar things i.e. determine the characteristic length but this time you use energy to make the deductions. I do not have a video about this but it could be a plan for the future.
Please Sir tell how to break porous metal material in abaqus
You would need to include a damage model yo your material model. I can suggest you consider this video: ruclips.net/video/gpSh2KLWnqk/видео.htmlsi=cNq4WBvj7EHVPlAL
Sir, what will be the characteristic length for a tetrahedral element?
I had a thought about it and the simplest way is similar to what I did here but replace the volume of the cubic element by the volume of a tetrahedron shape, but still take the cube root of the volume of the tetrahedron. Take volume of a tetrahedron = a^3/(6*square-root(2)) where a = edge length of a regular tetrahedron. I think you could assume that the elements you are dealing with are regular tetrahedrons where in more cases it would be irregular but deviations from edge length, 'a' can become an error measure which you can factor in in determining the characteristic length.
How to find the characteristic length for tetragonal element?
The simplest way is similar to what I did here but replace the volume of the cubic element by the volume of a tetragonal element, but still take the cube root of that volume.
Take volume of a tetragon = a^2c where a = edge length of a regular square base of the tetragon and c is height.
I think you could assume that the elements you are dealing with are regular tetragons even though in more cases it would be irregular but deviations from edge length, 'a' and 'c' can become an error measure which you can factor too in determining the characteristic length.
Really good video.
For fracture evolution, if I use the energy option, does the fracture energy inputted need to multiplied by the element size?
Yes, correct i.e. you need to determine the characteristic length but it is not so straightforward but I sort of get what you are asking.
Thank you immensely for your thorough explanation! 🌟
However, I have a question: Why use an explicit solver? What advantages does it bring? If the problem is tackled using Abaqus Standard, what changes? To my knowledge, Abaqus Explicit is typically essential when a problem encounters significant forces due to mass or intricate frictional forces-neither of which seem pivotal in this case. What, then, justifies its use here?
You are right... it could be easily done with ABAQUS Implicit and it would work.
However whenever there is significant possibility of nonlinearity, damage, impact or the examples you cited, it is usually recommended to include an ABAQUS Explicit step as its better suited to deal with the convergence of such unstable simulations. I hope it makes sense.
excuse me, Can I ask some question that is not relative with this video ?
My model is hybrid composite between KFRP and CFRP under low velocity impact condition but I cannot get force value or plot force and time graph. Can you explain how to I solve this problem.
I think this is possible. YOu need to track a set of node/nodes from which you extract forces or displacements. You have to kinematically linked a reference point to a section of the material. I have a few videos on this channel on RVE modellign where I showed this kinamtically linked *EQUATION approach of using a reference point node to load and extract properties from a model.
Sir, thanks for your sharing. It seems that you explain the initial point of damage，but we dont define the "initial point" in ABAQUS.....So what parameters the "fracture strain" in ABAQUS UI should be defined?
I define the fracture strain as the point of perceived fracture of the specimen. If you look at the video, you can see the argument I used in identifying and specifying it.
hello sir, thank you for your clean explanation. but there is something i want to share with you.
based on ABAQUS documentation, "Fracture Strain" means "Equivalent fracture strain at damage initiation". so there is a contraction between what you defined in the video and ABAQUS documentation. in another way, the strain you defined in the video, is more like "rupture strain" based on ABAQUS documentation. have I missed something ?
Thanks for the info. I believe the essence of the video is right, the specific way you or I interpret these boundary values might differ but you ahve to keep iterating your simulation until experimental data matches numerical data. This video is simply a guide for the beginner modeller in this area.
Great video! One thing you might have missed is converting thr density to tonne/mm^3 as opposed go kg/mm^3.
Not really...I was working in an SI unit with mass in a unit of kg. Also, the Pa is N/mm^2 and N = kg.mms^(-2). Therefore, it's better specified in kg. However, if your masses are in tonnes, then indeed you are right.
Thank you for this video, can you give us an example on the damage of an RVE as an example: Concrete with aggregats or Porous material.
Good question. I keep planning to make such video. Please watch the space.
please do a video on hybrid metal matrix composite RVE with fracture analysis
This is a good question. I have not done this before but should be interesting. What is the matrix and metallic reinforcement you are taling about. What is the hybrid material?
@@MichaelOkereke it may be aluminum matrix and TIB2 + SiC particle reinforcement. Analysis should be on damage initiation, crack propagation and interfacial debonding.
basically for understanding the deformation and fracture mechanism of MMCs.
