Acceptance limits for corrosion on threads during overhaul 1

[morganwz]

Does anyone know of any publicly available standards, recommended practices, documents, studies, etc. that address the inspection and disposition of corroded threads on in use equipment? We are in the process of writing our own criteria and are looking to see what information is out there that we could build upon.

To give some background, we are an OEM manufacturer of subsea equipment and our products utilize a substantial number of fasteners, nuts, and load bearing retention collars with stub acme threads ranging in diameter from 3.0" to 6.5". Materials are 4130 75Kys, 4340 130Kys, and 8630 130Kys. Our customers have equipment that has been in the field for 15+ years, and when it comes back to us for 5 year inspections and NDE examination, oftentimes a significant number of these connections will have some amount of corrosion damage to the threads. This is not necessarily an automatic rejection because the design has excess capacity far above that which is required by code.

The problem is our current method of dispositioning the threads on these parts is incredibly subjective and relies on the judgement and experience of the person doing the inspection. Basically to simplify things, if it looks like garbage or at least a certain percentage of the threads are not fully intact and undamaged, we advise them to replace the component. This has received substantial pushback from customers who want a less subjective acceptance criteria that is measurable/quantifiable.

They are wanting to use G0/No-Go gauging to make this determination or to measure thread backlash. I have been pushing back pretty hard on this as theoretically I could take a bolt, cut out half the threads and still get it to pass a G0/No-Go gauge test. As long as some portion of the thread remains undamaged they will always measure good.

 

[EdStainless]

And what happens when these fail?
If it isn't fatal then come up with a procedure to clean them (chemical, rinse, wire brush).
And then inspect.
I would discard anything with pitting more than a few thousandths, or threads that are undersized using a gage.
I used to build downhole equipment and our fasteners cost a lot more than yours (custom formed Monel K500).
We never reused fasteners.
$20k worth of bolts was chump change compared to a failure.
The field guys kept the old ones and used then for testing and non-downhole applications.

 

[Stress_Eng]

As a suggestion, you can put a procedure in place where the MRO supplies you with a report identifying the region of the threads that have been blended to remove corrosion / impact damage etc, such as a circumferential length of each damaged thread. I'm sure this will also include an appropriate NDT report. Where blending has been carried out, and including a % to encompass run-out etc (for example, 33% on each end of the blended thread length), a summation of the % of thread that can be assumed non-effective can be obtained (any blending will effectively render that portion of the thread ineffective as contact has been lost). The original thread analysis (thread shear, bending, jumping, SSC etc) can then be used to determine if sufficient thread remains. I would suggest some form of analytical approach, utilizing the original certification MS / RF calculations, to identify if the threads are structurally acceptable or not. Hope this suggestion helps.

 

[r6155]

Be practical. Use new fasteners and avoid wasting time on inspections, reports, etc

 

[morganwz]

Stress,

Definitely in agreement with you here. I actually wrote several versions of such a procedure to do exactly what you suggested. Part of the problem is we often see a lot of what I am going to term minor pitting that is pretty well dispersed throughout the thread length but the threads are in almost perfect shape otherwise. On paper these widespread but mildly pitted areas of threads should be considered non-effective which results in rejecting the component. The reality is that this is probably too conservative. A localized section of thread with mild pitting but otherwise good thread profile still contributes a substantial albeit reduced amount to the strength of the overall thread. Thankfully, fatigue is not a major concern here. On the other hand, we occasionally have components that will have several consecutive threads almost completely worn away but the component is accepted because the rest of the threads are in pristine condition and sufficient to carry the load. This results in the customer complaining about us rejecting what appears to be perfectly good threads but accepting ones they expect us to reject.

Basically, what would be useful is any resources or papers on the extent to which levels of corrosion damage affects the strength of a given thread for the purposes of coming up with derating factors.

r6155,

I wish I could tell them to just replace everything because the inspections and reports is going to be a nightmare. Unfortunately, these are not single use fasteners but massive bolts meant to be assembled and disassembled every couple months or so. Others are threads integral to major components and replacement requires cutting them off and welding on new ones. Replacement probably cost roughly $5,000-$10,000 a piece and we have several thousand units out out in the field with customers.

 

[Flying Whale]

Could you elaborate more on why a go/no-go gauge test isn't valid? If corroded threads are undersized and will pass a go gauge, get your next smaller go gauge of similar pitch and use that as a no-go.

