Continue to Site

Eng-Tips is the largest engineering community on the Internet

Intelligent Work Forums for Engineering Professionals

Register Log in
  • Congratulations KootK on being selected by the Eng-Tips community for having the most helpful posts in the forums last week. Way to Go!

Problems with Free-Free Rigid Body Check in COSMOS ?

Status
Not open for further replies.
johneng

johneng

Aerospace
Mar 19, 2011
6
Hi everyone,
I have a question relating to checking COSMOS models and in particular running a model free-free modal analysis.
A model has been sent to me at work with the results of a free-free modal analysis and there are 3 rigid body modes (translational but the following 3 modes are about 6 Hz, 14 Hz and 40 Hz ! but it is claimed that this is because there is a 'known problem' with rigid body modes when plate(and solid) elements are used.
Is this true and if so where can I find documentation on this limitation ? Has anyone else encountered this problem ?
Personally I find this hard to believe.

Thanks
john
 
Replies continue below

Recommended for you

Sort by date Sort by votes
Hi,
I maybe didn't put my question very clearly. The analysis is a free-free modal analysis, so it is expected to have x6 zero frequency modes. The analysis actually produces the first 6 modes, 3 with frequencies tending towards 0.0 but 3 others at 6 Hz, 14 Hz and 40 Hz. It is claimed these last 3 modes are not even close to zero because of a 'known problem' in COSMOS related to plate elements and rigid body modes.It is claimed this has been discussed with Solidworks/Simulation support people but I can't find anything related.
For me this type of 'FEM quality' check is standard in NASTRAN and there should be x6 frequencies tending towards 0.0 each one.
It's a check against over-constraint in the model (<6 zero fequency modes - as in this example) and/or mechanisms (> 6 zero frequency modes appear).
Any ideas ?
 
Upvote 0 Downvote
I suspect you are using solid elements so you would expect only 3 rigid modes because solid Elements only have three translational dof. If you have shell elements coupled to solid elements you have removed the 3 rotational dof in doing so.

That's my 2 cents worth.

TOP
CSWP, BSSE

"Node news is good news."
 
Upvote 0 Downvote
No, this can't be the explanation. Whilst individual points have no rotational dof this doesn't mean that the whole structure doesn't have rigid body modes !
There are both solids and plates in these models i.e. its a mixed shell/solid model.
Here is the exact 'explanation' given when the model was delivered :
"The Eigenfrequency results revealed a limitation in Solidworks 2010 where rotational rigid body modes are miscalculated for shell elements. This limitation is not believed to affect the results of the modal analysis.
This is not compliant in terms of the FEM requirements. However, it has been reported by Cosmos developers and technical support that this is a limitation of the solver itself when dealing with shell type elements and not of the Finite Element model"

Anyone any further ideas ?
 
Upvote 0 Downvote
Yep thats it. Stress-free =free-free
Non-zero RB modes means the damn thing is tied down somewhere when it shouldn't.

Whoa boy !
You're talking to a NASTRAN man here, used to problems (which all these damned problems have eh ... they all ought to be exorcised :) of a different kind.
I didn't even know there existed such things as 'draft elements'.
Do you boys actually deal with 'variable quality' elements ... nice concept but don't let the inspectors know.
I thank you for your comments anyway as its always nice to get people interested in helping people solve their problems.
No one else has replied so I suspedct that the reply supplied by the supplier is based on direct transfer of a comment from a Tech Support guy (stand up you guys you know who you are:) at Solidworks (probably 'mierda.de.toro'-ing him).
Just out of interest do you know if Solidworks Simulation (aka COSMOS) has a 'bug list' (otherwise politely known as the error list) 'a la' NASTRAN ? All these damn programs should be obliged to by federal law :). Maybe what is needed is someone to get a lawyer on the case. ' M'lord, thus because of the possibility of my client being completely fcucked by his customer because of the inaptitude of the <insert name of favorite FE program here> 'bug' number xx , I respectfully request the sum of $nn million in damages to my client. Just a thought.
Of course none of this would happen if engineers all considered their work as if peoples lives were at risk (stand up you Boeing/Airbus people).
Gets down from his HPC soapbox and slinks gracefully away to contsact a SW tech support guy (I don't know who you are (yet) but boy you had better have an answer :) )
 
Upvote 0 Downvote
A draft element is just Cosmos way of saying 3 node triangle or 4 node tet as opposed to a 6 node triangle or a 10 node tet. They call them "draft" quality to let the users know they might be good for testing boundary conditions and such, but use the higher order elements to get real answers. Same drill in NASTRAN (I'm a NASTRAN/ANSYS/ABAQUS type of guy). You may find documentation in the Cosmos/M manuals assuming they are still formulating the elements the same way.

