I have a specimen which is arc welded in a single pass. For simulation, I use a moving heat source and get the temperature vs time distribution. I have used quiet elements technique for the filler material.
During structural stress simulation, I use the same quiet element technique again, assuming small displacement theory.
However, during the structural stress simulation, the elements are born with some initial displacements, strains and stresses. Besides this happens for all those elements whose temperature has exceeded melting point.
Hence, I have a doubt regarding the following:
1) In my code, should I forcefully make the artificial displacements, strains and stresses as zero, for all elements which are not born?
2) For other bulk elements near the weld line, what should I do for displacements, strains and stresses, if the elements melts and becomes liquid?
(Note: I have assumed small displacement theory only, to keep my code simpler)
"Awaiting for kind insights from experts. Thanks"
Activatiing an element mechanically before it is melted (and is about to become solid again) should be avoided. This requires that you can separate the thermal and mechanical activation though.
Andreas Lundbäck Sir, Thank you for your expert insight into my query. I would model the mechanical element birth of filler, when their temperature starts dropping down from the melting point.
I have a small further doubt. There are a few elements, near the weld line, which belong to the main plates. I have not made these elements quiet, because they are not part of the filler.
During simulation, these elements of the main plates, also melt. So should I make displacement, stresses and strains as zero for these elements too, when they get melted?
Sandipan Baruah for your further doubt, once you performed full thermal analysis, select those elements which crossed the melting point of the chosen material and kill them in the restart analysis.
Shekhar Srivastava Sir,
Yes sir, I have exactly done that by considering the suggestions provided by Professor Lundbäck.
However, it seems that due to this improvisation, the stress distribution close to the weld center lines becomes highly compressive, which is not correct. Any suggestions would be helpful for correcting this. Thanks
Sandipan Baruah, the stress state of your activated element depends on how your software handles it. Is the material configuration assumed to be given in current configuration or in the original configuration? The result will differ quite a bit depending on how the solver treats it (and what you provide).
Professor Andreas Lundbäck , thank you Sir, for identifying this possible concern.
The material configuration is in the original state, since I have considered small strain analysis and no finite strains. The analysis is being done in MATLAB and no commercial FEA software is used.
After the non-linear incremental thermal analysis, I obtained the temperature-time history and so, I listed out all the elements of the main plates which had crossed melting point.
Before the start of the structural analysis, I deactivated all those elements of the main plates. The elements in the gap between the plates were already deactivated from beforehand.
So, now with all these deactivated elements (filler + melted bulk), I started the structural analysis.
In structural deactivation, I had made the elemental stiffness matrix small (without ill-conditioning) and zeroed out the displacements, stresses and strains which had possibly accumulated in the deactivated elements.
As the welding progressed, I started to sequentially activate the elements whose temperature started to cool down from the melting point. Stresses and strains started building up in those activated elements from the high temperature and progressed towards room temperature.
However, at the end, after full cooling, some highly compressive strains developed near the weld centre line, which seems to be incorrect.
Any insights or suggested corrections to the above described analysis would be helpful.
Thanks and regards,
Sandipan
You need to reset the deformations/strains to zero when an element is activated in “liquid” state. This is a little bit trickier in a small deformation formulation, but it is doable. Especially, as in your case, when you have the source code :)
- Badges
- Science topic
- Similar topics
- Biological Science
- Physiology
- Physiological Processes
- Stress