Applying Bolt Head-Clamped Part Friction 5

[bscotti]

Well, here's my problem. I'm trying to figure out a way to apply the frictional torque between a bolt flange and the clamped part to the clamped part, in SDRC. What I'm attempting to learn is how much stress is induced in the part from the bolt torquing procedure.

Thanks
Brian

 

[vonlueke]

Assuming thread and bolt head washer face coefficients of friction are equal, about 47% of the bolt installation torque T will be resisted by bolt head washer face friction, about 39% will be resisted by thread friction, and about 14% will be resisted by bolt axial elongation. Thus, total torque induced on clamped assembly (including wrench holding nut, if any) can be approximated as (0.47+0.39)T = 0.86*T. Therefore, apply a distributed total moment of 0.47*T to the clamped part underneath the bolt head, on an annular area equal to the bolt head washer face annulus. And apply a distributed total moment of 0.39*T to the surface area of the engaged internal threads.

To apply distributed moments on surface areas, use geometry-based boundary conditions, which must be applied prior to meshing and which might require free as opposed to mapped mesh. You also need to apply the bolt axial clamping force P = T/(K*D) to the appropriate surfaces. Good luck.

 

[DarrellW]

Vonlueke, what assumptions have you made to achieve such accurate and unambiguous numbers? I dare say the list is a pretty long one. In a real world where K varies from .11 to twice that value (at least), for you to state your bolt formula without qualification is frightening. Bolt material, nut material, thread pitch, thread fit, head face profile, lubricant quantity and quality, platings/oxides and coatings, mat'l yield in relation to contact stress? - the list goes on and on.

Bscotti, please find out what the full list of special circumstances are before using any such formulas for something as difficult to control as bolting. Any multipliers to two decimal places in a world that is plus or minus 50% have to be well understood and fit the application.

 

[CoryPad]

Firstly, I believe vonlueke's technique is a very good one.

Secondly, regarding the torque component distribution (pitch/thread/bearing), I have done an analysis using the following:

ISO 68 - metric 60 degree thread form
ISO 261 and 262 - preferred metric thread sizes
ISO 273 - hole sizes (medium)
ISO 4014 - hex head screws
thread major diameter verying from 6 mm to 100 mm
friction coefficient of 0.1 and 0.3

The pitch torque varies from 2.6% to 18.6% (vonlueke's = 14%)

The thread torque varies from 36.2% to 46.2% (vonlueke's = 39%)

The bearing torque varies from 45.2% to 51.2% (vonlueke's = 47%)

A friction coefficient = 0.3 and a major diameter = 100 mm are obviously extreme values. It is important to identify friction coefficient for the actual joint members, as well as their geometry and use torque components that match your joint. Just for clarity, vonlueke's numbers correspond well with fasteners with lower friction (~ 0.1 - 0.15) and medium diameter (16 - 30 mm).

 

[vonlueke]

Thanks, DarrellW. I gave Brian a median scenario since he didn't mention specific data nor a special case. Noting his Automotive discipline, I gave an approximate median scenario of mu1 = mu2 = 0.14 for bolt sizes D = 4 to 12.7 mm, metric and imperial, coarse and fine pitch. E.g., M8 x 1.25-6H/6g-N, mu1 = mu2 = 0.14, is approximately characteristic of this median. The reason CoryPad perceived I used a slightly different median might be because his range is for larger bolt sizes, though I haven't verified this.

And thank you, CoryPad, for your comments.

 

[AhChoo]

I am trying to model two plates (varied thickness) clamped with a bolt and nut to determine the stresses at the hole edge due to fastener load transfer.
I was planning mesh both the plate and bolts using solid elements but I am not sure how the interface conditions should be.
Given the torque load, would vonlueke's conditions above model my problem accurately?

 

[vonlueke]

Brian: Notice for the previous approach I mentioned herein, you approximate the applied loads by hand calculations and apply them manually yourself. If you're instead interested in a much more elaborate model in which you simulate the torquing and rotation of the bolt in an assembly, then perhaps try SDRC Technical Support to ask them to track down this model already created, in case an SDRC expert already made one as a demo; then you could ask him/her how one might go about creating it.

 

[vonlueke]

AhChoo: The approach I described for Brian herein is for attempting to manually approximate the stresses induced in the clamped parts while torquing the bolt, which is a different problem than the problem after the torque wrench is removed. For your problem, you would not apply either of the applied torque load components mentioned in Brian's problem.

For your problem, you could perhaps define the bolt material properties at an elevated reference temperature T1, then run your analysis at, e.g., room temperature T2 to simulate bolt preload, where T1 = T2 + M/(E*alpha*K*0.25*pi*D^3), where M = bolt installation torque, E = bolt material modulus of elasticity, alpha = bolt material coefficient of thermal expansion, K = torque coefficient, pi = 3.14159, and D = bolt shank nominal diameter. Now apply your applied tensile load(s) to your structure and run the analysis at T2.

If there are also applied shear load(s) on your bolted plate and you're assuming plate shear slippage (the usual assumption) causing bolt hole bearing stresses, I guess you would then need to make the above run a contact analysis to model the bolt-shank-to-bolt-hole contact stresses. Good luck.

 

[AhChoo]

Thanks Vonlueke,

But i am still unsure of something.
I am doing a stress analysis for bearing stress due to fastener load transfer. The two clamped plate is in tension of course but why is Temperature in the picture here? Is it common practice to simulate pre-load this way?
Could you elaborate on contact analysis?
I am at a loss on how to model bolt-shank-to-bolt-hole contact.

 

[vonlueke]

Reference temperature T1, in the menu where you define your bolt material properties, comes into the picture here only to make a room-temperature analysis contract (cool down) your bolt material and thereby induce the bolt installation prestress on your bolt solid model. Some programs have a special bolt element; others don't. Regardless, you can simulate preload yourself in any program with reference temperature, as explained above. I guess it's not that common, but I'm not sure. I haven't heard all that many people talk about it yet on this forum, so far.

Contact analysis can be nontrivial, and the instructions might be slightly different in each program. There doesn't seem to be much information on contact analysis on-line yet (?). If you're really interested in contact analysis and your program has this capability, you may need to locate the contact analysis instructions or tutorial for your program. Or maybe someone knows a clever way other than contact analysis for your solid model (?).

 

[vonlueke]

AhChoo: So that you clarify the statement of your given problem, do you have a shear load applied to your bolted plate? Or do you not have an applied shear load? The discussion on contact analysis (last paragraph of each of my two posts to you, above) applies only if your bolt is also subjected to shear.