Bolt make-up to ¦3% tension 5

[HotStab]

Guys,

API 17D (Design and Operation of Subsea Production Systems - Subsea Wellhead and Tree Equipment) has included in its second edition a requirement for flange make-up with bolts at 70% yield stress with an allowance of ±3% (see below).

“Closure bolting of all 6BX and 17SS flanges shall be made up using a method that has been shown to result in a stress range between 67 % and 73 % of the bolt’s material yield stress.”

Is there anybody knows a feasible and reasonable practice that could be used to assure a bolt tension in a so narrow limit?

Thanks for helping,
HotStab.

 

[SnTMan]

HotStab, I'm gonna say that accurately measuring bolt elongation is the only thing that will get you there.

Regards,

Mike

 

[HotStab]

Mike,
I fear that, for this narrow tolerance range, even using a very precise control, all the bolts in the flange shall be simultaneously made-up and stretch controlled.
All this, in my opinion, is far from a reasonable practice for an equipment assembly shop.
If I am wrong, please, let me know.
HotStab.

 

[fegenbush]

Lubricate the threads and faces appropriately, calculate the torque required to obtain the necessary preload, and use a torque wrench. That is about the only way I can think of to obtain accurate preload in a production environment.

 

[Cockroach]

Depending on the lubrication type, quality, torque will vary much more than 3%. I suspect this is HotStab's concern, which may be a bit over the top. I would suggest that bolt tighten to 70% using a torque wrench and call it a day. In my years of practice, I have NEVER seen an inspector come around and check torque measures on bolting. Doesn't mean you didn't do it, just says that to the best of your ability, bolts were torqued to 70% given the following set of circumstances. Outline and document those circumstances and your torquing procedure.

What I would suggest is that regardless of torquing method, you have a well calibrated torque equipment and have it certified for use by a recognized authority (standard). Your procedures and Quality Control Program would automatically dictate this under your QPI. That way, you have the method and the method of measurement suscribing to a specific set of standards.

Don't get all torqued up over some 3% allowance.

Regards,
Cockroach

 

[rb1957]

yeah, what's the point is carefully controlling torque when the preload, the thing you are trying to control, is determined as much by friction ?

i'd've thought it was an analytical exercise.

i'd've also thought that design guides would have said something more like "consider the effects of preload scatter, assume adverse preload; ie for gapping analysis assume the minimum preload is applied, ..." I've heard that torque wrench preload scatter is +- 1/3 (ie between 67% and 133% of target).

 

[drawoh]

HotStab,

Do you even know what the yield stress of the material is, within ±3%?

My first thought on reading this was that you mis-interpreted a requirement to strain the bolt by[ ]3%. Upon re-reading the OP, it is obvious that you did not, but perhaps someone else did.

--
JHG

 

[dhengr]

With all the discussion we have seen and heard about torque on bolts and field conditions, and all the variables involved and the scatter in results as far as final bolt stress, why use torque as the final tightening means? You know the pitch on the bolts, and of course they should be clean, new and in good condition; then use some min. torque, by mechanical means if you wish, to draw the flanges together, then use a turn-of-the-nut method to achieve your final tightening. This is probably a more accurate and consistent method. Of course, tightening sequence is important too. I don’t know about the +/- 3% or the 70% of Fy, that was probably dreamed up by some tech. writer who had never lifted a 6" long 1.5" bolt, or the wrench needed to tighten it before. Undoubtedly he/she thought +/- 3% would prevent all future blow-outs, never mind the important thing is to be sure all bolts get tightened.

 

[Cockroach]

Typically the valve/piping codes are blindly obeyed without much thought placed into what you are trying to accomplish. I greatly appreciate these recent points raised by the forum for this reason, and my experiences with third party inspectors. Following the code to the letter of the law usually results in frustrations and misunderstanding in meaning, it seems what you are trying to do falls a distant second.

Standard error in measurement is 5%, yet you're trying to hit 3% torque allowance. If you look at the point raised by DrawOH, material yield is based on the 2% offset method, so you now introduce much more uncertainty into the measurement. So it is ridiculous to expect someone to be 70% +/- 3% within torque measure for tightening a bolt. It makes me shake my head.

I would be thinking about the spirit of the code, not the letter of the code thereof.

Regards,
Cockroach

 

[dvd]

Use a Skidmore-Wilhelm torque/tension tester to correlate the tension in the bolt with the torque applied:

http://www.skidmore-wilhelm.com/bolt_testing.htm

You may also find Section 7 on Pre-installation Verification in Specification for Structural Joints Using High-Strength Bolts to be of some help -

http://boltcouncil.org/files/2009RCSCSpecification.pdf

 

[Tmoose]

The dozen or so API specs I've played with impressed me with their practicality.
The description the OP provided is hopelessly impractical, if the intent is to truly achieve that preload. Maybe they're insisting on great precision of an imprecise process "just because."

This site seems to say API has published torque specs.

http://www.woodcousa.com/about_17d.htm

" API Spec 17D requires closure bolting tension of 67% to 70% of the yield strength of the bolts (L7 / B7); torque tables included in API Spec 17D provide torque values intended to achieve this tension."

