Time stamps -------------------------------------- You can quickly navigate through the videos using these chapters. 00:00 Intro 00:40 ABAQUS Setup of Tesnile specimen case study 01:40 Approach 1: Localized stress-strain plot 06:32 Approach 2: Volume-averaged stress-strain plot 10:32 Approach 3: Experimental equivalence stress-strain plot 13:57 Analysis of experimental equivalence approach simulation results 17:36 Comparison of stress-strain plots from the 3 approaches 19:26 Outro
Your explanations regarding the tensile simulation are excellent. Could you please provide me with information regarding the strain rate or cross-speed utilized in the simulation?
Hello @Hemant, this simulation was a STATIC GENERAL simulation which means the load was applied statically. Therefore effect of strain rate is very minimal and more in the quasi-static range. For a quantitative data, I have applied a total strain of 0.30 over a time of 1 seconds. That means the strain rate is 0.3 per second. This is a quasi-static data but you have to take this with a pinch of salt. If you truly want to quantify a strain rate study, then you have to do a dynamic implicit or explicit simulation where loading effects are adequately considered.
Great job Dr. Okereke! If I want to plot Force-displacement curve of a crimping stent test to compare it with the real experiment, would it change something with respect to the third approach you described? Because I am struggling to do it. Thanks in advance!
Great suggestion! I have not tested it out but it is possible to use a reference point to link the surfaces of the stents where the radial displacement is applied. Then, using that extract the reaction force and displacement to get a force-displacement plot. I will try it out and it might work, but I cannot say for now until I try it out.
Hello Dr. Okereke, I am simulating a tensile test of CFRP using a shell element. In my first approach, I fixed one end and applied shell edge load on the other but I ran into a convergence issue. Because of that in my second approach, I applied shell edge load on both ends. I created history o/p and requested RF over the nodes on the two end edges but I got zero reaction forces although the Sim. ran successfully. Can you please advise me to follow the right approach? I also want to plot a Force vs Displ. curve Best wishes
Use the approach I use in my model using a referral point note and kinematically link it to the edge. Apply your load on that node and extract your history output from that reference point node.
Maybe you could compute the true stress - true strain curve from the experimental results. I think this new curve will fit very good with the others. Thank you. Nice video.
I think this is possible but it might not be the case here as the shoulder effect is causing the problem. If it was what you say, then the elastic section will line up together as others. Post yield differences arise from the plastic region rather than elastic section.
@@MichaelOkereke yes, you are right. I didn't take into account correctly the difference in the elastic region. As a matter of fact, I am looking for a numerical experiment to teach the difference between stress-strain and true stress-strain and I thought that your example could be one of them, but it isn't. Thank you for your answer.
Very nice way of explaining the concept sir. Sir, can we extract directly the stress strain plot from the reference point without first extracting force displacement plot from the history output as explained in 3rd experimental equivalence approach? Also, is the first two plots are true stress strain plots and the last one approach( experimental equiavence) is engineering stress strain plot?
Thanks for the comment. A really good question. You certainly cannot extract stress or strain variable from the reference point. This is because the reference point is a nodal point and stress and strain measures are variables extracted from an element set rather than a nodal set. Also, the first two plots generated from localization and volume averaging methods are true stress and strain values. To be precise the stress is a 'Cauchy stress' (an equivalent true stress) measure. It corresponds to the force per unit current area. For strain, it is a bit complicated as ABAQUS handles strain measures differently. The one shown here is what ABAQUS describes as an *integrated total strain* - which represents strain obtain by integrating the strain rate of the simulation numerically. It is a bit complicated but will depend on the type of simulation you are doing and the type of solver (ABAQUS Standard or Explicit). Other more common strain measures are the logarithmic strain (LE) and nominal strain (NE). For most cases, and like the case considered in this video, the total strain E11 (along the x-axis) can be assumed to be representative of a true strain but if you do want to test this out yourself, then do so. The values from my last approach (experimental equivalence) is definitely an engineering stress-strain plot as a force measure (though current as we tracked it with time) is divided with an initial area. In the problem considered, the area remained the same throughout the simulation. In which case, the current area was equivalent to the original area (since there was no change) hence the equivalence between this approach stress-strain plot and those of the other 2 approaches. What would be ideal is to track the area values as a function of time and the force values as a function of time (this later already exists). To do so for the area, a coding with Python script will do so. Currently, at the first time-step, I use my script to extract the area. I need to rather do the extracting across all time-steps or frames of the simulation. If this is done, then the force divided by that current area (per time-step) will yield a true stress-strain plot even with my third approach. It is a long explanation but I hope this makes sense to you. Let me know in the comments if there are further questions.
