Assembly Torque Values 4

[miningman]

I'm surprised I cant get the answer to this using Google... maybe I'm using the wrong search words.

I am installing 16 inch diameter steel pipe, to be used to pump water against 1000 feet head, so duty will be approx 450 psi at the pump outlet. Pipe has been engineered as Class 300, which I believe is rated at about 750psi, so no issues there.

What is the required torque value for the assembly of the flanges?? We are in North America, so no references to European standards please.

 

[HeviiGuy]

gasketguru, do you hear this: "mrrppph mrphhmrmm"? It's the sound of my foot in my mouth. I am very sorry for my rude insinuation. I had looked at other posts of yours, expecting to see references of the particular bolts which you had mentioned. None were to be found. Only, astute, non-commercial advice. Please accept my sincere apologies.

After seeing so many obvious commercial placements in these forums, one becomes too cynical :-(

With all seriousness now: Perhaps the fact that you experienced such atypical results were indeed due to some unfamiliarity with the process. A few things to keep in mind:

Unless all bolts on a flange are being tightened simultaneously (aka: "100% Tensioning"), different hydraulic pressures are required for successive placements of the tensioners. For example, when enough tensioners are used so that 50% of the bolts can be loaded at once (ie "50% Tensioning"), an "A" and a "B" pressure need to be calculated and applied. The former is higher than the latter to account for load transfer. It's a relatively easy calculation.

When less than 50% tensioning is performed, the calculations tend to get a bit more complex.

In any event, a very common problem experienced by people new to the technology is not energizing the tools up to three times at the same pressure. This must be done to ensure that the joint has properly "settled". If this isn't done, yes indeed, one may get unexpected results.

Don't give up on tensioning. As you say, it's a very fast method and, if done properly, it's also a very accurate method. One more thing - it acheives even gasket crush since the bolts are tightened simultaneously .

Ciao,

HevïGuy

http://www.heviitech.com

 

[unclesyd]

gasketguru,
How were you seating the nut after appying tension to the stud?

 

[gasketguru]

Trust me - after 25 years of doing this, I have a lot of experience with both torque and tensioners, and as you say the best way to ensure correct load is by measuring the extension, either by by micrometer, ultrasound or indicating bolts (it is just that those I mentioned have helped me over the years). The point is to stretch the bolt to the required amount over the desired clamping length(using a tensioner here really helps!), measure the loaded fastener length, and then try to replicate this length in the final settled installed condition. Within the accuracy of measurement method used, you then get pretty much the load you want.

The problem comes when simply replying on the recommended overload pressures from the tensioner manufacturers, as these are based upon an assumed loss per inch of bolt. The case I described was a high temp HX which had been a problem leaker for many years, and has now worked fine for the last two. With 60 bolts of around 2" diameter then the pattern was the usual A/B pattern you describe.

I do not know of any tensioner overload calcs that account for gasket/flange rigidity (like the ASME J factor). Integrating something like this into the overload calc might help, along with research into the loss per number of interfaces in the fastener (i.e. bolt vs. studbolt, and with or without washers, hard vs. soft gaskets etc.) as each piece creates some of the load-transfer relaxation in the joint. Maybe a good research project for someone like PVRC or ISO TC74 etc.?

 

[HeviiGuy]

I really don't want to flog this dearly-departed horse any longer but, I'd still like to add a couple of farthings: ;-)

There are two reasons for applying a higher hydraulic pressure to tensioners than what at first blush may seem necessary:

Although the constituent components of a joint certainly do have a great deal of impact, the calculated bolt load is a non-variable. And, since the area of a hydraulic tensioner is also fixed, the hydraulic pressure required to apply the bolt load is fixed. The reason for over pressure in this case is to address the issue of thread relaxation and consequent load loss. This is usually a very small amount.

The other reason for a modified pressure is to account for load transfer. Load Transfer could indeed result in huge variations if not accounted for. Certainly, if a small number of tensioners is being used (relative to the total number of bolts on the flange) one can apply a single pressure to the inter-linked tensioners by following a convoluted "pattern", much like torquing. However, since one of the most benefical traits of bolt tensioning is speed due to the ability to simultaneously tighten every bolt, a method was derived to compensate for this load transfer. Hence, the A/B pressures for 50% tensioning. Alas, this quickly becomes a potentially-complicated A/B/C/D/E... process whenever even fewer tensioners are used. One can see that if relying on a tensioner manufacturer's suggested operating pressures, the results may not turn out as expected if all prior assumptions are not valid.

Your points about the other compelling factors in the joint "system" are well taken and appreciated. However, my gut feel is that this atypical experience was due more to procedure rather than an inherent deficiency of this well-proven technology.

Suggestion:
An easy way to verify or modify the tensioner pressure ratios (again: not necessary at all with 100% tensioning)is to measure the resultant stretch (or twiddle the indicators) after the intial pass per set of tensioners. Then, if necessary, simply increase or decease the hydraulic pressure so that consistent loads are acheived. As long as everything remains constant (ie process, flange condition, gasket etc), these new tensioning pressures would be the ones to use for future assembly. One very important thing to watch out for is the capacity of the flanges. You'll need to verify that any additional load wouldn't result in flange rotation!

Ciao,

HevïGuy

http://www.heviitech.com

 

[mkkessel]

Late to the party, but here goes. The torque range is dependent on the minimum sealing stress and maximum load of the gasket. The following data assumes using a non-asbestos compressed gasket material, specifically Thermoseal (Klinger) C-4433 X 1/8" Thick Ring Gasket. Grade B7A Bolts. Flat face flanges. (450 psi X 16" pipe for hysdrostatic end thrust calculations.)