 

[Stress_Eng]

From a concessions / repairs background, we would receive reports identifying the locations along threaded joints where blending had been carried out to remove corrosion / damage, etc. These blending locations could be anywhere along the threads. I could see the potential for using a go/no-go gauge if the damage was guaranteed to always be at the free end of the threads, but could a gauge be used if the damage was in the middle of the thread engagement, or at the far end (undercut and shoulder / step change in diameter)? Going far-far back into the distant past (A/C fitting apprenticeship days), I remember something about a 3 wire check?

 

[dgeesaman]

I agree that a no-go gauge won't necessarily catch all situations because the metal won't corrode evenly.

If the bolts are not very heavily stressed at their operating torque setting, perhaps you can apply a tensile test that is between full proof load and operating load? Compare the expected elongation with actual and use that as a pass/fail criterion.

Also is crevice corrosion and cracking not also a concern?

 

[dgeesaman]

Be practical. Use new fasteners and avoid wasting time on inspections, reports, etc

Not all bolts are just an off-the-shelf bolt.

But you could put it to the customer as a rebuild option: base offering is visual inspection and replacement of bolts based on best guess, adder to replace all bolting.

 

[Stress_Eng]

I recall some Structural Repair Manuals (SRM's) stating a size of blend (rectangular surface area and depth) and a minimum defined distance between blends. Any blending that falls within this size and spacing criteria is considered structurally acceptable, and only to report any blending that falls outside this criteria. It sounds like you want to develop such a criteria (size / depth of pitting to frequency of pitting)? You'll probably have experience of different types of damages, each of which may be able to have it's own set of criteria. But the difficulty, as you say, is how to develop such criteria. You could come up with a test plan?

 

[morganwz]

The NO-GO gauge is not effective here because as some noted, they don't work with non-uniform material loss. In our case, the corrosion along the length of the thread usually occurs bad/good/bad/good/bad where the bad threads on the ends I don't really care about as we have always assumed they were non-effective. You can't get the NO-GO gauge past the good threads to check the potentially bad threads in the center. Also, if excessive material loss only occurs over a 90deg section of the threads for example, a NO-GO gauge won't pick that up. A NO-GO gauge can definitely have a place in the inspection process, the problem is there are involved parties that want to rely solely on a NO-GO gauge test (and surface NDE of course) to determine acceptability.

Internally, we have considered using the three wire check Stress-Eng mentioned. The downside is it is labor intensive if you do it for everything. In the case of using wires to measure only localized areas, if an area is questionable enough to merit measuring with wires, we probably are going to reject it on the visual inspection anyways.

Stress Corrosion Cracking is definitely a concern here. We require phased array UT inspection but I doubt it would be able to detect cracks in the early slow growth stage of SCC. Whether our recommended inspection intervals are short enough to detect larger more developed cracks before they propagate to total failure is another question we are addressing. Fortunately I am not aware of any instances of cracks being detected in these components. This does not mean they are not there, just that we have not found them.

To be clear, any pitting corrosion of significant depth especially in the thread root area we are automatically rejecting.

 

[morganwz]

Stress,

That is exactly what I am trying to do. I was secretly hoping an organization like DNV/ASME/API would have done the research and developed industry guidelines like they have with ASME B31G or API 579.

 

[Flying Whale]

The problem is our current method of dispositioning the threads on these parts is incredibly subjective and relies on the judgement and experience of the person doing the inspection. Basically to simplify things, if it looks like garbage or at least a certain percentage of the threads are not fully intact and undamaged, we advise them to replace the component. This has received substantial pushback from customers who want a less subjective acceptance criteria that is measurable/quantifiable.

This may sound like a comical suggestion but honestly this is what I'd do in your position: take a large random sample of corroded bolts, present them to multiple inspectors and have them judge whether or not each bolt is reusable, and get an average score for each bolt. Then, create your own systematic approach by taking measurements, data that differentiate the bolts from one another (such as % of total lost volume of thread, smallest major diameter in the bolt, etc) and correlate these data with the scores given by the inspectors. You should be able to find numbers (doesn't have to be exact and can be a conservative estimate) that you could then use in the future to bypass the subjective judgement and refer strictly to bolt measurements for making the decision. Customers can be given these numbers as justification for bolt replacement, they don't need to know where those numbers came from, do they? Besides, one could argue this method takes away the subjectivity if multiple inspectors are consulted and a large enough sample is used, even if all those inspections are subjective in of themselves.

 

[dgeesaman]

To add to that good idea, take that sample of corroded bolts and test them to failure after the visual inspection.

You may find that some individuals had very good correlation with actual results and tap their opinions for how they rated them. You will more likely find that there is no correlation and visual inspection / judgment is worth very little and then you'll have data to justify replacement or justify not warranting re-used bolts.