Bolt Prestress

No. You can't really translate rotational force into bolt elongation, which is how you know if a bolt is "tight".

The question becomes, how much elongation do I induce into this bolt to achieve the desired compressive force on the joint. Elongation is immune to surface finish of the fasteners, torque is not.

The easy way to get close is torque.

Hello Lyn

I think you're answering a question he hasn't asked. Agreed the bolt stress induced by a given torque is quite variable, but he hasn't got that far, he doesn't know what stress he should be aiming for in the first place. He hasn't said how he's going to achieve the right stress when he finds out what it should be.

According to my data, B7 bolts have UTS 860 MPa, yield stress 730 MPa (85% UTS) and tightening stress (back-calculated) about 90% of yield stress. For other bolts tightening stress given as 80% and 75% of yield stress. Some of this data is quite old and the later stuff has the lower figures, so I would suggest pdiculous963 goes for 75%.

As you say, there are ways of determining bolt stress other than from applied torque.

Cheers.........Codey

Yes,

You are right. I mis-read the post.

Thanks!

The problem with bolted joints is that too often one applies a "recipe" and is surprised that the bolt either broke under load or manifested a relaxation and allowed leaks. Even those so called allowable stress levels depend on friction and this is due to the fact many forget that bolts are at same time loaded in tension, torsion and some times in bending ! The combined stress (usually according to von Mises) should be under the yield stress and should take into consideration the load variations in order to avoid fatigue failures. The more the load can change the lower the level. But one of components is torsion which is function of friction in the thread so even for same axial load(and assuming bending=0) the stress level computed ONLY based on axial load can be inappropriate. I mentioned several times and I am not the only one that giving a basic rule for tightening without a complete view of the case is a dangerous exercise which can have fatal ends. In such a situation as the one explained in the OP it is better to say "look in a special source book" and recommend one (Bickford for instance or a German VDE/VDI norm which became ISO) in stead of giving recommendations which can lead to accidents. Basically 3 conditions have to be fulfilled AT SAME TIME : - maximal load leads to a stress < yield stress considering also the material property and geometry dispersions - the stress variation in the most loaded bolt section should be < the fatigue limit - the minimal joint compression, when pulling load is at its maximum, should be > minimal contact pressure required to keep the joint tight. If those conditions are OK the joint will work properly and for long times. In hot joints creep should also be considered as a 4th condition. This cannot be answered by a simple "bolt calculator" as it seems to be the OP intention.

Hello nick, been meaning to reply for a while

I'd agree it's always better to measure the actual stress if possible e.g. by bolt extension, but a few comments.

If a bolt is loaded in bending I would think that's usually due to bad design or bad installation e.g. tightening one bolt too much instead of tightening all of them a bit at a time.

Re fatigue - it depends on the application. If it's structural, usually the bolts can be tightened (under no-load) to > applied load. Then when the load is applied, the clamping force between the members is reduced by the applied load, but the bolt stress does not change (that's the theory anyway). So there is no fatigue and according to my data the bolt life is infinite, which does seem rather optimistic! but we know what they mean.

But it's different for a flanged joint. First there's a maximum bolting-up load to avoid overstressing the gasket. Then when pressure is applied, it would be nice to think (from the sealing requirement, but not from bolt cyclic stressing) that the gasket load would stay constant but the bolt tension increased. If we can't have that, one might assume the worst case would be the gasket load decreasing by the applied load, the bolt tension staying constant. But real life is even worse than that, the flange deforms elastically, the bolt load decreases and the gasket load drops by the applied load + the loss of bolt load. So there is a fatigue risk, but from flange data the recommended bolt torque is well below what's normal for the bolt, even for low grade bolts. Though it does say the bolts must be lubricated with automotive grease. If there's going to be a large number of pressure cycles in the design life, the designer should do a proper fatigue check, as you mentioned.

According to my bolt tightening data "The torque figures quoted are approximate and are applicable to bolts in the self-colour condition only. They do not take into account the effect of plated finishes, special lubricants or hard and smooth mating surfaces such as hardened washers etc". The torques quoted are based on bolt stress 80% of yield stress. I assume this is from tests of measured induced stress in the self-colour condition. It would be interesting to see data on torque needed for various other conditions, are you aware of any?

I've always greased bolts before assembly, including car wheel nuts, for corrosion prevention and ease of dismantling next time. That's with automotive grease, I don't know whether that counts as a special lubricant, but I've never had a problem of bolts breaking, thread stripping or working loose. My theory is the high contact load cuts through the grease so doesn't reduce friction much when it matters. Somewhat analogous to greasing battery terminals. One might think that would cause electrical problems due to increased contact resistance, but in practice any such effect is unnoticeable, I assume due to high contact pressure, and far outweighed by benefit of avoiding corrosion.

Cheers.......Codey

Hi Codey,

I hope you would not mind if I feel obliged to make a couple of remarks.

Because of the remarks I got lately I am very cautious in making remarks since there could be some to react again in an aggressive manner.

