The relationship between torque and the amount stretched (of/in a bolt) 2

[bneupa1]

Hey guys,

Let's talk about connecting rod bolts. We all know that it has to be bolted to a specific torque value during installation. The pump I am working with is requires 140 ft/lbs of torque on these bolts.

Torquing a bolt simply means stretching it so that the shank and the threads are under constant tension (the head of the bolt being under compression). So, my question now is, if I know that the bolt of a given length has to be stretched out by 1/16" (as an example), will I be able to calculate the torque required to do so? Basically reversing the known variable.

Of course, we would know the initial and the final length and the material properties; however, after going through a couple of forum, I am worried about my friction factor or the Nut factor.

Any help regarding this topic would be appreciated.

Thank you,
bneupa

 

[CoryPad]

If you want to equate torque and tension, then you need the nut factor for the short-form equation or you need to use the long-form equation (see faq725-536).

 

[MintJulep]

Conceptually, the relation between tension and elongation is given here:

http://www.engineeringtoolbox.com/bolt-stretching-d_1164.html

Note that the effective diameter changes between the smooth and threaded portions of the bolt shank.

The FAQ that CoryPad links to gives you a relationship between torque and tension.

So mathematically is it straightforward to relate torque to elongation.

In reality, there is so much uncertainty with friction and nut factors that the calculation approaches worthlessness.

If someone has taken the time to specify bolt elongation rather than giving a torque it is good bet that they have already determined that torque control is not satisfactory for that joint and they want you directly measure elongation.

 

[desertfox]

Hi

Try this link

http://www.torqueleader.com/linkservid/EAEA3A24-90F4-30D2-6FA4D3F056E832EE/showMeta/0/

 

[Tmoose]

I would start with a chart like this -

http://www.spaenaur.com/pdf/sectionD/D48.pdf

And compare my calculated results to their "standard" values and similar values of others.
page 32 here for instance -

http://www.holo-krome.com/pdf/Consolidated_Tech_Manual.pdf

As far as (justified) concerns of friction uncertainties, this moderately prestigious outfit (ARP) seems to believe that if I tighten and loosen a screw or a bolt/nut combo 8 times the clamp load will approach the same asymptote whether lightly oiled or lubed with very slippery stuff.
ARP has in my opinion been wrong on at least one detail of bolted joint practice they published on their website a few years back, but I am inclined to believe this one for now.

http://arp-bolts.com/p/arpultratorque.php

I guess The corollary is that //if// the popular assumption is torquing a fastener "to spec" the first time with super lube would overstress it, then buffing down the friction surfaces by torquing and loosening it 8 times when lubed with oil would damage it on the 8th and final assembly the same way too.

I would not read too much into that in regards to designing a bolted joint.

Rather, I try to design my bolted joints that the goal of the specified torque is more like scoring a field goal ( clear the bar ) rather than insisting Corn Bag In-The-Hole is the only way to win.

 

[ColinPearson]

@ the OP; You may very well know this, but torquing and tensioning are pretty different and IMO, MintJulep is right that if someone has spec'd out the elongation, you're probably going to require a tensioning procedure. I don't mean to tell you stuff you may know, but it's worth mentioning that bolting material appropriate for torqueing are not appropriate for tensioning.

 

[drawoh]

bneupa1,

Let's keep this simple.

There are simple, empirical formulae for calculating the torque needed to apply tension to a screw. They are limited by the fact that torque is very approximately 90% of that torque. The resulting tension varies accordingly.


--
JHG

 

[dvd]

The best way to correlate tension and torque is with a Skidmore-Wilhelm bolt tester. There are too many variables to calculate tension as a function of torque.

http://www.skidmore-wilhelm.com/collections/bolt-testers

The Research Council on Structural Connections has some good commentary on torque tension testing in sections 7 and 8 of the link below:

http://www.boltcouncil.org/files/2004RCSCSpecification.pdf

 

[drawoh]

torque is very approximately 90% of that torque.

I meant friction is very approximately 90% of that torque.

--
JHG

 

[bneupa1]

Guys,

First off, thank you for the responses. There is a lot of good information from physics point of view and from market available devices on bolt torquing in the above comments if anyone is interested.

To summarize my understanding, relating the amount of torque applied on a bolt to the amount of elongation the bolt undergoes is not a simple thing as its formulation depends greatly on the friction present/ nut factor. This is true to a point that it begs a question if this conversion process is even worth carrying out for practical purposes.

On the other hand, torquing and tension-ing may not always be the same process and hence cannot always be substituted for one another.