@@MichaelOkereke Thank you very much, sir, for your excellent and clear explanation of my query. The way by which you are explaining the concept is really impressive.
Dr first i want to send my regards for your comprehensive explanations and second i have a question. How can i extract the stress on a surface middle of the model or RVE. To be exact think of a surface that is cutting the RVE or model into two parts now we want the average stress in this surface. Please help
Hello, thanks for the query. This is possible. You simply need to identify a set of elements in the surface you want to. Associate them to a set and then use history variables to track them later.
Thank you for your explanation, I tried to draw the stress-strain curve of a representative volume element RVE in the elastoplastic state with the Comsol software, but I did not succeed, because this software is very different from the Abaqus software.
Hello Saad, I am not sure what is going on with COMSOl but I believe the principle is right. I cannot really comment on how you are trying to apply the same for a COMSOL software, sorry!
Yes I can, just access for it via: cmvig.cmvideos.org (you need to sign in first) and then you will be able to download the files for the model as ABAQUS input files.
Hi Michael, Thank you for these brilliant work. I realised that in your experimental equivalence approach, when you combine the RF1/120 and U1/(30*10), your plot still comes out as Force vs Displacement. You didn't convert that to Stress vs Strain before combining them with the Localised and Volume Averaged Stress - Strain Plots. Is there any particular reason you didn't convert the Force vs Displacement for the ExptEquiv plot?
Hello Josh, this is a brilliant observation. This is a bit of problem with ABAQUS. Although I make the conversion from force to stress and displacement to strain, ABAQUS still plots force in place of stress and vice versa. This is because it does not recognize the conversion I have made but plots the variables based on the original values. This can be sorted if you take the data into say Excel and plot it yourself. Just to make sure in this case that what I am saying is true, look at the original force displacement profile and compare it quantitatively with the stress-strain plot and you will see they are different. I hope the above makes sense to you.
Yes, I realised that Michael, though the plot is stress-strain, the labels are Force and displacement. I managed to change the labels within Abaqus but yes, one can also do the same in Excel. Much appreciated
I believe the approach shown here will still work. You probably will be better off using the Equivalent Stress-strain approach where you model the two dissimilar materials and extract their reaction forces and grip displacements using the reference nodes approach. This way, whatever is happening internally within the dissimilar materials following your study will result in a reaction force which is captured at a reference node. This is what I suggest you consider. Good luck with the study.
Time stamps
--------------------------------------
You can quickly navigate through the videos using these chapters.
00:00 Intro
00:40 ABAQUS Setup of Tesnile specimen case study
01:40 Approach 1: Localized stress-strain plot
06:32 Approach 2: Volume-averaged stress-strain plot
10:32 Approach 3: Experimental equivalence stress-strain plot
13:57 Analysis of experimental equivalence approach simulation results
17:36 Comparison of stress-strain plots from the 3 approaches
19:26 Outro
This is by far the best video I've seen on the stress-strain acquisition. It is very detailed, scientific, and systematic well-made
Glad it was helpful!
great job! best video I've ever seen about stress strain. Thank so much!
Glad it was helpful!
Great Job Dr. Okereke.
Thanks
Great job! Thank you so much! Looking forward to seeing more wonderful works from you!
Thank you so much!
Great Job Dr .Michael
Cheers @Di.
