Required / suggested NDE for in-service compressor suction separator and stage 1 suction bottle 4

[SNORGY]

Client has an existing compressor proposed to be relocated from Site A to Site B (Alberta, Canada). To achieve suitability in the new service, the nameplate pressure needs to be increased from about 4520 kPag @ 149 C to 4960 kPag @ 55 C. An in-service inspection is planned prior to relocation, but the AI's documentation with respect to extent of NDE is disappointingly vague or silent, other than making references to API standards (e.g., 510 - which I have ordered...).

In absence of better data, my plan is as follows:

(1) Thorough in-service visual to verify general condition, confirm nameplate against Manufacturer's Data Reports, etc.
(2) Shut down for internal visual inspection using boroscope as required to look for signs of corrosion / pitting.
(3) UT traverse for 360 degrees of the circumference along entire length of vessels / bottles, 4" (100 mm) spacing for thickness examination.
(4) UT on Category D nozzles.
(5) MPI (external) on circumferential weld seams, with any indications measured against ORIGINAL Edition of Code Of Construction (I don't want to be telling the folks in the field to start chasing out crack indications that, had they been found during original fabrication, would have been accepted).
(6) Re-run vessel design calculations based on measured thicknesses to see if re-rate is acceptable according to CURRENT Edition of Code Of Construction.
(7) Submit application with records and calculations to Jurisdiction for approval.

Any thoughts or suggestions regarding the above? I am kind of flying by the seat of my pants here but, like everything else I am dealing with, it's a schedule crisis...

 

[SNORGY]

Regards Point 6 - calculations in accordance with CURRENT Edition of Code with allowable stress values from ORIGINAL Edition of Code.

 

[metengr]

Before I would do anything, submit the above plan to ensure you have captured all that the Jurisdiction requires. Having been down this road, you need to have your plan approved first versus doing all of the above and finding out you missed an additional inspection requirement mandated by the Jurisdiction.

 

[TGS4]

SNORGY - please contact me off-line. I've been down this road before with ABSA. I can help. I definitely echo what metengr said!

 

[SNORGY]

metengr,

Thanks...I was afraid of that very scenario. In both of the Provinces closest to me, it's sometimes a challenge to satisfy a Regulator.

TGS4,

I believe I can track you down tomorrow morning. If you don't mind, I would welcome the opportunity to talk about this one.

We have added a hydro to the plan...but that depends on what the NDE and ensuing calculations conclude.

 

[DekDee]

SNORGY,
Interested in your reasoning for evaluating using acceptance criteria from the ORIGINAL code edition ?
No idea what your code of construction was but if you had a crack that was acceptable in 1960 (hypothetically) and you are now re-rating the compressor to a higher pressure would that crack not be evaluated as per the latest code edition acceptance criteria ?
Interested in your / metengr / TGS4 thoughts,
Cheers,
DD

 

[SNORGY]

I have actually had two bad experiences with this.

About 15 years ago, I was involved (against my will) in the evaluation of an 8000 bhp compressor package being relocated from Broken Arrow (USA) into Alberta (and then on to Saskatchewan) as part of a fireflood project expansion. One of the things ABSA mandated at that time was to examine the pulsation bottles for cracks and repair them as required. The package was dismantled for a thorough inspection (cylinders, etc. by a compressor shop; bottles by a vessel shop). I got a call from the vessel shop foreman, asking me how far I wanted to start grinding and chasing cracks out of the bottles, because the more MPI they did, the more indications they found, and the more grinding they did to chase them out. In that instance, the shop called in an AI, who told them that the flaws they were grinding out were flaws that would have "passed inspection" according to the Code edition that had been in force at the time of construction, so it was the AI who told us all that they would accept inspection pass-fail criteria on that basis. Interestingly enough, they subsequently reassembled the machine and - despite my commissioning plan that told them to NOT PICKLE THROUGH THE MACHINE because the bottles were internally coated - they decided that they knew better than me and, since it was too difficult for them to isolate the piping from the bottles when they got it on site, they pickled through the machine anyway and deposited a rather annoying blob of internal coating goo into the cylinders. So, about another million dollars later, after they took everything apart and cleaned it all up and put it back together again, they were up and running.