If so, Then the quest is for a torque wrench with 3% accuracy and torque to the middle value.

 

[dinjin]

this was also in the reference that tmoose referenced
CAUTION: API Spec 17D requires flange connection bolt tightness (tension stress in the bolt) equal to 67% to 70% of the specified minimum yield strength of L7/B7 bolts. When using some lubricants and/or anti-corrosion coatings on bolts, tightening bolts consistent with the torque tables shown in API Spec 17D may in fact produce tension closer to the actual yield strength of the bolts. Operators should use an alternate method to measure tension in the bolts to confirm the applicability of this table.

 

[gasketguru]

As everyone else has sugested, torque only measures the turning power on the nut, not the actual bolt tension - you have to have a lot of faith in whatever friction coefficient you may have on the threads and under the nut. Typically, torque vs. tension will give a scatter of around 30% (on a binomial curve). The only way to get good accuracy is to axially load each bolt to the force required and accurately measure their extended lengths and then replicate this on the installed joint. (The turn of nut method mentioned above would be difficult, as the flanges start with a 1/8" gap before the ring is deformed inwards.)

When BX joints were first devised in the 1950's for 15k flanges the design premise was for 50% yield of B7 / L7, though the later higher loads for subsea items are presumably indended for higher security and to take account of the use of newer corrosion resistant alloys with higher strength.

The API standard gives torques based upon assumed friction, though to achieve a higher % accuracy, then the easiest way is with a load-indicating bolt (e.g. "RotaBolt" and others can be found from various manufacturers), where a mechanical strain gauge is installed in the bolt which is set to lock a cap at the desired axial force. These are used sub-sea with a "divers wheel" so can be checked whilst wearing gloves.

 

[WolfHR]

I'm by no means an expert, but I suspect the key to this is the wording in the requirement: "using a method that has been shown to result in a stress range between 67 % and 73 % of the bolt’s material yield stress." Albeit not being a native English speaker, I see this as: you do a series of tests with flanges, bolts and lubricants and procedure equal to those used in this application and establish a torque value that will produce required stress (say, X ± y Nm). Then it's up to the operator to tighten bolts to specified torque, and should it be required, test results will show that this procedure results in required stress.

 

[Cockroach]

You could also use a "bolt stretch" device, not sure what the technical name is. Basically it is a hydraulic device which grabs the end of the bolt and literally stretches it to a preset value, say 70% of the material strength. The nut is then tightened hand tight to the flange and the bolt let go of. The nuts are now tight to 70% the material strength with nut make-up.

I've only seen it in the oil patch a few times, believe it comes from the aircraft industry. Regardless, it would give you the accuracy you seek without the friction worries due to lubricants.

Regards,
Cockroach

 

[drawoh]

WolfHR,

The famous Mitford sisters were instructed to prepare budgets to enable them to survive on 500[£] per month. This would have been quite some time ago, so it should not have been difficult. Nancy Mitford was supposed to have started hers with the line item "Flowers:[ ]490[£]".

I think the material specifications are blowing your error budget. Everything else is added on to that.

--
JHG

 

[MintJulep]

Direct Tension Indicators, or "squirter washers" claim +/- 1.5% accuracy.

http://www.turnasure.com/

 

[CoryPad]

Ultrasonic control is much more accurate than torque. Here are two suppliers:

http://www.innovationplus.com/index.php

http://www.intellifast.de/

Hydraulic tightening is another option, many suppliers out there.

 

[HotStab]

Guys,
The concern is not an inspector check. The standard states that the accuracy of the make-up method shall be proven. In fact, not only the method used but also the operational staff shall be qualified. As stated:
“Written procedures, incorporating the features of these provisions shall be developed for use by the qualified connection assemblers. The applied torque/tension in the written procedures shall be qualified for some relevant bolt sizes with actual material, coating and lubrication.
...
The use of qualified assembly procedures and qualified assemblers is analogous to the general requirements for welds, where use of qualified welding procedures and qualification of welders is present industry practice.”
Ok, let’s qualify a method. The problem is that, for my concern, there is no practical method for this.
Any torque procedure is out. Even the temperature effects on grease shall represent more than 3% deviation.
DTIs (direct tension indicators) do not achieve the required precision and usually only focus on a minimum tension value.
Let’s be precise and use hydraulic tensioners. Only if all bolts are tensioned at the same time the scatter of tension values might be narrow enough. But this would require one precious and rare feature in subsea equipment: space. To be simple: it is not feasible!
Even if a NASA technology were used, like a ultrasonic method for stretch control, the precision would be covered up by errors cause by deviations in material properties (elastic module of the materials or yield stress of bolt material).
API is requiring a miracle and for nothing. Because there is no reason for a so tough requirement. These are my conclusions.
Thank you all for so valuable comments.
HotStab.

 

[MintJulep]

DTIs (direct tension indicators) do not achieve the required precision and usually only focus on a minimum tension value.

Good point about the minimum.