Wonderful Explanation Sir, Thank you so much🙏
You are most welcome
Wonderful tutorial Prof. It is very useful for my research. Thank you so much
Glad you found it useful.
Great work! your videos are beneficial for my research work. Kind request you to post more.
Thank you Nithurshan, I will try my best.
Thank you, Sir
Cheers @Johnson. I hope you enjoy the video.
Thank you very much sir.
You are most welcome @benmelouka.
Your explanations regarding the tensile simulation are excellent. Could you please provide me with information regarding the strain rate or cross-speed utilized in the simulation?
Hello @Hemant, this simulation was a STATIC GENERAL simulation which means the load was applied statically. Therefore effect of strain rate is very minimal and more in the quasi-static range. For a quantitative data, I have applied a total strain of 0.30 over a time of 1 seconds. That means the strain rate is 0.3 per second. This is a quasi-static data but you have to take this with a pinch of salt. If you truly want to quantify a strain rate study, then you have to do a dynamic implicit or explicit simulation where loading effects are adequately considered.
wonderful video Dr, pls can u make videos on thermal analysis of concrete under fire with stress and strain data extracted
Hello, this sounds a lot of requirement under one video. I do not usually take up such projects. Good luck.
Great job Dr. Okereke! If I want to plot Force-displacement curve of a crimping stent test to compare it with the real experiment, would it change something with respect to the third approach you described? Because I am struggling to do it. Thanks in advance!
Great suggestion! I have not tested it out but it is possible to use a reference point to link the surfaces of the stents where the radial displacement is applied. Then, using that extract the reaction force and displacement to get a force-displacement plot. I will try it out and it might work, but I cannot say for now until I try it out.
@@MichaelOkereke thanks! If this works I will put you in my acknowledgments. I am working on bioresorbable shape memory stents.
Hello Dr. Okereke, I am simulating a tensile test of CFRP using a shell element. In my first approach, I fixed one end and applied shell edge load on the other but I ran into a convergence issue. Because of that in my second approach, I applied shell edge load on both ends. I created history o/p and requested RF over the nodes on the two end edges but I got zero reaction forces although the Sim. ran successfully. Can you please advise me to follow the right approach? I also want to plot a Force vs Displ. curve
Best wishes
Use the approach I use in my model using a referral point note and kinematically link it to the edge. Apply your load on that node and extract your history output from that reference point node.
That great work video really, would you how to create anchor, deviator with a saddle-shape on T-beams?
Sorry can't help
Maybe you could compute the true stress - true strain curve from the experimental results. I think this new curve will fit very good with the others. Thank you. Nice video.
I think this is possible but it might not be the case here as the shoulder effect is causing the problem. If it was what you say, then the elastic section will line up together as others. Post yield differences arise from the plastic region rather than elastic section.
@@MichaelOkereke yes, you are right. I didn't take into account correctly the difference in the elastic region. As a matter of fact, I am looking for a numerical experiment to teach the difference between stress-strain and true stress-strain and I thought that your example could be one of them, but it isn't. Thank you for your answer.
Very nice way of explaining the concept sir. Sir, can we extract directly the stress strain plot from the reference point without first extracting force displacement plot from the history output as explained in 3rd experimental equivalence approach? Also, is the first two plots are true stress strain plots and the last one approach( experimental equiavence) is engineering stress strain plot?
Thanks for the comment. A really good question.
You certainly cannot extract stress or strain variable from the reference point. This is because the reference point is a nodal point and stress and strain measures are variables extracted from an element set rather than a nodal set. Also, the first two plots generated from localization and volume averaging methods are true stress and strain values. To be precise the stress is a 'Cauchy stress' (an equivalent true stress) measure. It corresponds to the force per unit current area.
For strain, it is a bit complicated as ABAQUS handles strain measures differently. The one shown here is what ABAQUS describes as an *integrated total strain* - which represents strain obtain by integrating the strain rate of the simulation numerically. It is a bit complicated but will depend on the type of simulation you are doing and the type of solver (ABAQUS Standard or Explicit). Other more common strain measures are the logarithmic strain (LE) and nominal strain (NE). For most cases, and like the case considered in this video, the total strain E11 (along the x-axis) can be assumed to be representative of a true strain but if you do want to test this out yourself, then do so.
