relation between friction due to pressure and torq moment on bolt 2

[jossaad]

Hello gents,

Sorry for my poor English, I will do my best to describe my case I have to resolve.

I have a steel ring devised by 4 pieces I need to tighten them around a concrete column.
The 4 parts are bolted to one another (say 6 bolts between each part).
The height of the ring is (say) 500 mm = 20 inch
The ring or called bracket supports a vertical force/ piece (say 20 Ton/ part).

I need to know how torque I need to apply to tighten the bolts to be sure that the bracket is tighten enough to transfer the vertical force by friction to the column. Besides, The vertical forces are offset (1m), which means I have Vertical forces plus a moment.

I will appreciate any help

Many many thanks

 

[tbuelna]

jossaad,

The ring segments will only be able to resist the vertical loads applied through friction.

The breakaway friction can easily be calculated if you know the circumferential hoop tension produced by the total fastener preloads, the hoop stiffness of the (hollow?) column structure, and the coefficient of friction at their interface. The offset loading point will cause an overturning moment about the bracket section, but this will only cause the bracket upper edge to unload while increasing the load proportionally at the bracket lower edge where it must react. So I don't think the friction breakaway force would appreciably change.

The only issue you must account for is the uneven loads applied to the bolt patterns that result from top to bottom in the bracket flanges.

Be conservative when calculating frictions. And be sure to take into account any thermal expansion mismatch between the various materials.

Good luck,
Terry

 

[jossaad]

Dear Tbuelna,

I appreciate your clear reply. Actually I totally agree with your analysis. However the point I have no idea about it is how to be sure that the friction is maintained. In other words, how can I assess the tension on the bracket by tightening bolts to ensure the friction? Also, how much shall I estimate the torque on bolts to be sure the bracket parts are well fastened?
For your information, the column is not a hollow one.

Again many thanks

 

[kelowna]

I guess you will determine from your design the tension required on the bolt, and you want to know the torque required to achieve that tension.

The relationship between torque and tension in a bolt is not simple. Some codes give you a relationship, but it depends of so many factors.
Is it the bolt galvanized, oiled, rusted,... you could get the bolts from the suplier (usually lightly lubricated with shipping oil) and get a relationship. You leave the bolts out for a couple of days and you get something else.

That is why I prefer the nut method to tension the bolts, you can guarantee a pre-tension of 70% (if I remember right, I do not have the code handy). However, in your case, you probably have a gap between plates, so this does not work for you.

In fact, thinking about it, and considering you have a complex system, and there is a gap between plates (or there should be) the only way to guarantee you have a given load is not to rely on torque but tension the bolts directly, with a minijack.

 

[rb1957]

are there 4 20 ton loads (one per piece) or two ?

the friction force (gosh, i'd hate to support 40 (or 80) tons by friction alone) reacts the load (including whatever factors apply ... safety factors, siesmic, wind, ???)

the normal force (creating the friction, if you will) is the vertical load * coefficient of friction (any guess ??)

the normal force is the clamp-up pressure * the contact area of the ring.

hoop load = p*R*w,
p = clamp-up pressure
R = radius of ring (of the bolts to be fussy)
w = width of ring.

and this load is created by 6 bolts.

so ...
required Friction force, F, = Load*factors
required Normal force, N, = F*k = k*(Load*factors)
k = coefficient of friction
required clamp-up pressure, p, = N/A = k*(Load*factors)/(w*pi*d)
w = width of ring
d = diameter of column
requied hoop load, P, = pRw = k*(Load*factors)*R/(pi*d)
R = radius of bolt location (= PCD/2)
required bolt preload = P/6 = k*(Load*factors)*R/(pi*d*6)
required bolt strength = preload/0.7

 

[BobM3]

Is this standard practice for structural supports? Seems real risky to me to support that load only with friction.

 

[jossaad]

Hello everybody and really thank you for ur support.

What I igore is this formula about Hoop load, would you please expand or give me a site explaining it.

Also, I do not understand "R = radius of ring (of the bolts to be fussy)
" !!! why you take the radius of the bolt

to clarify my case , the bracket is divided by for pieces and each piece support 40 T. the parts are connected to one another by 6 bolts, the diameter of the column is 2.5m. the height of the bracket is 0.8m.

Besides even if I have a feeling it might be done I have a doubt about the whole faisability but your help is priceless.

 

[rb1957]

the 2nd time i explained it as ...
R = radius of bolt location (= PCD/2)

basically i'm saying that the bolts form the loadpath, acting like a hoop (clamp), their radius is their PCD/2.
clear as mud ?

hoop stress (refer pressure vessels) is a more familiar expression, the circumferential stress generated in a cyclinder under pressure, stress = pR/t
but in your case hoop load is more appliciable to your problem .. load = stress*area = pR/t*(wt) = pRw

 

[jossaad]

Dear rb1957

After googling PCD I got it.

Sorry for that actually is simple but as I m not familair with english vocabulary.... sometimes it appears I have nothing to do with structure design

Nevermind, It makes me laugh.

Your reasonning is good.

I will check further about the hoop stress.....
It s because it might be fabricated so no room to risk..
Thanks a lot everybody.

 

[Ron]

jossaad....be sure that there is a large gap between each of your sections. You do not want them to "bottom out" so that the bolt clamps the two pieces flatly together. If that happens you will continue to put tension in the bolt, but not in the "hoop".

You will need to develop a torque-to-tension value using a Skidmore-Wilhelm device, or you can use one of several torque-tension bolt relationships (search this site...it has been discussed before).

 

[jossaad]

many thanks Ron.
I actually have 15 mm gap. may be i shuld rise it to 2cm. why not

 

[GregLocock]

Why not? because in the worst case you'll encourage high stresses locally in the column. Not necessarily a big issue, but something to watch for.

The big unknowns here are the friction coefficients (you have two of interest).


What you are doing sounds reasonable to a mechanical engineer, but you are dealing with something of the size of a structure, it may be that there are practical reasons why structures are held together using the shear strength of bolts rather than the friction developed at the interface.



Cheers

Greg Locock

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

 

[unclesyd]

Isn't he going to have a large moment on each flange especially at the top bolt hole?
Depending on the ring plate thickness the flange might have to be stiffened to prevent bending.

 

[rb1957]

yeah, end-pad bending is just one of the many things to be checked ... i think he'll have a god-awfull big bolt load what's a reasonably coefficient of friction ? 0.1 ??

 

[unclesyd]

Here are various sources for COF values of steel vs concrete it looks like a value of .40 to .5 is used.

Thread167-11539

http://www.supercivilcd.com/FRICTION.htm

http://cedb.asce.org/cgi/

http://WWWdisplay.cgi?9000548

 

[rb1957]

this is the wrong way round isn't it ?

"required Friction force, F, = Load*factors
required Normal force, N, = F*k = k*(Load*factors)
k = coefficient of friction"

shouldn't it be Normal force = F/k
(the lower k is, the higher the normal force is to generate the required friction force)

 

[Ron]

Greg...we actually use both types of connections in buildings. Non-friction is a bearing connection while friction connections are slip-critical and require tension development in the bolt.

I agree that too much gap will cause a concentration at the contact point of the hoop, while also potentially inducing bending stress in the hoop that could actually reduce the friction interface, but some gap needs to remain to assure that you don't bottom out before the hoop tension is developed. This is the point that unclesyd made...induced moment at the bolt connection.