Induction Bends in 42" ASME B31.8 Pipeline

[GL431]

Which standards actually apply for the dimensional tolerances of the OD at the ends of induction bends to be used in an ASME B 31.8 42" pipeline?

API 5L (43rd ed.) table 8 gives tolerances for the OD of pipe-ends, which are quite tight. I can see that these tolerances ar fine for straight pipe, but I would have thought that more leeway for induction bends would be acceptable because of the manufacturing process of these bends.

ASME B16.9 (wrought butt welding fittings) give more comfortable tolerances (table 2). But induction bends are not wrought, are they?

The bends in question were first manufactured as longitudinally welded pipe according to API 5L. Then they were bent by the induction process. Due to the bending process, the wall thickness will decrease on the extrados side and increase on the intrados side. As per industry standard, the initial pipe had a slightly heavier wall to allow for the thinning of the extrados wall upon bending. I would naturally have thought that the tight API 5L tolerance would not apply anymore for the finished bend.

Follow on questions are: why do wrought fittings get a laxer bevel end OD tolerance than other fittings and what effect does the difference in the two tolerances have on the field welding process?

Thank you in advance for your input.

 

[BigInch]

You must hold API minimums on induction bends. Due to the variation in yield stress permitted in the yield stress rating, since only some pipes actually get a tensile strength test, the yield stress of any given pipe is not totally guraranteed and its possible that a given pipe has a small percentage chance of having a SMYS slightly less than its yield rating. The min wall thickness (supposedly) makes up for the possible lesser yield stress.

I've never really thought past the obvious about the rest, perhaps its the same logic.

http://virtualpipeline.spaces.msn.com

 

[GL431]

BigInch, thanks for your reply. In my present case the actual wall thickness exceeds the minimum required by API 5L by far. The problem is in the outside diameter at the bevel. The actual is larger than what the + tolerance of API 5L would have allowed (by 4 mm on 42"). But it is within the range which ASME 16.9 would have allowed. I think it is rather a problem of weldability out in the field than a problem of yield stress. There is in fact more metal than API 5L wants, not less. So, the question is about the OD, not the wall thickness at the bevel.

 

[BigInch]

That's what I meant about "the obvious" possible mismatches in weld-up and undercutting or overlapping. Is there more than 3/16" misalignment in the wall faces? If so, not good.

http://virtualpipeline.spaces.msn.com

 

[GL431]

I think I understand now. Thanks.

 

[BigInch]

Sorry I can't be more help. Let us know if you find out something about the different criteria between those specs. It never was a problem for me, so I never had the "opportunity to investigate".

http://virtualpipeline.spaces.msn.com

 

[SJones]

Try ISO 15590-1 and see how that may have helped a bend manufacturing specification.

Steve Jones
Materials & Corrosion Engineer

http://www.pdo.co.om/pdo/

 

[GL431]

Thanks to both of you. I had the ISO 15590-1 sitting there on my computer, but it did not enter my mind to look for it. That is a problem if one is chasing standards in different systems. The ISO states a tolerance in the ballpark of the API 5L spec for line pipe (i.e. far stricter than the tolerance for wrought fittings). So I will concede that induction bends have to have the same tolerance as the line pipe. ISO states that tolerance should be applied either to OD or to ID, depending on client spec (i.e. leaving one degree of freedom). I still have to ascertain whether API 5L states something similar.

Which leads me to a real problem. The factory has delivered bends which are therefore out of tolerance (based on measurements out in the field), but equivalent to the other wrought fittings (e.g. barred T's). I can now advise to continue to weld up and to perform extra welding quality control, or to pare the bends down at the bevel (would be about 2 mm and a process which I have not investigated yet), or to say treat them like the barred T's which have a wider tolerance.

 

[BigInch]

As long as there is minimum metal left over for allowable stress, I don't see why you couldn't do a tapper bore on the inside or the outside, or both.

http://virtualpipeline.spaces.msn.com

 

[GL431]

Yes, that seems to be the most appropriate fix. 170 bends of which approximately 50% have too wide OD's (by 2mm). The construction contractor is probably going to make a casino out of this. Let us see whether he has the equipment to do it. Is this a normal event ("taper boring" or "tapper boring") during pipeline construction (42") and can one reasonably expect the construction contractor to have such equipment on site?

 

[BigInch]

It is done a lot at transition pieces between class wall thicknesses; usually not so many pieces are involved.

With so many pieces, it might be better to try to find a shop and do them there rather than in the field. Shop rates are cheaper than field rates and QC is much better.

http://virtualpipeline.spaces.msn.com