Analysis of thread in Mechanica

[szofer]

Hello!


Is it possible to analyse in Mechanica thread connection between bolt and nut? I want to obtain stress in threadalong high of nut.


I want to do thisby 3d model but I can't choosehelicalsurface to mate constraint. Following message is appear:


"Not valid surface!
Select mating surface or datum plane on one part"


Why thread surface is divided on several pitch pieces (when I try to choose it)?


View attachment 2720


Is it possible to do that analysis in Integrated Mevhanica?


Maybe I have to take 2d analysis? (is itpossible in Pro/E-Mechanica?)
Edited by: szofer

 

[Mloew]

Although it was not specific to Pro/MECHANICA, I did compile a FAQ on this topic on theEng-Tips forum:

<a href="http://www.eng-tips.com/faqs.cfm?fid=894" target="_blank">3-D Modeling of Fastener Connections</a>

I hope this helps.

 

[rmckinley]

try using "tangent" as your mate constraint type.

 

[Mloew]

This is a contact problem. You will need to define contact pairs between the male and female thread surfaces. The assembly constraints are not relevant. Just use any assembly constraint that does not result in any interference prior to simulation.

 

[szofer]

Thanks for advices!


If somebody has something else I'll be grateful and sure villingly read it.


Tangent constraint don't work properly. It could be make like speakMloew (contact). At the moment I've done 2d analysis. Constraints in assembly is only auxiliary (I've done it on planes).I thing 3d will take a lot of time and I afraid it might do not finish in my computer.


Sorry about very long pictures

View attachment 2722


View attachment 2723
Edited by: szofer

 

[Mloew]

That looks very good. You can see why the conventional wisdom is that the first three threads take most of the load. Some of the links I put into the FAQ talk about comparing axis-symmetric to full 3-d simulations.

Make sure your mesh is fine enough (even with P-elements).

 

[nagabhaskar_m]

very nice , by seeing this i got the idea of doing analysis of thread , thanku very much

 

[szofer]

Is there any way to make more thick mesh? (the simplest in Integrated Mechanica) I suppose it would make more accurate results. Am I right?


And another thing:


What is it possible to make graph with plotting grid equalsfor instanse 10 points when modelalong this line contain only 1 or 2 elements.Where fromsystem obtain these data?Is it reallyneedful to make moredetailed mesh?


View attachment 2731


Below my 3d model. Unfortunetly I can'tdone it withthreadas helix (problem with add contact region).


View attachment 2732


Best regards.
Edited by: szofer

 

[szofer]

Is there anybody able to explain problem about graphabove?


It's very mysterious for me. This graph isn't linear.Along this curve system makes 10 nodes?


How it possible that Mechanica show nolinear graph inside one element? In FEA values of parameters (displacement etc.) are determine only in nodes. But there are nodes only in corners.


Is this graph reliable?
Edited by: szofer

 

[ndk]

There is no problem with the gragh. FEA in Pro/Mech using P element method. Each element (that you have seen as in the mesh) have higher order (in Pro/E, from 3-9) and so forth has more interpolated points in 1 element. To know this gragh is reliable or not, youshould check the convergence of the Von Misses stress.


NDK