ASME B31.3 335.1(c)(1), Flanged joint alignment

[timsch]

(c) Flanged Joints. Unless otherwise specified in the
engineering design, flanged joints shall be aligned as
described in subparas. (1) or (2), and (3).
(1) Before bolting, mating gasket contact surfaces
shall be aligned to each other within 1 mm in 200 mm
(1/16 in./ft), measured across any diameter.

If I'm interpreting (1) correctly, then the tolerance on the flange alignment for a flange with a 12" OD is +/- 1/16". Similarly, for a 6" OD flange, the tolerance is +/- 1/32", and this is regardless of distance between the flanges. Please let me know if this interpretation is incorrect.

If it is correct, how does someone deal with the challenge of the alignment tolerance of small diameter flanges with C-C distance of 50' for example or even greater between mating pairs of flanges?

 

[TGS4]

I'm completely confusing about what you are asking. In order for a mating flange to work, they need to be, essentially, in contact. The distance between the flanges should be the thickness of the gasket.

I don't understand your question with regards to flanges 50' apart. Do you have a 50' thick gasket?

 

[timsch]

Sorry for the confusion due to my poor wording. By 50', I mean linearly between two flanges on a single length of pipe that have to mate to two other flanges. If the flanges must be aligned to within 1/16" per foot (of diameter of a flange, if I'm reading the code correctly), then the 50' center-to-center dimension must have a tight enough tolerance itself that the flanges mate such that the alignment tolerances of each pair are within the alignment tolerance. Say, for example each flange has a 6" OD, then the alignment tolerance is +/- 1/32" for each flange. That would be hard to maintain with such a length of pipe.

 

[TGS4]

The tolerances are for the mating flanges ONLY.

Let's use the following exemplar...

FF-------------------------------------------------FF-------------------------------------FF
Flange Pair 1                                      Flange Pair 2                          Flange Pair 3
A/B                                                A/B                                    A/B

The tolerance applies to between 1A and 1B, not between 1B and 2B. It would apply to 2A and 2B.

Does that make sense?

 

[timsch]

Using your example, if vessel A has flanges 1A and 2A 50' apart nominally and are to have pipe B with flanges 1B & 2B connect, both 1A/1B & 2A/2B have to meet the alignment tolerance. Say the flanges are welded to their vessel & pipe respectively, then the 50' center-to-center for both the vessel and the pipe during manufacturing have to be very tightly controlled to be able to meet the alignment tolerances for each set of flanges since once welding is done no further adjustment is possible. It seems to me that controlling this 1/32" tolerance for 6" OD flanges with such a distance between them would be very difficult, which makes me wonder if I'm interpreting the code requirements for alignment correctly.

 

[KevinNZ]

A 50' spool will need site weld/s. Regardless of the flange tolerance, the two connection points are not going to be set out accurately in the field.

 

[LittleInch]

If you are fixing a pipe between a pair of flanges 50 feet apart where those flanges are very rigid and have no room for movement, then yes, a 6" spool piece pre made will be very unlikely to fit.

That's why you allow for some flexibility and a requirement to cut to fit within the iso, usually in three planes. thinking about how you're going to make it is all part of design and allowing some flexibility in a piping system required to avoid overstress on changes in pressure, temperature and fluid content

Ask / watch a pipe fitter at work. I think you need to get out and see some piping construction to understand how it works in practice.

It would be unusual to have such a long length of dead straight pipe between two rigid fixed flanges. Stress on temperature change would be large and even fitting it with a few degrees difference might eat up your tolerance limits.

The length of pipe between flanges is irrelevant. It's what you do with the flanges at the end which is key.

Remember - More details = better answers
Also: If you get a response it's polite to respond to it.

 

[TGS4]

Shop-welding such an example would require some tight tolerances and some high-quality measurements. My company has dimensional control (think high end surveyors) that have successfully measured field locations and checked the shop fab. It's do-able, but requires a high-precision mentality. This type of work is also done in the off-shore oil platform industry, where field welding is not permitted.

 

[timsch]

Thank you all for your replies.

In our shop we manufacture magnetic level gauges, frequently with a C-C for the two flanged process connections around 20' on a straight length of pipe through which a float with an internal magnet passes. We don't have the high precision mentality or measuring equipment needed to achieve the tolerances as specified in B31.3, although I do run the pressure calculations using that code. We deliver a complete product, so there is no pipe fitting on site for these. While we can keep a 1/8" tolerance on this C-C dimension, much tighter than that will be difficult for us to achieve using our current methods. It looks like it's time for me to look into new methods.

 

[KevinNZ]

If the floor of your workshop is flat you can make sure the connecting flanges are parallel to the floor and in the same plane.

 

[LittleInch]

That's a bit different.

If I was buying a probe to fit between flanges I would expect that either you offer it in set lengths to a pretty tight dimensional tolerance (+/- 1mm say) or that following construction someone would measure the length and provide you with an exact dimension or hopefully the site fit of the fixed length instrument has some site fit dimension.

