Shear transfer via fasteners in torsion in built-up, thin-walled, open sections

[SteveAero]

Hi all,

I am trying to find a method of fastener shear transfer between a thin-walled open U-section and skin (both are aluminium, with skin attached on both sides of the vertical walls). From the grid point force balance output from the FE model, I know what torsional moment is being carried at the cross-section. However, as my section is modelled in FE as shells, I have simply increased the thickness of the vertical walls to integrally include the skin. Since shear flows in a loop due to torsion in a thin-walled open sections (assuming no warping), what shear force is transferred between the U section and the outer skins via the fasteners? I have modelled the skin attachments separately with fasteners in FE and get pretty strange results. I have not been able to find any theory of this means of transfer online. Any practical help would be greatly appreciated.

 

[rb1957]

"I have simply increased the thickness of the vertical walls to integrally include the skin" ... this may not be correct if the torsion is being transferred from one element to the other.

However if you extract the torque from the model, and this is being applied to the U channel then you know the internal loads (shear flows) this creates. Now if this torque is being transferred to the skins , well ok the rivets transfer the shear flows.

But as I think of it you could have the torque applied to the composite section, but this doesn't really load the rivets. So how much torque can the U channel accept ? How much torque can the skins carry ? But for rivet loading, you need to transfer the torque into or out of the skins (or the U channel).

Something to consider is force balance. You may want (need?) to reaction the torque as a couple on the vertical sides of the U-channel (rather than all three sides).

A picture might be helpful !

"Hoffen wir mal, dass alles gut geht !"
General Paulus, Nov 1942, outside Stalingrad after the launch of Operation Uranus.

 

[SteveAero]

Hi rb1957,

I have attached a sketch of the assembly as per your request. I have thought of applying the torsional moment as differential couple force and splitting the shear force according to relative area-as you suggest. But there must be a more analytical method...using the shear flow...somehow...

 

[WKTaylor]

Rb... I concur...

In this case, a picture is the only way to clarify the question/issues...

U-channels and fastening patterns come in a lot of flavors...

AND typically U-channels are used for panel stiffening... rarely torsional loading... except for incidental secondary reactions.

Regards, Wil Taylor
o Trust - But Verify!
o For those who believe, no proof is required; for those who cannot believe, no proof is possible. [variation, Stuart Chase]
o Unfortunately, in science what You 'believe' is irrelevant. ["Orion", HBA forum]
o Only fools and charlatans know everything and understand everything." -Anton Chekhov

 

[rb1957]

given your section, couldn't you say that the U-channel is doing all the work, the the skins are "going along for the ride" ?

you could say that the torque is being reacted/applied as a couple (on the vertical faces) and the rivets transfer the proportion of the load equal to the thickness ratio, yes??

"Hoffen wir mal, dass alles gut geht !"
General Paulus, Nov 1942, outside Stalingrad after the launch of Operation Uranus.

 

[SWComposites]

First, that is not a pure U channel, it is some sort of fitting.

If I understand correctly, you have a torque on the U/fitting, which is reacted by shear forces in the attaching fasteners, vertically up on one side, down on the other.

Once you have the fastener forces, its simple to perform hand calcs for the various joint failure modes - bearing, shearout, net section, fastener shear.

 

[SWComposites]

And if that fitting is a actually a long part (please show the entire assembly), then only part of the torque will be transferred to the skin. In which case, model the fitting and skins with separate shell elements, connected with fasteners (cbushs if using Nastran)

 

[SteveAero]

Hi,

Thanks everybody for the replies. I have provided more detail in the attached sketch.

What is interesting for me is that there is no established method to transfer the torsion loads through the fasteners to the skin. For transverse shear, we have a V*Q/I x pitch type approach. I can understand the philosophy of using differential bending couples and distributing the force to the skin using relative area. What is interesting however is that from my 'simple' FE investigations, it seems to indicate that much more force is actually transferred through the fasteners-which tells me that something else might be going on here....most likely just highlighting my inadequate FE modelling abilities...

 

[SWComposites]

You have extracted the "torsion" for the combined section. This is not helpful. You have a complex redundant structure and need to separately model the individual parts, with fastener connections between the parts.

 

[rb1957]

ah! that explains a lot ! so you're transferring the torque from the top rib down through the skins to the lower part of the structure, yes?

so the torque on the rib, becomes a couple in the skins. so ...
1) how many fasteners to transfer the couple force ?
2) is this reasonable given how the torque is measured/applied ? over how much of the rib ?/
3) is the torque coming in through the three pairs of holes at the top, and is the rib distributing this over more of the skins ??
4) what are the stresses in the skins ?
5) what is the allowable, particularly in compression, of the skins ?

"Hoffen wir mal, dass alles gut geht !"
General Paulus, Nov 1942, outside Stalingrad after the launch of Operation Uranus.

 

[rb1957]

For what its worth (which probably ain't much) I think you model, though very detailed is missing key features.

The primary question you're trying to answer is how doe the rib distribute the applied loads into the skin. By combining the rib walls with the skins, you are defeating the FEM. You need to model the rib and the skins separately (if this is what you're after).

I would measure the applied torque as the load into the rib, like at the holes/fasteners.
You can use this model to say so much of this torque is reacted locally, and so much is distributed along the rib.
I would probably use unit loads to figure this, and a bunch of hand calcs.

Maybe you should've modelled the rib as a beam element (and saved a tonne of electrons the inconvenience of working for you.

"Hoffen wir mal, dass alles gut geht !"
General Paulus, Nov 1942, outside Stalingrad after the launch of Operation Uranus.

 

[Stress_Eng]

Just for clarity, can you show how the model is being loaded and restrained. Is the torque shown the only load? Having a full picture will give a better understanding of the local load path through the structure and how the parts interact. In the picture, is the single torque applied as differential loads at the two closest holes, to the skin and/or the fitting? Are the other holes being loaded at the same time and in the same manner?