Hi all
I want to simulate a free object (i.e. without fixation) in contact with the ground. This object is empty hemicylindrical and receives a vertical force towards the ground. I decided to simulate the ground using the virtual wall function in "contact between contacts between sets". I added to the model plane supports on the edges of the hemicylinder. When I run the simulation, an error message appears: "The iterative solver has stopped".
So I added an extra contact (which could distort my result) by blocking the point of contact with the ground (=virtual wall). The simulation starts however I can see that the part has crossed the virtual wall because the point is indeed in its place (on the surface of the virtual wall) but the part is below the wall...
Do you have any answers? Thank you
Hello
Without knowing anything about your problem, it looks like an oversight to put without penetration if the object passes through the wall.
You should also see a point that worries me is when you write [[ I added to the model plane supports on the edges of the hemicylinder. ]]
So if the simu "bananas" it is certainly because of too much displacement (and also the possibilities of results much higher than the maximum elasticity and the maximum possible displacement). Remember that SW always remains in the elastic domain and stops at this limit.
Can you show you an image (as an attachment please) of the parameters of your simulation.
if you can also post images of the assembly or the ASM itself because it's not easy to know what has been incorrectly configured.
Kind regards
2 Likes
Thank you for your answer, here is the part I am modeling and the parameters entered
For the virtual wall I set a coefficient of friction of 0.9
piece.zip
@chablozortho
Could you post your piece with the elements of your simulation. We should have at least developed the imposed displacements so that we understand your logic.
Do it with the pack and go, specifying that you are attaching the elements of the simulation.
Indeed I have a little trouble understanding what you want to do with supports on both sides and nothing under the hemisphere portion.
If I understand it's the round part that is in contact with the ground but I don't see any contact at this level. I also think that your strength is wrong. This is why we should be told which other piece holds (or leans on) the one we see.
On the other hand, I see imposed movements on both sides, which is weird (although I don't see the direction of the arrows.
I'm waiting for your files to be able to help you a little more :-)
Hello, thank you again for your help.
The purpose of the simulation is to observe the crushing of the hemicylinder on the ground, the ground is guided to the sides (hence the plane supports), the force is applied to the center band, where the support is located.
Attached are all the files, thank you
copie_de_test_sur_sol_3_cylindre_entier_pv.zip
1 Like
Hello Chablo
It seems to me that there are several errors for the simulation and an error of logic.
Error in simulation
1°) You fixed the two flanks while in your message you say that they are guided. If they are truly guided, it means that they can go up and down depending on the deformation.
2°) You consider that the contact of your hemisphere is made on a point, it's true but it's false for the simulation as soon as you have a deformation that exceeds a quarter of the thickness of your part.
3°) SW doesn't like it at all when you put the constraint on a point, not to mention that your virtual wall is of no use and especially the friction coefficient is useless at first
Logic error
If you put a force like in your simu it means that the deformation will be done first on the first non-cut and then the non-cut on the right and left. We have to do things differently. I'll try to do this to you in the logic of a tread.
To show you the problems to be solved , I made you a mini video. In my test I didn't put a dot but a tiny extruded rectangle 0.01 mm thick to have a flat surface. This is what appears as a rectangle in the deformation. Be careful in the video I only put 300N and not 500 N like you. This means that the deformation will be significant.
To go further (only if you want to) you need to confirm the following points
a) Do the side sides slide up and down like in a slide.
(b) Are the upper flanks (parallel to the ground) fixed or guided in any way or can flare or narrow freely depending on the deformation due to force.
Kind regards
2018-06-05_13h51_36.mp4
Great thanks for these elements,
1) the edges are effectively guided by a structure, the goal is to simply reconstitute the contact with the ground
2) so without the virtual wall I can't find the contact of a flat surface with the cylinder... Is that why you made a rectangle of 0.01? It then spreads?
3) What you call non-cutouts are the parts drawn by the cut-out lines, right? I designed it to have the band on which the force is applied, it is possible that this is not the right way indeed...
3) The sides slide well up and down in slide
4) The upper flanks (parallel to the ground) are effectively free
Thank you very much for your help
1 Like
Hello Chablo
Here is the result I obtained, not without difficulty, but hey!
I'm attaching a video of the result of the trips.
For the how to do it, I would have to make you a mini tutorial because it is too complicated to explain in the message. On the other hand, I don't promise to do the tutorial during the day.
Clearly the problem did not come from the virtual wall but from the uncontrollable excessive movements, especially the fact that your room was badly fixed and that the shape was like the souk. I did what it took to make the deformation controllable. (See the tutorial)
Kind regards
PS: you can close the topic if the result suits you.
hemispherique_et_contact_au_sol_v6___sw_2018.mp4
Thank you very much Zozo, this is indeed the simulation I wanted to do.
I'm waiting for your tutorial, thank you again! :)
Hi Chablo
For the tutorial I'm sticking to it :-)
In fact, the whole problem comes from the shape of the part which, if we don't make a few adapted modifications, wants to crash almost infinitely. What is needed is simply define the part of the envelope that could almost become flat.
In doing so, you go from a "great displacement" unmanageable by the simu to a deformation as you see on the last video.
To be totally honest :-) :-) I didn't find out how to do it right away.
Kind regards
Hello
As promised, the tutorial is available HERE http://www.lynkoa.com/node/885656?from_tutos=1
I hope it will be useful to you and maybe you can either learn new things or help you for your future simulations.
Feel free to let me know all your remarks so that I can modify the tutorial if necessary.
Kind regards
OOppsss
in the Tutorial area, I can't put both a Powerpoint file that can contain videos and a PDF file for colleagues who don't have Powerpoint available.
If anyone knows how to do it by maybe adding a chapter to the tutorial???
So in the meantime I'm joining him here
Kind regards
hemicylindre_simulation_deformation_v1.pdf