Hello,
I established an FE model for a sleeve type seat air spring. Step 1 is to inflate the air spring and step 2 is to move the piston upwards to the design height by 70 mm, as shown in the attachment. General contact is defined for the reinforced bellows and the piston. Friction coefficient is 0.3. All the elements are shell elements and the simulation is completed in Abaqus/explict.
Can anyone give me any suggestion about why there is "abnormal" deformation in the bellows near the piston?
Thank you.
Looking at it, it might be just too course mesh to capture that large curvature. Other reason could be excessive mass scaling (do you use any?). You could try to refine the mesh locally in the curvature direction.
By the way, this model looks like something that you could model as an asisymmetric model, it could save you a lot of trouble, and you could have a lot more elements for lot nicer results. Just a tip. Obviously if you need to simulate off-centre conditions its a no go.
Hello David,
Thanks for your suggestions. For now, I did not apply mass scaling. The element size near that "abnormal area" is 2 mm and I refined by 0.6 m. It didn't help a lot.
At one second, the deformation of the bellows is smooth. After that, it deforms likes the attached figure. I am looking into whether it works if I control the stale increment time.
Regards
Ok, well if it happens suddenly, it can be something like a snap-through exciting an unrealistic vibration mode, or there is some error. Does the analysis complete, or is this the last step where you get results and then it aborts with an error? Time increment should have no effect on this, ideally your model has to be time-step independent, so if its so sensitive to time that you get good/bad solution based on the step value, that's no good really.
Hello David,
Thank you.
It could be. The analyses are complete without an error and the force the model predicted is acceptable.
For now, the time period for step 1 and step 2 are 0.02 second an and 0.1 second (the displacement is 70 mm). What I want to do is to increase the time period to, for example, 2 seconds or even 10 seconds to make the simulation to be a quasi static process. I am not an experienced user in abaqus/explict. Hope you can give me more suggestions.
Regards,
- Badges
- Science topic
- Similar topics
- Developmental Psychology
- Attachment