Pressure Vessel Fatigue

[SG12345]

My company has supplied several vessels that are now approaching their -3 sigma fatigue life as defined by ASME Section VIII, Div.2 which our client has been made aware of. The vessels are made from SA 516 Grade 70 and swing between 0 psig and 170 psig every 2.5 minutes. The operating temperature is about 150F.

Some of these vessels have issues with weld peaking and weld misalignment. As such, an API 579 FFS analysis was done to de-rate the fatige life of the vessels.

My questions are as follows:

1) Once a vessel reaches its -3 sigma life, what procedures could be used to determine whether or not a vessel is still fit for service.

2) Does fatigue usually result in crack intitiation?

3) If cracking is the result of fatigue, can Part 9 of API 579 be used to determines suitability of the vessels for existing service conditions?

 

[TGS4]

You've moved beyond fatigue into a fracture-mechanics approach. First, determine your minimum-detectable crack size, usign the NDE method of your choosing. Then, determine, using Part 9, the number of cycles it will take for that crack to propagate to 80% of the wall thickness. That number of cycles sets the maximum inspection interval. I usually halve that and that becomes my inspection interval.

Every time you inspect it, it resets the clock, because there are no flaws of a detectable size.

The fatigue approach to life has been useful to get you here. Now, walk away from that, and use the fracture mechanics approach. The vessel could operate many times its current life with this approach. And, when/if you find flaws, fix them, or declare end-of-life.

 

[Duwe6]

Internal Inspx by an API-510 Inspector who has been made aware of the Fatigue Life problem. Inspector will have magnetic particle testing MT performed to look for crack initiation.

No cracks = Fit-for-Service. Simple.

The problem will be trying to establish a good Intertnal Inspx interval. Do not randomly pick a calender interval -- make the vessels tell you. Meaning, with all of them in "Similar Service", get into the one(s) with the worst weld peaking & misalignment [hi-low] ASAP. Keep doing 1 or 3 a year until the first crack is found. Set up your intervals based on 1/4 or 1/2 the time it took for those first cracks to initiate.

Being fatigued, the 1/4-time to crack interval will keep getting shorter, thus the out-of-service plus Inspx costs will escalate. But early fatigue cracks are very shallow and fairly short. They will 'buff' out easily, with no need for weld repair.

The biggest thing is to make certain that management understands that Benign Neglect will lead to "BOOM".

 

[SG12345]

Thank you both for your replies. The response was much like what I had expected.