I would though expect the instrument to be straight and with flanges at exactly 90 degrees to the tube in all planes.

Actually looking at your OP again. The alignment is basically an angle which applies to all flanges. Now in strict measurement terms yes, the allowed misalignment for a 6" flange is 1/32", but in angle terms it is the same as a 12" dimension with 1/16".

Length is irrelevant o how you fit up the flange to be 90 degrees to the axis of the pipe. A simple jig should fix this in place surely?

Remember - More details = better answers
Also: If you get a response it's polite to respond to it.

 

[timsch]

If it's a little bit different, I apologize for not presenting the issue better. Regarding the tolerance, I assumed that alignment of the flanges as described in 335.1(c)(1) was a linear dimension in the plane of the flange face, but is it rather an angular tolerance? Is it (c)(3) that would govern the linear (and also rotational) dimensions?

We do control angular alignment of the flanges during fabrication with precision levels, so that's controlled well enough. Set lengths are measured with tape measure, so I can't say that we hold the C-C dimensions to within +/- 1mm in all instances, but I will look into what we need to do if that is what the code specifies. It all comes down to whether C-C is governed by 335.1(c)(1) or 335.1(c)(3).

 

[DekDee]

timsch,
I will give my opinion based on a previous life (a long time ago) as a pipe fitter.
The reason for the tolerances noted in 335.1 (c) are basically to cater for 335.1 (a)."
335.1 (a)
"Piping Distortions. Any distortion of piping to bring it into alignment for joint assembly that introduces a detrimental strain in equipment or piping components is prohibited.

All flanged joints (weld necks or slip-ons) should be fabricated as close as possible to true square in both planes and this is not always easy to achieve.
As an example - a pipefitter cuts a pipe that is not true square, adds a weldneck flange and tacks the flange on so it is perfectly square to the pipe in all planes.
Only problem is the root gap is now larger on one side than the other due to incorrect initial cut so there is a strong possibility the flange will "pull" on one side once welded and the flange is now out of square.
That is why pre-welding inspections are required on critical flanged spools.

Now to put my B31.3 hat on.
we have a 50 ft long spool with flanges on either end that are hopefully welded square (or as close as possible to) to the pipe.
This spool is then bolted to an existing flange (in accordance with 335.1 (c) (2).
The mating flange at the other end of the spool must now be within the tolerances of 335.1 (c) (1)to avoid detrimental strain as noted in 335.1 (a).
If it was a 1 ft long spool piece the flange gasket faces cannot be misaligned more than 1/16".
A 50 ft long spool piece can have up to a bit over 3" misalignment.
By moving the pipe across or up/down 3" over 50 ft the committee has deemed unacceptable strain has not been induced in the joint.
This is applicable to all diameters.
335.1 (c) (3) is related solely to bolt hole orientation.
In the USA your bolt centres are on the vertical centreline so the centre of top hole cannot be more than 3 mm left or right of the vertical centreline.
In the Southern Hemisphere our bolt holes straddle the vertical centre line but it is the same premise - 3 mm max, one way or the other.
Hope that helps,
Regards,
Shane

 

[timsch]

Shane, with your explanation using spool pieces, I think I see now that this section of the code applies to piping where connections are made with flanges on the end of the pipe. I failed to note that for my application, the flanges are welded to the sides of the pipe, either at extrusions or very short nozzles, with the axes of the flanges perpendicular to the axis of the pipe.

 

[DekDee]

timsch,
I did not explain very well as I got a bit carried away with "mating gasket contact surfaces".
It seems to be easier to understand in the B31.3 2002 edition where the same formulas are used but it refers to the "design plane" instead.
(c) Flanged Joints. Before bolting up, flange faces shall be aligned to the design plane within 1 mm in 200 mm (1⁄16 in./ft) measured across any diameter;

I used the 2002 edition as there is a 1999 interpretation that I thought was applicable.

Question (I): In accordance with ASME B31.3-1996 Edition, para. 335.1.1 (c), does the specified tolerance apply to each individual flange, measured to the design plane, or to the total misalignment between two mating flanges?
Reply (I): It applies to each individual flange, measured to the design plane.
Question (2): In accordance with ASME B31.3-.1996 Edition, para. 335. 1.1 (c), do the requirements apply to all pipe sizes and pressure rating classes?
Reply (2): Yes.

This will still be applicable to your situation.
If your nozzle or extrusion is 400 mm long you are allowed max 2 mm of misalignment vertically and/or horizontally.
If your nozzle or extrusion is 200 mm long you are allowed max 1 mm of misalignment vertically and/or horizontally.
If your nozzle or extrusion is 100 mm long you are allowed max 0.5 mm of misalignment vertically and/or horizontally.

Cheers,
Shane

 

[timsch]

Thanks Shane. I appreciate the clarification.