Assembly constraints within Mechanica...

[Vesh]

How does Pro/E treat assembly constraints in FEA (Mechanica)?

If for example you "mate" face 1 (from part 1) to face 2 (from part 2), do all the elements in part 1 flush with face 1, 'bond' or 'glue' to all the elements from part 2 flush with face 2?

How does it treat placement constraints, like standard alignments, axis alignments etc.

Any help would be appreciated.

Cheers...


Edited by: Vesh

 

[deepakbhat_nie]

Hi,


The asm constrains has to be given again in mechanica using the options like connection(fastners, weld etc) .


the normal asm constrains will not have any importance in mechanica.


Regards,


Deepak Bhat
Edited by: deepakbhat_nie

 

[Vesh]

I done an analysis earlier and simply constrained/loaded the asm file but didnt add any connection types as mentioned and it seemed to work, ie no errors in the model. Why is that?

How doea the analysis work without connections if you are saying they are required for FEA?

 

[deepakbhat_nie]

Hi,


It is only the connectivity b/w nodes which has allowed you to complete the analysis.now try the following example.


take two flanges which are bolted together and apply atorsion load on one. check out whether you will get the desired result.


in order to get a proper result you have to assemble the two faces by mating it at a particular distance and apply bolting connection.and then apply the torsion load.


I hope I am in line with your question. let me know if I am giving an absurd answer.


Regards,


Deepak Bhat



Edited by: deepakbhat_nie

 

[Vesh]

Yes absolutely your answers are in line, thank you.

I dont entirely understand "...it is only the connectivity b/w nodes..." sorry.

So you are saying I need to mate faces with an offset of some value, and then add connection types to constrain them to each other, like weld, bolt, etc?

Is this offset need to be an exact value? 1mm, 5mm?

 

[deepakbhat_nie]

Hi Vesh,
there is no standard offset value for the offset.maintain somewhere around .05mm the logic is the elements in one component should not be connected to the elements in the other.by connectivity b/w nodes I mean that when a node of an element of one component is same(or connected)as the node of another element of another component then the load gets transferred directly through nodes.which will not be the requirement in many situation.


Regards,


DeepakBhat

 

[Vesh]

OK thank you.

So when in assembly mode I should never "coincident mate" any components as this would connect nodes.

What about when pin jointing a pin/bolt through a hole, does this connect elements?

Overall are you saying the correcct way to do FEA on an asm file is to ensure no surfaces from different parts touch each other? Then in mechanica mode you create all the 'real world' connections like bolts/welds etc.

Cheers...

 

[deepakbhat_nie]

Hi,


For bolts you have the option of connection.it takes care of all the things. you dont have to worry abt it.just put the connection and leave the rest to mechanica.and the rest you have understood it right.


Regrds,


Deepak Bhat

 

[chrisu]

Hi,

I was asking very similar question recently and I am slowly moving forward, but somekind of tutorial on connections would be very helpful. Do you happen to have anything like it? Or maybe you know the source other than official sites where it costs around 500$ . Plz let me know because I have a quite complex model and I am getting frustrated

 

[Vesh]

Try this,

http://www.staffs.ac.uk/~entdgc/WildfireDocs/tutorials.htm

It has some very helpful tutorials. Not sure if has much on connections specifically though.

 

[chrisu]

Thanks for this link, but I have already practised with these tutorials and I need something more precise on connections and meshing. In fact I have even written to the professor Cheshire, founder of this website for more tutorials but he hasn't got any more )

Good luck and if you find something for ProMech don't hesitate and share it )

Cheers

 

[ghgarzon]

I belive this should help. Come right from the pro-E website technical support link.


2006-08-28_091137_Transferring_assemblies.zip


Unless the modeler has changed, (I'm not aware that it has)assemblies with no gap between the mating parts become 'welded' in Pro-Mechanica


ghg

 

[chrisu]

The only problem is that my assembly is composed mainly of pipes and all fasteners are suitable only for parallel surfaces. I try to deal with it cutting out very small parallel surfaces where I need to point references but it is not very professional I suppose(because there may be different stress distribution in these areas). But for this moment I don't have other ideas. I any of You think of something or have any clue please let me know.

Regards,
Chris

 

[ghgarzon]

Chris,


What is the purpose or goal of your analysis?


Is it stress? deflection? or is it modal?


If it's stress concentration at the joints, you may have to do the analysis in two steps:


1) A complete assembly in order to find the areas of the model with higher stresses


2) Then model a 'cut out' of the region of the previous model with the objective of predicting more accurately the max stress.


For step # 1 you may use the flat surfaces as you're currently doing. (Have you tried rigid links? I can give you additional details to use them correctly if you like).


For the second step you may have to model the weldments or bolts in a more detailed way.


On the other hand, if the goal of your model is modal analysis; what you're currently doing might be just fine, because modes and natural frequencues are global properties, less dependent of the way you connect the components. Not that you don't have to model the connection correctly but that small deviations from the ideal model have a lesser effect on the results beacuse the objective is to determine a global and not a localized property or result, such as stress or deflection.


As for the person whowrote that Pro-mechanica only joins the elements at the nodes for two touching surfaces, that is not correct. The two adjacent volumes become as welded as you can get it at the common surface;one advantage is that you can give different material properties to the two touching volumes, if that happens to be the case. Of course, since the purpose is to model as close to the real object, then modeling as bolts or weldscan bemore appropiate.


I hope all this rambling helps.


Sincerely,


ghgarzon

 

[chrisu]

Thanks for showing interest in thi subject ghgarzone,

If you would be kind enough to send the materials you mentioned on the rigid connection I would be very grateful. Right now I am trying to solve my problems in the mechanica independent mode, because it has more options and you can investigate more by yourselve than in the integrated mode where everything is automatic.
One can see the elements that are causing problems for mesher and so on. But I think I'll start doing the same thing in ANSYS to compare few things.
Cheers,

Chris

 

[ghgarzon]

Hi Chris,


You didn't answer my question as to what was the main objective of the model. Stress analysis or Dynamic/modal analysis? That's OK, you don't have to.


Here is a link that I have found has very good tips for modeling joints. It hasn't been kept up to date; however, most of the suggestions are still valid.


http://www.tsdengineering.com/


I applaud you for goind to Mechanica Independent. I used it for 7 years before switching to Pro-M/Integrated. I still use it whenever I run into diffcult modeling situations in the Integrated environment.


Sure, try Ansys if you have access to it; it will reassure you that you're getting the right answers.


Regards,


Guillermo

 

[chrisu]

Hello,

Forgive me my ignorance. The subject of my analysis should be stress and deflection, because it's point is to optimize the construction of a wheelchair. So I think both characteristics are useful.

Thank you for the link, for sure I will use that information.

Regards,

Crs