You still haven't answered my question as to whether you have the same problem with all the solvers and element types or just in one particular set of circumstances.

I am well aware of the shortcomings of CosmsosWorks. Do you have Cosmos/M installed along with Simulation? Sometimes it is. If it is, you can export your model to that and have all the bells and whistles at a price in user friendliness. As always, it is the engineer's job to verify the answers are correct.

TOP
CSWP, BSSE

"Node news is good news."
 
Upvote 0 Downvote
I can't answer your question.
This model (and related analysis) comes from and is analysed by a supplier, I am just reviewing the analysis work. We have a detailed checking procedure which they must satisfy and which they must satisfy.
We do not have solidworks/COSMOS at all ! Hence we cannot run anything, I'm just trying to find out if what they say (see my last post) is actually true because basically I don't believe it.
 
Upvote 0 Downvote
Gets down from his HPC soapbox and slinks gracefully away to contact a SW tech support guy (I don't know who you are (yet) but boy you had better have an answer
Click to expand...
They likely won't know what you are talking about. And if you aren't on subscription they likely won't want to talk to you.

#1 said:
I have a question relating to checking COSMOS models and in particular running a model free-free modal analysis.
Click to expand...
#2 said:
We do not have solidworks/COSMOS at all
Click to expand...

I'm trying to understand these two quotes. How can you have the model to check and not have the software? There is no bulk file or any other way to view the model. Are you aware that Cosmos can refer to Simulation and it can also refer to Cosmos/M, a standalone FEA package that predates Simulation/CosmosWorks? Many people that have Simulation have Cosmos/M and just don't know it. Are you saying you don't have SolidWorks?

SolidWorks Simulation is specifically designed to hide from the user almost all the internal workings of the software. At one time in the far distant past there was a small attempt at documenting bugs a'la ANSYS, but those days are far past and it was for Cosmos/M.

TOP
CSWP, BSSE

"Node news is good news."
 
Upvote 0 Downvote
Hi kellnerp,
just a line to say thanks for your effort to shed some light on this matter, even if we havent found the reply !
Just to reply to the points in your last email , NO I dont have solidworks. All I have is a mech analysis report with a statement concerning the model checking.

As for tech support, whats new in any of these progs ..... the only exception was about 20 years ago when NISA was starting to market in england ... very good tech support from the guys in Southampton .....
What you said about hiding away all the complicated stuff inside the program reminds me of a quote from someone a very long time ago who said ..... "Any fool can fill out a -n_ASTRAN deck, run it and believe the answers ... "

Thanks anyway for all your help ....

johneng
 
Upvote 0 Downvote
Johneng

If all you have is a "paper" report and it doesn't answer your questions you can ask for clarification from the provider of the printed report, such as:

"Please clarify/show boundary conditions."
"Please show all rigid body modes."
"Please show reactions."
"Please show convergence study as a graph for for first three modes vs characteristic avg. element size."
"Please show reported model error."
"Please show analysis options used."

TOP
CSWP, BSSE

"Node news is good news."
 
Upvote 0 Downvote
Status
Not open for further replies.

Similar threads

oharag11
  • Locked
  • Question
Rigid Connector proper application
Replies
10
Views
461
oharag11
oharag11
IdanPV
  • Locked
  • Question
Apply Moments in Solidworks Simulation
Replies
2
Views
260
IdanPV
IdanPV
Anurag K
  • Locked
  • Question
Modeling clamping effect in SolidWorks FEA
Replies
2
Views
256
FEA way
FEA way
IdanPV
  • Locked
  • Question
Create Custom Results Plot in Solidworks
Replies
2
Views
204
IdanPV
IdanPV
samim09me
  • Locked
  • Question
How Solidworks Motion Analysis function builder are used with examples
Replies
0
Views
192
samim09me
samim09me

Part and Inventory Search

Sponsor

Back
Top Bottom