B31.4/ Z662 restrictions on attachments

[nipra03]

Hi,
An issue I have is the restrictions that the pipeline codes impose on structural attachments and the need to have full encircelement sleeves. For example B31.4 states:

404.9.1
..... If the pipe is designd to operate above 20% SMYS all attachment welded to the pipe shall be made to a separate cylindrical member that completely encircles the pipe and this encircling member shall be welded to the pipe by continous circum. welds.....

Z662 Canadian code has similar wording but with 50% stress ratio.

I want to know why this requirement is specified.My concern is with regard to partial saddles that are welded to the pipelines at support locations. A coleague of mine mentioned that it has to do with the stress cocnetrations at the longitudinal welds that attach the saddle to the pipe ( i.e. welds parallel to axis). The code however does not state such.

Is there reason to be concerned for the longitudinal weld any more then a partial circum. weld ?

 

[BigInch]

Your bud is correct. Pipeline codes devote more of the total allowable stress to meeting hoop stress requirements, therefore there is less of the total available strength to accept stresses from concentrated stresses, and residual weld stresses, near directly welded attachments. If hoop stresses are less than 20% of SMYS they consider that sufficient excess strength exists to handle whatever additional stresses might develop from the direct load and welding, heat and shrinkage stress combinations.

The codes don't often explain many of their provisions to the full satisfaction of readers, but none the less, the provisions must still be followed. Actually even make it more imperative to do so.

I hate Windowz 8!!!!

 

[nipra03]

Thanks Big Inch.
The concern for local stress concentrations due to attachments is known.
The issue of concern : why is the longitudinal weld considered more severe than the circum weld say in a welded saddle.
I can email you a pdf of the saddle if you like. I dont know how to post one online.

 

[LittleInch]

You can add attachments via the "...or upload your file to engineering.com" link under the posting box. It's very easy.

BI has stated quite clearly that because hoop stress is normally the biggest stress in a pipeline (usually by a factor of 2 compared to axial stress), there is less available "spare" stress allowance compared to longitudinal. How many pipe failures have you seen that are not longitudinal failures - very few I would expect other than ones where it failed due to shear or axial stress. Therefore adding local stress concentrations is not good for the pipeline.

Are you talking about a planned attachment or one you already have?

If you need to fix your saddle then you normally just clamp it circumferentially or attach a collar like the code says and weld to that instead and everyone is happy.

The code bans these things because it is very difficult for the code writers to differentiate between what could possibly be called a tack weld and a major fillet weld. There are examples of longitudinal things welded to pipelines, normally with small fillet welds, but there is a process and procedure that needs to be gone through (including physical testing) to demonstrate that in your particular instance the code recommendation can be overruled. However for a pipe support there are many other simple solutions so normally no one bothers.

My motto: Learn something new every day

Also: There's usually a good reason why everyone does it that way