Flange Peening Technique

[Bambie]

Yamoffathoo has asked me to post the following question: A class 150, 3"nps cast aluminum valve on a 90 psig x 140 degF seal oil system has 3/4" thick FF flanges bolted to 1-1/2" thick wrought steel FF pipe flanges. The valve flanges have four threaded 5/8" 11 UNC holes and O-ring seal grooves inside the bolt circle and one of the flanged joints is leaking, the suspect being a degraded O-ring.

Peening the perimeter of the aluminum flange into the steel flange is proposed, in addition, replacing the existing bolts, one at a time, with longer, high strength bolts, nuts and washers is also proposed to transfer 'wedging' loads directly to the fasteners thus protecting the aluminum flange threads.

Yamoffathoo is looking for any written procedures, standards, calculation methodology or documented experience with chisel peening flanges. He has checked in with Furmanite, RJ Stacey & CEDA but only got verbal assurances like: "That should work".

 

[Bambie]

Picture of the peening candidate

 

[MikeHalloran]

I thought using aluminum in hydrocarbon piping was illegal.



Mike Halloran
Pembroke Pines, FL, USA

 

[Bambie]

Mike, would you care to cite a reference supporting your remark?

 

[MikeHalloran]

Forty years ago, I was asked to modify a fuel oil pipe in New York. I think some applicable reg that I have long forgotten just flat required steel as the material.

Regardless of what's required by statute, or where it is, it appears that the aluminum valve body in question is telling you that it's not happy.

Peening a flange to compensate for a failed o-ring is something you do in the middle of the night in the middle of nowhere when there is no alternative.





Mike Halloran
Pembroke Pines, FL, USA

 

[Bambie]

Mike, Yamoffathoo advises me that the code of construction is B31.1 and the seal oil is for a turbo-generating unit for which there are no specific restrictions on aluminum piping or components. Degassing the generator and replacing the O-ring would be very costly.

A sealant injected enclosure has been considered but determined to be too complex and the risk of sealant ingress to the process too high. Peening aluminum into the 'sandwich' created by the steel flange and grade 8 fasteners is considered to be low risk.

The question being asked is whether the additional strain imposed on the fasteners can be quantified if the extent of plastic deformation it is controlled?

Is it a primary compressive load equal to the flange perimeter x depth of peen x poisson ratio x ultimate stress of aluminum? Would this force divided by the total bolt tensile area then indicate the increase in fastener stress?

 

[dicksewerrat]

My concern would be the crystalizing the aluminum. Or striking at a bad angle and breaking one of the ears off. What will happen if the fittings start squirting oil? Defering proper maintenance costs more than doing it right the first time. Who gets blamed when the plant goes into emergency shutdown?

Richard A. Cornelius, P.E.

http://WWW.amlinereast.com

 

[MikeHalloran]

We come from different worlds.
E.g., when you make a purely economic argument for maintaining the status quo, or rolling out some band-aid fix, to the FDA, the root issue is instantly settled, and all that remains to decide is who gets to write the recall letter for your product.

... or, in this case, schedule shutdown for appropriate repairs.


In this instance, a leak of unstated rate exists, of a fluid that I think is fairly thin and maybe a little volatile. Striking a chisel in the vicinity, or whatever abrasive machining is involved in converting blind holes to through holes, might be considered 'hot work' in some environments, so deserves some caution.

Temporarily removing one bolt from a four-bolt flange basically leaves half the flange cantilevered, which makes me nervous all by itself, and I can make and refute mental arguments for doing any chisel peening before or after bolt replacement, w.r.t. relieving any strain built up by the peening, distorting the sealing face, overstressing the bolts or the internal threads, etc., so I'm kind of dizzy from the exercise.

It's not clear if any of the blind internal threads in the aluminum have actually failed, or if the modifications are being proposed as a belt and suspenders sort of thing. Rather than all the implied die grinder work, I'd suggest replacing the bolts in turn with studs, one at a time, or maybe just leaving the bolts in place, depending on what could be determined about how they were installed in the first place and what kind of shape they're in.

Before doing any of that, if the oil is not actually spraying everywhere, I'd consider applying a thin wicking Loctite to the periphery to see if capillarity would pull it into the places where it needs to be. Okay, long shot.

Too damn bad that the valve has an eared flange. If the valve flange were round, I'd try wrapping the 'seam' with TFE tape or cord and squeezing it into the joint by wrapping it with a Band-It clamp.

Do consider Dick's advice about who gets blamed when/if repair attempts make the leak worse, or start a fire, or cause some catastrophe.


Mike Halloran
Pembroke Pines, FL, USA

 

[Bambie]

Yamoffathoo has consulted with many non-standard repair specialists (listed in the first post) and confirmed that the risk level is low.

A small, rounded slot chisel in a pneumatic peening gun directed at the edge of the aluminum flange will cause very localized plastic flow and create a compression ridge next to the thick steel flange, which will stop the leak.

O-ring replacement will be performed during the next planned outage.

Engineering has been asked to quantify the increase in bolt tension - that is the question being asked of Eng-Tips members.

 

[MikeHalloran]

A little extra separating force should not be felt by the bolts, if they're already properly preloaded.

I'd consider checking that the bolts are preloaded at least a little, e.g. by applying a torque wrench to verify that below some calculated torque they don't move, but beyond that I wouldn't mess with them. ... for fear of making the leak worse.

We here still don't know how bad the leak actually is, because you haven't quantified it, and you haven't told us if it was present immediately at startup, or if it appeared after some hours of satisfactory service, or after some plant upset. All of that might make a difference in guesstimating what's actually leaking and why.





Mike Halloran
Pembroke Pines, FL, USA

 

[DLite30]

Sounds like a BP operation to me.

Peening won't work on that connection, especially with cast aluminum. Good luck with that.

The only thing I would trying as a last resort would be envoloping the entire joint in a tomb of epoxy, and vacuum bag it until it cures.

 

[Bambie]

Yamoffathoo has updated me on the current situation;
- Bolt replacement was successful on one side of the flange, however, loosening either the top or bottom bolt on the other side caused the leak to increase, so they re-tightened and backed-out.
- Inspection of the pipe supports revealed that a nearby u-bolt guide was tight against the side that leaked and was prying open the flange.
- Maintenance is considering loosening the u-bolt to release the bending moment
A picture of the support arrangement is attached.