Leaking Spiral Wound Gaskets 3

[uglyspec]

When performing an ASME code hydrotest on a heat exchanger with body flanges designed with spiral wound gaskets I have noticed that during the initial test all joints are leak free. But once the pressure is reduced and the test is repeated the following day those same leak free joints appear to now have a small leak.
Can anyone explain this occurrence? Is it simply bolt relaxation?


Thanks,

 

[racookpe1978]

That should never occur, and is definitely NOT "bolt relaxation".

You might - very, very low probabity! - have a bad gasket. Failures happen. But I've never seen a leak fail like due to bad gasket since my hydro test in 78.


How many joints are leaking?

You almost certainly have a bad assembly <- From bad assembly workmanship.
Wrong final torque values.
No intermediate torquing being set.
Bad torque wrench, or no torque wrench at all!, or out-of-cal/broken torque wrench. (or other torquing mechanism.)
Not tightening in a star pattern.
Using a sledge hammer.
Bad alignment.
Movement of the pipe or heat exchanger; bad foundation or bad pipe support twists the flange apart when the HX is drained or filled.
Bad original fitup -> So the flange never was parallel and at the right gap.
Bad fitup: Excess gap, no gap at all (the gasket is pushing the flange apart.
Scratches or cuts in the gasket face, or across the gasket.
Dirt or rust or gunk on the faces.

UNDER OBSERVATION! Take the flange apart.
Clean it, verify alignment, verify "no movement" happens.
Verify assembly and insertion position and handling and care care of the new gasket. (DO NOT re-use a gasket that has failed. Ever.)
Verify the torque values assigned for that size flange, those bolts, that brand name of gasket.
Verify the torque wrench. Prohibit "assistance" or extension handles or breaker bars or chainfalls or pry bars or other "we have always done it that a way" claims. If the foreman persists, fire him.
Verify original starting torque is correct. The 1/3 torque, 2/3, 90% and final torques are correct, are correct ON EACH BOLT in the pattern, and are properly using the torque wrench.

Workmanship - It ain't for weenies nor macho's.

 

[uglyspec]

To answer your question racookpe1978, typically only larger (>12" NPS) body flange and nozzles leak upon re-pressurization.

Thanks for the insight!

 

[weldstan]

How long did you hold the test pressure? If for a minimum time, the leak path may not have had time to pass through. Often times new bolts and gaskets are required to be put in place after hydrotest. See racookepe1978.

 

[uglyspec]

Hold time was 1 hour.

 

[ricklts]

Spiral wound gaskets compress under bolt loads and flange rotation. All bolts/studs will experience load loss due to thread embedment and relaxation. Once this occurs, especially if the sealing surface is closer to a 125 finish than a 500, a leak can happen during a second pressurization. Also, your assemblers may have been overly enthusiastic during make-up, and the flanges could have been yielded, cupping the seal surface and unseating the inner contact surface of the gasket.

Rick

 

[Guest102023]

Welcome to the weird and wonderful parallel universe of Leaking Gasket! Its right next door to that of Bolt Tightening Procedure.

"If you don't have time to do the job right the first time, when are you going to find time to repair it?"

 

[Duwe6]

At >12NPS size, the bolt torque was probably barely sufficient to hold the hydro pressure. Make sure that the bolt-up crew actually uses a torque wrench, actually follows a good pattern, and uses a 'good'* established torque value. If the problem persists, 'hot torque' the studs -- retorque them after pressurization, with pressure on the flange/exchanger.

*the Flexitallic catalog gives good, 'middle-of-the-road' torque values for making up flanges using spiral-wound gaskets.