The second bad experience was about a year earlier than that, when I was begrudgingly involved in the evaluation of a surplus sour separator (6'-0" D x 16'-0" S/S) and sour condensate stabilizer (30 trays, 3'-0" D x 80 ft) proposed to be relocated from Clioent Site A to Client Site B in Alberta. In the case of the condensate stabilizer, my "superiors" (my boss included among them), had based their fitness for purpose decision on the basis of opening up the manway on the bottom and seeing that there was still a chalk mark on the inside of the shell, so "GOOD ENOUGH, EH?". SNORGY, stupidly, said, "What about the rest of it, like the trays and in between the trays?". SNORGY's superiors got angry and sent SNORGY to the field to pay for his insolence by climbing up through the still erected stabilizer (after Operations had taken half of all the trays out of it) under mask and circling anything that looked like a defect with bright red chalk. It was a rather involved climb; the air hoses got pretty heavy after I got to about Tray 15, but I was harnessed and there was a safety watch top and bottom, deploying rope as I progressed through the tower. I circled a total of 48 unacceptable pitting sites and identified 3 trays for complete replacement. On that basis, SNORGY - again stupidly, apparently - asked if it would be prudent to ship the tower to a vessel shop, strip off the insulation, and do a visual and UT inspection externally. This resulted in 168 external repairs and a complete re-sandblast and refinish of the vessel before it was finally declared fit-for-purpose, at a total cost of $20,000.00 more than the purchase price of a new vessel. Everybody involved (except the vessel shop) was pretty mad at SNORGY. But it didn't end there; the relocated separator was accepted as "fit for purpose" even though the slug volumes in the new service would require us to increase the length from 16'-0" to 44'-0", so SNORGY, again as punishment for earlier having suggested that perhaps more inspection in the field might be in order for these vessels, was dispatched to visually examine the vessel under mask internally. SNORGY came out of the vessel and declared, stupidly: "The bottom quadrant of the vessel is internally coated, but there are places where it looks blistered.". As it turns out, "blistered" was an unfortunate choice of words for a vessel in sour service, and a $60,000.00 map scan UT on 4" grid spacing ensued in the presnce of the Client's PV expert and the AI. It was discovered that the plate on one side of the middle circular seam was full of laminar reflectors (indicators), the ends of which (as you tracked them through the shell thickness) were in a discouraging alignment. The plate on the other side of the circular seam was fairly clear, and the contrast was remarkable. The AI declared that, in accordance with the original edition of the Code of Construction, it is likely that the plate would have been accepted, whereas in accordance with the Code edition in force at the time, HIC plate would have been requested and the plate would not likely have passed the HIC test. However, the AI accepted fitness for purpose at the new location on the basis of the facts that the plate material met the requirements of the Code in force at the time of construction and that the new service was sweet. It was then that SNORGY shipped the vessel to a vessel shop, cut it in half at the circular seam and inserted a 28'-0" can in between the two halves. Again, interestingly, SNORGY had to prepare a data sheet for the vessel shop, so he drew it on 1:50 scale by hand with a pencil and a ruler onto an 11 x 17 piece of paper, whereupon the vessel shop photocopied it and sent it back on their title block as the Issued For Approval vendor print, so at least SNORGY saved his superiors some money in drafting.

In summary, what I am trying to avoid in this instance - in far, far less severe and risky service - is unnecessarily committing the Client to refurbishing fleet equipment to "as-new in accordance with the latest edition condition" if I don't need to. The AI might accept that, but in retrospect, it might be a fall-back position that one resorts to in the event that it technically falls short of the current Code edition.

 

[SNORGY]

To clarify...it is only the interpretation and acceptance criteria for the NDE that, in the event they fall short of current Code acceptance, I propose to be measured against the original Code requirements. The calculations and pressure testing, etc. would be to the CURRENT Code rules, based on the lower of allowable stresses between the new Code edition and the old Code edition. It's not a big jump in pressure to achieve the re-rate, and the requirement for the re-rate is driven by process constraints external to the compressor itself. Namely, the alternative is to put the machine into service as is without changes (after the inspection effort) and resize the suction PSV for the incremental gas breakthrough (fail open) rate across a 6"-ET control valve at the difference between 4960 kPag and 4520 kPag, which for this gas puts us into a 4x6 or 6x8 PSV with anything from an "M" to a "P" orifice, which the plant flare system cannot handle.

 

[Duwe6]

. . uh, guys:
Sect VIII Div-1 UW-51 (1) & UW-52 (1) radiography "any indication characterized as a crack or zone of incomplete fusion or penetration."

Appendix 6 Magnetic Particle Examination 6-4 (a) . . .free of relevant linear indications*

Appendix 12 Ultrasonic Examination 12-3 (a) "Indications characterized as cracks, lack of fusion, or incomplete penetration are unacceptable regardless of length."

In other words -- no $%#^ cracks!. Wonder what criteria that AI was working to ??


*cracks are both a 'linear indication' and over 1/16" 1.5mm long are 'relevant'

 

[SNORGY]

Alas, I can only recount my vivid recollection of a past assignment that I never wanted to inherit (buying used equipment of unknown quality because the Client said it was a good idea).