Now, let me purpose a solution and see what you guys think. I will start off in the next comment below so that it's easy to follow

 

[bneupa1]

I have a pressure vessel (cylinder). The top cap and the bottom cap (or base) is held together using 12 different foot long bolts, in circle along the edge. The body of the cylinder (with lesser diameter than the top and the base, thus inside the cage that these bolts make) is now under a constant compression since these bolts are tied down. It is important to know the torque value before I disassemble the bolts and hence the structure, so that I can torque it to the same amount after I put the vessel back together. I don't know the torque value, but I have a foot long stretch gauge.

If I can find out by how much the bolts stretch out as I loosen them, I can perhaps bolt them to the same value later while assembling.
I know the elasticity of the bolts is one major concern here, but do you think I can assemble the structure to "appropriate"/ "applicable" value given the bolts stay under it's elastic limit.

Thanks,
bneupa1

 

[MintJulep]

In theory, if you can measure the installed length and the removed length then you know the elongation, and you tighten your new bolts the same way - by measuring elongation.

In practice the head and shank end of a bolt are rarely flat and parallel to each other. This makes getting accurate and repeatable length measurements difficult.

 

[rb1957]

you could take and retorque the bolts as installed, gradually increase the torque setting untill you can click on some more torque.

but i think it's better to find the assembly drawing, so you can see the torque that the designers intended.

or you can calc your own preload; you know the pressure > the load on the end cap > the load in a bolt. T = Pd/5*1.5 ... 1.5 to cover the scatter. now what does 1.5*P do to your bolt ? less than Fty is better than less than Ftu, more than Ftu is bad ...

another day in paradise, or is paradise one day closer ?

 

[Screwman1]

I would serialize the bolts so you can get them back into the same location, center drill both end of each one and then measure the installed length of each bolt using a ball end micrometer. for reassembly you just re-tighten to the same bolt lengths and you will have very close to the same clamp load on each bolt. This will be much more accurate than any method that starts to involve torque into the process.

 

[Tmoose]

In the OP the OEM's torque spec for the conrod bolts is known. Unless perhaps the pump rpm are increased substantially I would expect that assembly method to work just fine if threads and underhead surfaces were in good condition. If I was curious as to the OEM's desired clamping force I would consult a few generic torque tables for a practical first approximation.

If I was concerned about real or imagined bolt fatigue failures or loosening in service I would compare those generic clamp loads to be sure they significantly exceeded my calculated estimates of maximum joint tensile loading. If all appeared well I would find a way to measure actual elongation, AND ensure that threads and underhead finishes were in good condition on the next pump I built, as crummy surfaces bring the risk of embedment which can easily lead to loss of preload and subsequent fastener loosening and fatigue. Devices like Loctite and safety wire may resist bolt/nut rotation, but do nothing for preventing loss of preload due to embedment, and preload is what keeps fasteners tight and reduces risk of fatigue failure 90% or so for free.

=====================
ARP, a premium aftermarket automotive fastener supplier says Loosening and tightening a bolted assembly a few times can smooth up the threads and other counterfaces and thus help assure full and proper bolt clamping.
page 9 here -

https://www.northernautoparts.com/_f...esentation.pdf

It is interesting that ARP says the end result is the asymptote of either the first assembly with very slippery bolt lube, or 8 loosen/tighten cycles.
One of the often mentioned concerns of using anti-seize or moly on bolts or other fasteners is that the the fastener will be overstressed if the original dry or lightly lubed torque is used.
ARP's chart suggests the 8th time the nut is tightened the stud will be (over) tensioned just as bad.

===============
Next Jumping to the OP's hydraulic cylinder example.

It would be my luck to take heroic measures to duplicate an improperly built cylinder tie bolt assembly.
As part of the investigation I would attempt to get the OEM's tightening method.

I believe some name brand industrial cylinders are sometimes designed to be assembled successfully by using tie rod torque.
Page 6 = last page here -

http://www.parker.com/literature/Industrial Cylinder/cylinder/Serv/1110-M3.pdf

One real nice thing about a long tie rod (unlike a con rod bolt) is the elongation will be relatively immense. A few thousandths of embedment will not change the clamp load much at all.

There was a time when ski lift towers used long spacers on their anchor bolts. These days I expect the installation spec might include a turn of the nut spec.

http://www.newenglandskihistory.com/lifts/NewHampshire/attitash/summittriple-2014-0914e.jpg

As I rode the lift I would impress my cute captive audience by theorizing the intent was to provide lots of fastener stretch to accommodate embedment etc and maintain preload. For some reason most skied off quickly to "catch up with their friends."

 

[desertfox]

Hi bneupa1

Well now you've mentioned a pressure vessel I have several questions:-

1/ is the vessel designed to a code.

2/ How do you know it's okay to reuse the studs, they may need replacing.

3/ is there sufficient length of thread for the studs to be tensioned with a tensioning tool.

4/ how are the end caps and cylinder body sealed, by a gasket?


If the vessel is designed to a code then the stud preload will be calculated from that code and that's the value you need