Fastners connection in FEA! 1

[johnsmith2]

Hi,

I am the user of both abaqus and Pro/Mechanica. As fastner connections was introduced in the version 2.0 in pro/M. I found compared with abaqus, it is rather simple to make connection between two parts by screws or bolts, what you need to define is edges of the holes, materials, diameter of the bolts, or/and spring stiffness properties, etc. Most important, I don't need to model the real physical bolt parts in Pro/mechanica, I don't need to define the contact pairs between bolts and the parts.

However, I am not sure about the results in Pro/M, especially the contact pressure between bolts and parts. Of course, in Pro/M, bolts take all shear, so, no friction between parts exist.

In abaqus, you need to define surface-to-surface or node-to-surface contact, you need to model the bolts in beam elements, define the master and slave surface.

If my purpose is to ensure that the bolts are strong enough to hold or support under pressure/force between two parts, do you think Pro/M can fulfil my requirements. Thanks for ideas and opinions.

 

[GregLocock]

Do some simple test cases and find out, by comparing with a hand-calc or test. Use coefficient of friction of 0.15 to 0.3.

Cheers

Greg Locock

Please see FAQ731-376 for tips on how to make the best use of Eng-Tips.

 

[killerbee]

Friction is one of the more inconsistient factors in the response of a joint. It depends on clampup forces that have a tendancy to decrease during the service life of the joint. There is a good reason that you get no credit for it in aircraft applications with respect to load transfer in a shear joint. So going the extra mile to simulate the friction load transfer is not worth the trouble. If you can not guarentee it in service you can not take credit for it. So for shear transfer I'd say the friction is not worth considering.

In fact I can think of no practical example where joints would be sized including friction. Maybe as a research excersise but not when it counts.

 

[cbrn]

Hi,
Killerbee, I may have misunderstood something in your post, but as I read it I have to totally disagree.
In turbomachinery, just to make an example, friction couplings between rotating parts are employed everywhere and everytime, except some very rare cases. All the shear load, coming from torque and other direct loads, is carried by the flanged parts through the friction. That's why special powders (silicium carbide especially) are put on the interface in order to have a stable and reasonably known friction coeficient (mu=0.5 +/- 5%, generally). Recently, pre-built films (silicium carbide in a very thin resin matrix) have been used instead, making erection on site even easier and controllable. Bolts are preloaded with a very high preload/live-load ratio. In this case, a careful and complete calculation of the joint (bolt+clamped parts, embedding effects, verification of the threads, of the bolt's resistant area, etc etc...) is mandatory. In this case, as the O.P. was asking, it would NOT be allowed to study the joint itself with a system like Pro/Mech.
If, instead, the need is to study the parts which are bolted together, then a "connection" like the one made possible by Pro/Mech (or CosmosWorks, just to mention another - in fact, every FE program has this opportunity, be it more or less difficult to set up) is, in my opinion, perfectly adequate. The designer will then have to separately calculate the joint itself.
Clamping forces vary in the life of a joint, that's a matter of fact, but:
- there are design techniques in order to make it less sensitive to preload loss in case of alternating loads
- there are calculation methods which allow to carefully account for all these factors (e.g. VDI2230)
- there are applications for which preload-bolted friction joints are not the best practice, or may even proove dangerous, but these are not the majority of cases.

I have examples of friction joints in hydraulic machines which have worked for something like 45 or 50 years, so you can imagine the degree of reliability we can achieve NOWADAYS on joints like that, with the improvements on material science and calculation techniques...

Regards

 

[killerbee]

Disagree away It was the airframe industry I was referring to. What I said is perfectly correct in that application.

 

[rb1957]

i think cbrn and killerbee are talking about two different things. certainly there are devices designed to use friction to transmitt load; mind you, i think the clamping force is controlled (hydraulically?) so the joint maintains it's capability. else there's a bunch of maintenance to control it; the point being that the device is designed to work with friction.

however, i think killerbee is talking about the preload in a bolt which is due to the friction on the threads, whihc isn't very well controlled ... and so it's of questionable value.

i'd just model the connection, which would give me the external load in the joint, and then design a bolt preload (including all the scatter factors and possibly thermal effects) to suit. And then you know the maximum clamping force, so you can design the pieces being clamped.

 

[killerbee]

Could not agree more.

 

[051174]

Pressure or force applied beetween part connected by bolts on the flanges connection line is not relevant for bolts. If beam give compression is because you cannot simulate contact of flange but bolts will be surely unloaded in reality.Bolts have not to be sized for this case.In general for this cases nobody could ask to you what is the bolts loads.

 

[Qshake]

Just to add fodder to discussion, civil-structural applications use friction connections all the time, they are called slip-critical connections. However, I will add that the ultimate scenario is always checked: that of friction failure and bolts engaged in shear.

I say all the time, but mainly with heavy infrastructure. Most building applications are mostly bolt shear connections.

Regards,
Qshake

Eng-Tips Forums:Real Solutions for Real Problems Really Quick.

 

[brep]

Just for your reference, here's a good description of modeling preload in bolted connections. Actually it's an ABAQUS patent:

http://www.patentstorm.us/patents/5920491-fulltext.html

 

[cbrn]

Hi,
Killerbee, Rb1957, I agree. Thank you to Rb1957 for having "bridged" what we were talking about.

Regards

 

[johnsmith2]

Thanks for all information and explanation above. Brep (in Pro/Mechanica forum) is very true to point out what I should try to do in the problem.

1) Load transfer between components?
2) stress distribution around the fasteners?
2) stress in fasteners itself?

After reading the manual, in Pro/M, I think what they can do is (1,2), not (3). Because we don't need to model bolts as solid or shell elements. So. we have no idea to see whether they are strong enough to hold external forces. Also, bolts take all the shear, without friction between parts involved. Moreover, linear analysis cannot catch the behavior where the preload is overcome and the joints start to open up.

It seems that most people rely on the hand calculation to predict the bolts stress. For example, based on the tensile area of bolts, the bolt stress can be calculated based on the force or torque applied to bolts.

thanks again to make me think!