Bolt preload in NX 8.5

[christophersrikanth]

Has anyone used NX 8.5 for bolt preload? NX 8.5 models the bolt as solid element-anyone has experience using this functionality? Any video/tutorial for this will be very useful

Christy

 

[Michel1978]

I also have this question. Did you find anything?

I use NX8.5
HP Z600 Intel Xeon E5520 2.27GHz Dual
Quadro FX 3800 5GB
9GB Memory
Windows 7, 64-bit

Groeten, Michel

http://www.kvi.nl

A leading Dutch institute in atomic and subatomic physics

 

[rb1957]

in the olde days preload was simulated with thermal loads ... cool the bolt down (i think) so it contracts against the surrounding material

Quando Omni Flunkus Moritati

 

[izax1]

Yes, that is correct for the old days. I think ABAQUS has a dedicatrd bolt preload element. In NASTRAN I think you need to cool the bolt down (1D Bar element with rigid connections in the ends.)to impose preload like rb says.

 

[Michel1978]

I found it. Just make a bolt connection in your fem model.
Then assign a cross section and material to the CBAR or CBEAM 1D element which represents the bolt shaft.
And in your sim file you can easily add a 'bolt pre-load' with 'load type'. Select the CBAR or CBEAM elements and add a pre-load.


I use NX8.5
HP Z600 Intel Xeon E5520 2.27GHz Dual
Quadro FX 3800 5GB
9GB Memory
Windows 7, 64-bit

Groeten, Michel

http://www.kvi.nl

A leading Dutch institute in atomic and subatomic physics

 

[ninjaz]

What happen when we do not consider bolt preload in the simulation.

 

[izax1]

Well, that depends on your objective. If you do not consider bolt preload, you might as well ignore modelling the split face. That could be correct. But if you have a connection that is subject to repetitive laods (fatigue case),thermal loads or vibration/shock loading, disregarding bolt preload could be fatal.

 

[loki3000]

when you apply the preload, how do you account for where in the bolt the preload ends? (the Kf factor)

http://lace.uni-mb.si/strojni_elementi_1_UNI/slikovno_gradivo/6 - Vijacne zveze.pdf

pages 24 and 25 (of the 50page pdf) here.
the first image on the 24 is the ideal, the second two are the actual scenario. on the pg. 25 there are more pics of how the preload is determined by the flange design... (look for the Kf)
also, how do you call that in english - where the preload starts, from the Kf factor (just look for it in the eqs)

 

[loki3000]

(or in ansys, abaqus), if it's not possible in nx...

 

[izax1]

A bolted connection is a very complicated and highly non-linear system, so I tend to not do this type of analysis with FEM. The type of FE appraoch to use is also dependent if you want to do a bolt connection analysis (verify the connection) or a bolt analysis (verify the strength and stiffness of the bolt) which sometimes requires very detailed modelling of threads and fillets.
Anyway, the load distribution in a bolt connection is dependent on the stiffness ratio of the bolt vs. the clamped area. If you model this correctly, the FE model should take care of the stiffness distribution and ditribute this correctly and thus the load introduction plane (and Kf) is taken care of. But remember: The FE method is a numerical approximation approach, and the only thing you can be 100% sure of is that the answer is wrong!! How wrong depends on the skills of the FE user (and to a very minor extent the SW system used)
Sorry for sounding very pessimistic, but in general I sometimes get scared of engineers relying 100% on the FE answers. But if the analysis is done correctly, your answers will be reliable, and probably good enough.