Trust me, guys, I don't make this stuff up.

 

[SNORGY]

In the case history described, the "cracks" were those indications on the external surfaces on and / or near nozzle welds that were revealed through WFMPI, and the ensuing surface grinding to (using the shop's terminology) "chase them out" were not successful.

 

[TGS4]

Duwe6 - you are correct. The catch with this is the technique used to look for the cracks. Every technique has a threshold for detection - and the important thing to differentiate between a fabrication crack and an in-service crack is to ensure that the inspection techniques match.

Sure, you can easily say

No $%#^ cracks!

but then I will turn around and ask you: what is a crack? If I use a high-powered SEM, I am sure that I could find linear indications that resemble "cracks", but is the size truly significant? Every flaw characterization must have a size criteria, and that size criteria is often a function of the expected service: cyclic vs non-cyclic, service environment, operating stress state, temperature, etc.

This actually comes up a lot in evaluating piping welds. A certain weld joint efficiency can be assumed based on a reduced inspection extent. Now, what do you do if/when you find a "rejectable defect" 30 years later when you accidentally catch a bit of that old weld (which was never inspected) on a tie-in weld x-ray?

 

[Duwe6]

You may be overthinking the 'crack' thing. Fatigue cracks are shallow at first. As are Environmental Cracking stuff; IGSCC, Caustic SCC, Chloride SCC. If it is an 'acceptably small crack', it will easily buff out with a sanding disk. I have personally removed a few DeAereator Tank stress cracks, and some caustic ones. You don't even have to remove 1/32" 1mm of metal to remove the 'acceptable' ones, and the buffing to remove them gives a nice shallow, smooth taper back to undisturbed base metal.

Try buffing out the cracks you find - most will be very shallow [today]. And then try to figure out when the last inspection for cracks was, and shorten that interval by half. Catch 'em while they are tiny and buff them out. Then you don't have big, ugly ones.

 

[Duwe6]

As for RT/UT evaluation of old, existing welds that;
". . accidentally catch a bit of that old weld (which was never inspected) on a tie-in weld x-ray?"

Now it is time for some of that hard-to-find "Good Engineering Judgement". If the joint efficiency is 85%, will that old nasty weld give you 85% or better strength?? Usually the answer is 'yes'. But there will always be that 'nervous Nellie' that just cannot allow an identified flaw to exist. Those undiscovered flaws are the reason for the 85% E. If you just have to have 'excellent welds, everywhere', you have to be willing to pay for 100% RT/UT. Show Nervous Nellie the anticipated bill for 100% RT/UT, bet he will not approve it.

The whole concept of reduced joint efficiency based on RT of 5% of the welds assumes that there will be small flaws left in the 95% of welds that did not get tested. Doesn't matter if you find them the day the welds were made, or 30-years later. Use some Good Engineering Judgement, and you will find that most, if not all, of those existing flaws will still allow 'Continued Good Service'.

 

[DekDee]

Sorry to deviate SNORGY, don't mean to hijack the post.
We radiographed two welds on a replacement spool during a critical shutdown in a meat processing plant in New Zealand, an Ammonia line.
Made the mistake of overlapping the film on to an adjacent circumferential butt weld that had been in service for over 20 years.
The existing weld had 100% LOP so even though we hadn't meant to radiograph it we had to report what we saw on the film.
Extended shutdown to fix something that had been in service for 20 years - we were not popular with the Owner,
Cheers,
DD

 

[SNORGY]

No worries about the hijacking, all of this is valuable to me.

I thought I had posted an update but maybe it was red-flagged and deleted. In any event, in my particular case, we will be ordering replacement vessels due to other mechanical limitations that would have rendered the desired re-rate impractical, so the NDE became irrelevant.

 

[marty007]

The cost of inspection/re-use of old equipment often surprises management types (especially when relocating across borders). They think it should be easy to relocate from site A to site B and plug and play.

I've seen it too many times where building a replacement would actually be cheaper, so I'm often very relieved when I can find a cut and dry technical reason that disallows re-use, such as pressure rating.

 

[Duwe6]

It's a little late now, but minor Incomplete Pennetration [even 360°] is structurally acceptable. The circumferential joints just have to be 51% as strong as the pipe seam to have the seam rupture first. Next time, see what it would cost to either have Phased Array done, or the weld depth-of-penn evaluated by an expert* using manual UT. If the depth-of-penn stayed above 50% all the way around, I would've recommended that that weld be left alone. And signed the recommendation.

*Anybody with API QUTE certification is expert, as well as EPRI cert'd folks. YMMV on the folks from the local x-ray shop, but some are experts.