The values from my last approach (experimental equivalence) is definitely an engineering stress-strain plot as a force measure (though current as we tracked it with time) is divided with an initial area. In the problem considered, the area remained the same throughout the simulation. In which case, the current area was equivalent to the original area (since there was no change) hence the equivalence between this approach stress-strain plot and those of the other 2 approaches.
What would be ideal is to track the area values as a function of time and the force values as a function of time (this later already exists). To do so for the area, a coding with Python script will do so. Currently, at the first time-step, I use my script to extract the area. I need to rather do the extracting across all time-steps or frames of the simulation. If this is done, then the force divided by that current area (per time-step) will yield a true stress-strain plot even with my third approach.
It is a long explanation but I hope this makes sense to you. Let me know in the comments if there are further questions.
@@MichaelOkereke Thank you very much, sir, for your excellent and clear explanation of my query. The way by which you are explaining the concept is really impressive.
Dr first i want to send my regards for your comprehensive explanations and second i have a question.
How can i extract the stress on a surface middle of the model or RVE. To be exact think of a surface that is cutting the RVE or model into two parts now we want the average stress in this surface. Please help
Hello, thanks for the query.
This is possible. You simply need to identify a set of elements in the surface you want to. Associate them to a set and then use history variables to track them later.
@@MichaelOkereke for making the set i need to make a section in middle ? Or there is other ways.
Thank you for your explanation, I tried to draw the stress-strain curve of a representative volume element RVE in the elastoplastic state with the Comsol software, but I did not succeed, because this software is very different from the Abaqus software.
Hello Saad, I am not sure what is going on with COMSOl but I believe the principle is right. I cannot really comment on how you are trying to apply the same for a COMSOL software, sorry!
Great tutorial! I am wondering if you can share the Abaqus model to replicate yours.
Yes I can, just access for it via: cmvig.cmvideos.org (you need to sign in first) and then you will be able to download the files for the model as ABAQUS input files.
Hi Michael, Thank you for these brilliant work. I realised that in your experimental equivalence approach, when you combine the RF1/120 and U1/(30*10), your plot still comes out as Force vs Displacement. You didn't convert that to Stress vs Strain before combining them with the Localised and Volume Averaged Stress - Strain Plots. Is there any particular reason you didn't convert the Force vs Displacement for the ExptEquiv plot?
Hello Josh, this is a brilliant observation. This is a bit of problem with ABAQUS. Although I make the conversion from force to stress and displacement to strain, ABAQUS still plots force in place of stress and vice versa. This is because it does not recognize the conversion I have made but plots the variables based on the original values. This can be sorted if you take the data into say Excel and plot it yourself. Just to make sure in this case that what I am saying is true, look at the original force displacement profile and compare it quantitatively with the stress-strain plot and you will see they are different. I hope the above makes sense to you.
Yes, I realised that Michael, though the plot is stress-strain, the labels are Force and displacement. I managed to change the labels within Abaqus but yes, one can also do the same in Excel. Much appreciated
SIR HOW TO SIMULATE TENSILE SPECIMEN FOR DISSMILAR MATERIAL WELDING AND HOW TO USE ACTUAL STRESS STRAIN DATA IN ABAQUS
I believe the approach shown here will still work. You probably will be better off using the Equivalent Stress-strain approach where you model the two dissimilar materials and extract their reaction forces and grip displacements using the reference nodes approach. This way, whatever is happening internally within the dissimilar materials following your study will result in a reaction force which is captured at a reference node. This is what I suggest you consider. Good luck with the study.
@@MichaelOkereke thank you
"Why is the pattern of my graph coming out the same, but the values are half in the Abaqus software?"
Hello @Deepak, it is difficult to say without more information. Is your area correct?