Minimum Seating Stress @ 450 PSI Internal = 5511 PSI Flange Load
Maximum Allowable (Gasket) Stress = 13808 PSI Flange Load

Torque values assume mineral oil lubricated bolts seating against hardened washers. Three tightening passes in a star pattern to achieve a (theoretically) even load. Also want to shoot for a reasonable bolt stress to achieve a "tensioned" flange unit, capable of compensating for creep relaxation and possible unloading of the flanges from external forces.

@5511 PSI the torque value is 374 ft. lbs. bolt stress is 18% (Min sealing stress)

@13808 PSI the torque value is 905 ft. lbs. bolt stress is 46% (Max allowable stress on gasket.)

A reasonable range of bolt stress is 30-35% which equates to a range of 590 to 688 ft, lbs applied torque. This correlates to an effective gasket stress range of 9132 and 10654 psi gasket load.

The caveat here, of course, as stated in arguments above is the accuracy of applying bolt tension via torque to achieve the gasket stress. The values given here are strictly for gasket sealing purposes.

Values provided by the Klinger Expert gasket software program.

"Tighten it until you hear the bolt snap, then back off half a turn."

 

[miningman]

Mkessell, late maybe, but a very timely and usefull contribution. My planning for this is progresing rapidly, and I know I am looking in the 600 -1000 foot lb range. I also now understand that the optimum torque is dependent on both the allowable stresses in the gasket as well as in the studs, as well as some sound engineering judgement. You clearly have the background to make a reccomendation and I would be gratefull if you could repeat your analysis using Class 300, ASME B16.5 Spiral Wound Gaskets / flanges, grade B7 studs. Pressure will be closer to 500psi. I do not have enough background in this area to apply the appropriate engineering judgement.

The Garlock website seems to suggest a preffered torque of 1000 foot lbs but unless I have misunderstood some of the previous posts, I find it significant that this is a "preffered" rating as opposed to "recommended".

Am I mis-reading something, or does good engineering judgement suggest a slightly lower value.?? I would hate to have my crews user a higher than necessary value and have leaks develop, say because the studs were over stretched.

 

[mkkessel]

The program I was using is strictly for Klinger materials and their sealing properties. Spiralwounds are a different animal in terms of the necessity of compressing the winding to affect a seal, using the outer guide ring as a "stop." Theoretically the stop prevents over-tightening however on b16.5 150# class flanges the Y value exceeds the point where flange rotation becomes a potential. Probably not valid with 300# flanges.

Flexitallic offers a pretty good technical manual which should get you in the ballpark. Check it out and I believe you will be able to calculate the necessary torque values for your application. I will attach the file.

Finally, the Garlock recommendation is also material dependent and their suggested torque range for a similar material is 401 ft. lbs. min and 912 as "preferred." I can't comment on why they would use the term preferred, other than again, ballparking a bolt stress range. It's instructive to recognize that both Garlock and Klinger's interest and expertise lie with the affectation of the seal rather than the science that effects the bolted joint.

 

[mkkessel]

Sorry, it appears as of the file did not upload, here's the URL:

http://www.flexitallic.com/pdf/designcriteria.pdf

 

[HeviiGuy]

C'mon folks: Although you're getting torque values from the "experts" you've got to take this with a huge grain of salt: "torque" does not and cannot be considered as a metric of applied load!. It's "preload that you must be concerned with. The most practical way to verify preload is to measure the resultant stretch of the applicable fasteners.

Ciao,

HevïGuy

http://www.heviitech.com

 

[mkkessel]

Hevi: You're absolutely correct, and your reasoning is why 85% of gasket failures occur. I know the corresponding error rates of hand tools vs. calibrated torque wrenches, vs. hydraulic tensioners. I'm aware of the fact that surface rust is rarely cleaned from nuts and bolts in the field. Nine times out of ten USED bolts nuts and washers are re-used, with little due regard for condition. Loosen the bolts, spread the flanges, remove the old gasket and replace with a new one. Next to expansion joints, the bolted flange is the weakest link in any piping system because sophisticated engineering rarely translates into field practices.

Other than M&Y values, which is at best sorely misunderstood and outdated, the second most popular question coming from the field regarding gasket installation is exactly what was asked right here. The results are legitimate, come from published known physical properties of gaskets and fasteners and represent good science. In that regard the opinion is indeed, "expert."

Your argument can't be against attainment of these recommendations, because utilizing any method even as primitive as a hand wrench might attain the necessary goal of creating a uniform leak tight seal, with appropriate load applied. Therefore your argument must be for "verification" by way of the accuracy in the measurement of bolt strain.

I agree 100% that your argument embodies best practices, however is in reality a "control" issue in terms of installation. Given the variables likely to be encountered in the field, the issues of both verification and control are well outside the purview of the materials manufacturer. Ultimately these practices must be adopted by the end user or installer.

 

[unclesyd]

I agreed with the previous post in that the only way to go is measure the elongation but in reality you have to look at torque values are going to be used.

Here are some torque tables eluded to in previous posts. It gives the complete range of values to seat a gasket from the min. to max.
As I've mentioned before look at the values for the 3" & 8" flange on the lower Class flanges.

One thing not mentioned in most literature is the requirement that you have enough flange to do the job. Our piping codes don't allow the use SW gaskets with store bought Class 150 flanges. Having said that we do allow SW gaskets on some home built flanges if they are of the many small vs the normal few large studs and calculate out.