The figure shows the most usual model for an assembly. The lines represent:

- from left up to right down the flange + gasket in a field deformation - force

- from left down to right up the bolt in same coordinates.

Their crossing point is the position where the pre-loading is obtained.

The "rounded " zones at zero force are the zones where the surfaces are not yet "clamped" but only in contact over limited regions. The rounded zones at the higher forces are the transition zones from elastic to elasto-plastic behavior. In fact neither the bolt nor the flange + gasket do behave in a perfect elastic manner but this approximation is good enough.

What will happen if the flange under the bolt heat is pulled by a force ? This force will AT SAME TIME load the bolt AND unload the flange + gasket. How much ? Till the force between the 2 lines will be equal to the external acting force. As you see the force IN the bolt will go up and of course the force which compresses the gasket will go down. So the theory does NOT assume that the bolt tension stays constant.

If you look at the 2 lines right of the crossing you will notice several aspects:

- If the force is too big either the bolt will be stretched over its yield strength or the gasket will be pressed with a too low force and loose its capability to maintain the assembly tight.

- If the force on the assembly presents changes due to ANY external load then the force on the bolt will vary and if the VARIATION leads to a too big stress change then the risk of fatigue failure is real.

As you very well put it down there differences between structural and "mechanical" bolted joints. But all of them are tightened without load.

The biggest risk for flange+gasket assemblies is the gasket setting under load which leads to a lost of "geometric pre - loading" and thus to a decrease of the force pre - loading. At high temperature this setting is also connected to some creep.

Now with respect to bending. In a flange assembly bending can occur if a too thin flange was too tightly pre-loaded since the bolts are outside the gasket and this generates a bending moment. Due to thread bolts are very sensitive to such loads (eccentric) since the thread has a high notch effect. Since all structures deform, more or less, those deformations can affect the connecting elements and "lateral sliding" can lead to bending of the bolts.

In fact using grease the "assembly" friction goes down but the grease is expelled and the friction after a quite short time under stress is high enough to maintain the assembly tight. You are very right to use it. The grease MUST have such a viscosity that it can be expelled. This is the reason not all grease types can be used with success.

With respect to electrical contacts I can give you an even better example. Torque transducers for bolting have most of them a strain gage bridge on a rotating shaft which is connected with the stationary part via a sliding ring with 4 sliding contacts.

Those sliding rings transmit very small currents at low voltage differences, any problem with the contacts could lead to high errors in the torque measurements. If the contacts are dry it follows a high wear but if the contacts are "greased" wear is quasi nil and the obtained repeatable accuracy is under 0.25% of FS!

When you grease the bolts you obtain a positive result because all of them have almost SAME friction so that at same torque you obtain same pre-loading! Uniformity is a very positive aspect. Consider that usually friction can vary ± 10 to 20 % and this reflects directly into the pre - loading.

I hope it is of interest for you and if you have any other question please feel free to ask or discuss any point you consider of interest.

OK thanks nick, I'll study your post in more detail when I have a few minutes and maybe come back if I think of any more questions/comments.

Cheers.......Codey

I expect you will never find a good answer. The percentage of yield probably can be stated fairly easily, but the torque to achieve it cannot because the frictional components are very variable. But I would expect the percentage to be higher than 75, because some tightening methods use turn-of-the-nut which I think is intended to yield the bolt a bit. Measuring the elongation may be the best method, but a torque wrench is the easiest! Either have to be calibrated.

Application would change the percent of yield strength too:as example,a 34% is taken sometimes for reference of fatigue limit,not being exact.In fact that limit depends on things like simetry of efforts.When load is really constant you could get closer the highest value and consider creeping effects...I guess the Eiffel tower was made with a little bit more of faith than science knowledge.Is it?

What u mean by exact bolt torque?

U can find the max. bolt torque to be applied.

First determine the allowable bolte torque (depends upon bolting material)

E.G. for for B7 bolts,

Allowable Bolt stress, ó = 30000psi

Bolt Preload, Fb = ó .Ar = 18172.32 lbf

Torque, Tb =Fb .dn ·µ = 257.441 ft lbf µ = friction factor, dn= nominal diameter.

The applied bolt torque should be less than or equal to Tb.

Your equations are uncomplete. Be responsible do not give counsels when your knowledge is limited with respect to the required.

It is impossible to have an answer to 6 or 7 significant figures when an input value is good to only 2 significant figures (30000.)

From the text I used many years ago Fi = 0.75 Fp for reused connections, Fi = 0.9 Fp for permanent connections. Fi is preload, Fp is proof load derived from proof strength. Hope this helps.

This is a link to the specification for high strength structural bolts

You may pull some guidance from it.

Thanks for everyones response, the more and more I research, the more I learn that there is no one answer. I understand that bolt torque does not completely indicate how tight it is, but I think what my company is looking for is a good approximation. But we use many different size flanges and different gaskets that make the question much more difficult. Thanks again.