lack of penetration in a 30 year old vessel 4

[Bartholomew69]

I have a 30 year old vessel that was recently modified (one section was cut out and replaced)- it has operated without problems up to now. Once cut it was discovered that the longitudinal seam had lack of penetration due to the weld preperation (double vee with central 'shoulder'). It was in the central shoulder region that there no fusion. There is no crack propogation into the weld material

The vessel was designed to ASME VIII Div 1. The current code does not allow lack of penetration UW-9(b); UW-35(a); UW-37(a); UW-51(b)(1). Now that it has been discovered, must this weld seam now be repaired or since it has operated for so long without any propogation of cracks, can it be left as is

 

[metengr]

Yes, I believe it should be repaired regardless of past experience. I am assuming that since this is a vessel that is in service, the National Board Inspection Code (NBIC) would be mandated by your Jurisdiction or Insurance Agency that insures the vessel. The NBIC is an in-service inspection, repair and alteration code that would require weld repairs to be performed in accordance with the original code of construction - in this case ASME Section VIII, Div 1. As you pointed out in UW-9 b) and other applicable parts of UW you need to demonstrate complete fusion and complete joint penetration.

The only other option you would have is to discuss your findings with the Jurisdiction or insurance agency to determine if they would waive repairs to the existing seam weld(s).

 

[jte]

Bartholomew69-

I feel your pain! You are probably sitting there having to 'splain to some manager (or worse - an MBA)why this vessel which has served reliably for 30 years needs additional repairs based on the discovery. This is now a large discovery job in the middle of a planned turnaround which may be critical path. Just keep in mind that just because it didn't fail doesn't mean that it is safe.

I concur with metengr that you need to discuss this with your AI, Jurisdiction, and insurance agency. Depending on the details of the problem (is it 90% penetration or 50%, no nozzle penetrations or other non-internal pressure loads on or near the seam, etc) you may be able to take a FFS approach using API-579 with this. You might have to derate the vessel to make it work, though.

One approach which might pacify your management is to figure out what your corrosion rate is and how much future corrosion allowance you need to get to the next opportunity to take this vessel out of service. Add some margin to the FCA to be safe (maybe 50%). Take UT's of the shell along the seam to be confident you know the existing thickness. Determine what your actual tmin is and see where you fall with regard to the FCA (see API-579 4.5) after reducing the available thickness by the lack of fusion percentage. This may put you into a situation where you can FFS the repair until the next scheduled outage so that it can be planned and prepared for (welders and weld rod on site, scaffolds erected, PWHT crew and equipment ready, etc.) It may even be possible to justify just replacing the vessel at the next outage if you can buy a new one with a long lead time.

A paper was published at the ASME Pressure Vessels and Piping Conference in 1994 by David Osage et. al. titled "Use of Nonlinear Analysis Techniques in Fitness-For-Service Assessments in the Refining Industry." This paper includes a case study where (during a plant outage) a two piece head with a seam which was "ostensibly joined by a double pass [groove] weld and subjected to a 100% RT... However, the inner pass was never welded, leaving a groove approximately 3/4 in. deep and 1 1/2 in. wide..." This is on a head with a tmin of 1.6 inches. Long story short, a quality FFS analysis was performed, the vessel placed back into service at the end of the outage, and a planned repair was done at the next outage. If you provide me your contact info (phone or email), I can get you a copy of this paper, since I couldn't find PVP-Vol 277 at the ASME site.

jt

 

[Bartholomew69]

jt - thank you for this valuable information.

The lack of fusion is approximately 1,6mm on a 12,7mm wall thickness, but it is continuous along the whole seam. I will do min thickness calcs and use the FFS method to try buy some time as you suggested.

i would appriciate if you could send me that article. My e-mail address is bart.bartholomew@telkomsa.net

regards

Bart

 

[JStephen]

It seems to me that something is missing from the above discussion.

If you build a structure with spot radiography and a reduced joint efficieny as typically allowed in tank and vessel codes, you accept that areas of weld that are not radiographed are not necessarily free from flaws. You are trying to establish a reasonably high quality of welding, not to confirm perfect welds at all locations.

Suppose this was a newly built structure with spot radiography. And for some reason, you chose to make an additional shot in excess of that required by code. Is that additional shot also required to be perfect? If it is, then you are raising the standard of welding that the structure is built to; essentially, you're saying that welds made with spot radiography should be just as flaw-free as welds that are 100% radiographed. But if that's the case, you'd use a higher joint efficiency, too, right? (And you'd also pay for that higher quality.)

It seems to me that if that vessel was spot radiographed when it was built, visually inspected, etc., then there should be no reason to try to go back and start making repairs based on a higher level of inspection, which is essentially what this amounts to.

 

[metengr]

JStephen;
As a National Board Certificate holder, when I make or direct weld repairs to our boilers and pressure vessels, I perform these repairs with the intention of being able to conduct NDT to the NBIC and/or code of construction requirements. As in this case, if I perform a weld repair, how can I attempt to obtain a clean shot with RT or even using UT to assure a quality weld repair with a defective butt weld?

The argument that it was built this way or use lesser acceptance standards is not relevant to our discussion. The question was posed that a repair was made and a problem was observed with the original vessel weld(s). As per the requirements of the NBIC, you need to perform some method of NDT to assure integrity of the weld repair. If this means removing 100% of a vessel weld as part of the repair SO BE IT. I am not advising to perform a 100% inspection of the vessel welds, this only pertains to the repair weld location.

I suspect that the vessel was constructed with minimal NDT, which for vessels built many years ago was common practice. In the spirit of the ASME B&PV code I would have a hard time believing that a vessel seam or girth weld, as a fusion weld, would be intentionally designed this way and given a higher efficiency factor. I suspect that someone either forgot to back gouge or assumed that the welding process would achieve 100% penetration.

 

[arto]

MIL STD 2035A has a section on "Inadvertant Radiography"
Sounds kinda like here - finding out bad news accidentally.
par.5.2.1.10.c:

"for welds which are not required to be radiographed, which are inadvertently radiographed, all defects shall be evaluated. Defects that are determined to be cracks shall cause the weld to be rejected."

A 1/16" "crack" running the length of the vessel = get out the grinding wheels!

 

[JStephen]

But lack of penetration is not a "crack", is it?

 

[arto]

It depends on toughness/temp.
Harvey Ch 5.25.1 has a section on "Leak-before-break"
stress = K/sqrt(pi*t)

for Kic=sqrt(5*Sy*(CVN-0.05Sy)) ==> 54ksi*in^.5 for 36ksi & 18ft# Charpy's
stress = 43ksi - it'll leak before break

if you're below Nil-ductility Temp:
for Kic = 0.45*Sy ==> 16.2ksi*in^.5
stress = 13ksi = fast fracture possible

[Kic=0.45Sy from L.R.Hurst "Brittle Fracture of a Brick Lined P.V." Mat'l Perf. March 86 pp24-26]

 

[metengr]

Lack of penetration can be considered a "notch" and is a stress riser. This type of joint would not be considered "good engineering design" in pressure vessel code space.

 

[JStephen]

Metengr, looking back through this, it seems you are saying that since they have cut into this weld, they are going to weld back across it, and then when that repair weld is radiographed, the original lack of penetration will show as a flaw. And then if you gouge a certain distance of that original weld and reweld it, you'll still find a flaw at the end of it where it runs into the original weld again. Is that how you're figuring it? That makes more sense to me. But the idea of radiographing the whole weld looking for problems doesn't.

We have run into similar issues, not on existing welds, but with customers who, for whatever reasons, elected to do more radiography than the structure was contractually required to have.

 

[metengr]

correct.