"Large displacement","Equilibrium satisfaction not achieved" 3

[xhbg]

Hi, I am a newbie in COSMOS. I am working within linear range, but sometimes I have the message saying turn on the large displacement flag at the end of the run. What should I do to solve this problem if I dont want to move to non-linear analysis?

Another message says "Equilibrium satisfaction not achieved. Save the results up to current iteration". Why is there such message and how can I solve this? And if I click "Yes", are the results I get all wrong?

Please advise me. Thank you so much in advance.

xhbg

 

[corus]

The large displacement request indicates that Cosmos thinks you need to use this feature. Using it will add a little more to the run time, but unless you know you have severe large dispalcement where the direction of load will have a major influence on the results, then you can ignore it. It is preferable to use it when Cosmos suggests it though.

If you don't get a solution but you save the results up to the current iteration, then the answers will most likely be wrong. Looking at the results up to that iteration might help you to see what is wrong with the model.

corus

 

[2design]

Try a modal (frequency) analysis on the model first to make sure everything is attached. Unfortunately you will have to redo your mesh to do this. Sounds like there are constraint issues in your model.

 

[xhbg]

Thank you Corus and 2design for your help.

But if I expect my model to behave within linear range, I should not activate the large displacement flag, should I? As COSMOS would automatically run it in non-linear, is that right?

I have also encountered a problem with "No penetration" in boundary condition. Even though I set the two surfaces of two parts as "No penetration", "surface to surface" option, the result still shows the interference between the two parts. I tried with mesh control, small element size, but it doesn't help it seems....

I do appreciate your help.

xhbg

 

[2design]

Are you modeling it with solids, shells, or both? Penetration works pretty well with solds but not so good with shells and even worse with both.

 

[xhbg]

I am modeling it with solids only. But I still face this problem. Is it because i have gaps between parts that might cause the problem?

 

[2design]

Make sure that the contact areas between the two are identical and normal to each other. I have had problems with this in the past.

 

[2design]

I meant parallel

 

[xhbg]

I have found solutions for most of the problems I face, but there is one which I am still stuck with. I have this warning at the end of the analysis. "Load/Restraint produce large displacement. Would you like to stop the program and rerun with large displacement flag activated? Yes/No".

How can I avoid this warning without activating the large displacement flag as it does take very long time to run? There is one plastic part in my assembly, but I could run with high load before without this warning....What could cause the problem?

Your help is appreciated and thank you so much in advance.

xhbg

 

[cbrn]

Hi,
you said in a previous post that you have some initial gaps between, I suppose, at least two components of your model: in this case, the first iteration will see a Rigid Body Motion for some nodes, and this is not compatible with the small-displacements hypothesis. By the way, generally speaking, analyses which start with a RBM will not solve at all: in your case, probably you're lucky because other surrounding restraints (contacts...) avoid this RBM to cause displacements higher than the internal stop-limit. But the program is aware of it and prompts you for the message.
The only way you have to avoid it at all is to "cheat" with the initial position(s?) of the component(s?) in order not to have any "initially-open" contact (i.e. gap). Note that, by "closed contact", I mean "at least one node is in contact": this is enough (even if, in case of 1-node-only initial contact, the first iteration will experience convergence difficulties).
Hope this helps...

Regards

 

[xhbg]

Thanks cbrn,

You said "cheat" with the initial positions of the components in order not to have any "initially-open" contact, did you mean that I should change the geometry of the components to make sure there is no gap in between or I can just simply change in "Contacts/Gaps"? And if so, how should I do it?

Thanks a lot.

xhbg

 

[cbrn]

Hi,
well, I haven't been using Cosmos since 2 years, so I don't know of the latests improvements in the "Contact" settings (which were extremely simplicistic up to version 2004). Of course, the most "realistic" way of proceeding is putting the components with the CAD in the initial touching position they would assume in reality (because of tolerances and gravity...), or near this position (sometimes you have to "guess" !!!) but always touching at least in a point / edge. If it's not possible because, for example, the model isn't yours and you have no way to receive a model prepared as you need, then probably you can adjust some settings in the Contacts (something like "initial gap = 0", or "close gaps"...).

Regards