What about Lode Parameter ?
Good question!
Lode parameter is definitely essential for describing plasticity of materials. I have not specified it here as the ductile damage model that comes with ABAQUS, and which was reviewed in this video does not have need of this lode parameter. Just like the stress triaxiality, the lode parameter gives and indication of the dominant stress state in the model.
Lode parameter, typically represented by mu, can be: mu = 1 (pure shear stress), mu = 0 (uniaxial stress) and mu = -1 (corresponds to a stress state where principal stress, σ2 =σ3 (triaxial compression). It is similar to the stress triaxiality term which gives us an indication of the stress state (dominating) the simulation.
If you want to learn a bit more about the lode parameter and the like for a triaxially loaded system, then watch this video of mine: ruclips.net/video/HJ3JzPUhz1U/видео.html).
Sir, in your video you refer the "Fracture Strain" set in ABAQUS as the value when it totally damaged. I think "Fracture Strain" in ABAQUS might be the fracture strain at damage initiation, which you called epsilon zero.
Was going to ask this as well. Because I believe equivalent plastic displacement is then used to get from this value to the point of failure in the damage evolution module.
Hi, good question. If you look at the graph, the initiation aof damage is at the UTS and there D = 0, which suggests that damage has started and this will evolve with changing D until D = 1 (complete damage). Complete damage occurs at the fracture/separation of the specimen. The strain at which this happens is the fracture strain.
It could not be at the epsilon_0 (which is equivalent to start of the plastic strain), as this corresponds to the UTS. The specimen cannot fracture at this point rather it will be the final point of yielding. Eventual fracture/separation of the specimen happens at what I have called fracture strain.
Within a single element, when damage initiates, we need to know when a failed element will be deleted from the model. This corresponds to the fracture strain through the displacement at failure. Once this is obeyed, the element is then deleted from the model since element deletion is swithced on. Other elements around the failed element might have D values not up to 1 and so not approached the fracture strain. Once they meet this condition, the elements would then fracture and be deleted from the the model.
If you do not believe me, I asked ChatGPT to define fracture strain and here is the answer:
"In the context of continuum damage mechanics, fracture strain is defined as the strain at which a material undergoes fracture, leading to the complete loss of load-carrying capacity in the material. It is The strain value at which a material experiences fracture due to the accumulation of damage, resulting in a complete loss of its structural integrity."
I believe this is consistent with what I have tried to explain above.
@@MichaelOkereke Thank you sir, that's very clear. But in Abaqus manual Ductile Damage Section, the "Fracture Strain" is defined as Equivalent fracture strain at damage initiation, not the end. I tested it in abaqus and plotted the stress-strain curve, and the "Fracture Strain" input value is really just the initial failure strain.
In which case you go with ABAQUS... the documentation will not lie.
I think the displacement at failure is calculated wrong. It should be the product of the characteristic length and (plastic damage failure strain and plastic damage initiation strain). Per Abaqus keyword documentation:
Data lines to specify damage evolution for TYPE=DISPLACEMENT, SOFTENING=LINEAR without the MIXED MODE BEHAVIOR and the RATE DEPENDENT parameters
First line
Effective total or plastic displacement at failure, measured from the time of damage initiation. (Units of L.)
You could be right... I need to check carefully... the principle remains though.
perfect!
thank you so much sir
Most welcome
Great video. I am though wondering if it is possible to do the same simulation (with the same ductile damage model) in ANSYS workbench as well.
Yes, absolutely.
When using mm, I believe your density needs to be tonne/mm^3, not kg/mm^3. In a sense you are applying a mass scaling of 10^3 when you use kg/mm^3
Not really, if you look at my analysis of the density, it shows it is right. There was no mass scaling.
Greetings sir! Can you share your contact details pls may be email or something, so that I can contact you , I know you have a very good understanding of abaqus, I need some help or guidance on model which I am doing. Pls sir can you help 🙏🥺
Hello, thanks for your interest. Normally, the best way is to subscribe to my CM Videos Insider group here: cmvig.cmvideos.org. You will get a welcome email and then contact me via that. I support people who come through the Insider Group.
@@MichaelOkereke You are very great man sir ! I expect help from you 🙏🙏🥺 thank you for your kindness. God bless you, keep inspiring us.
e_pl_true = e_true - S_true / E
You have drawn a line at S_UTS that exactly shows the relation above
Yes in deed... sort of!
Damage initiation criteria asks for e_fracture that is e_pl at UTS not epsilon at specimen fracture. You reach that giving the plastic displacement