Modelling miter welds along a pipeline

I'm currently performing a study to check the stresses in a fabricated expansion loop which has a 1.5° - 3° miter weld between two pipe joints.

I'll like to know how it should be modelled in AutoPIPE. Can it be, for example, a 1.5° deviation in the second pipe joint from the original alignment of the first pipe joint which causes a kink, or is there any other approach i'll have to consider?

Any advise is greatly appreciated.

  • Adding on, I have a more definite question. For a 1 cut miter joint, should it be considered as a widely spaced miter as there is no sections and there is no comparison of S and Sb (according to the AutoPIPE manual)?

  • Hello Mwan,

    I apologize for the inconvenience, however please see the following AutoPIPE help section:

               Help > Contents> Search Tab> enter "miter bend" (include the quotes),

               press List Topics button, double click on the selected topic "miter bend" from the list

               provided to see more information.

    On this page is a Note:

              Note: The user is responsible for predetermining whether the miter bend is closely or widely spaced.

              The current version of AutoPIPE does not trap an incorrectly specified miter bend.

    Therefore, you must decide the type of miter bend, as the hoop stress calculations vary from code to code based on this decision.

    Again, I apologize for the inconvenience.


    Mike Dattilio

    Bentley TSG Support Analyst


  • For miter bends there are closely spaced and widely spaced. Please advise how to determine S (chord length) for a 1 cut miter.

  • For two common codes, consider the following:

    B31.1 104.3.3 Miters

    A widely spaced miter with theta < 9*SQRT(tn/r)deg shall be considered to be equivalent to a girth butt-welded joint, and the rules of this paragraph do not apply.

    B31.3 304.2.3 Miter Bends.

    An angular offset of 3 deg or less (angle in Fig. 304.2.3) does not require design consideration as a miter bend.

    This means that for B31.1 you need to calculate the value for your particular wall thickness and radius.

    For B31.3 it is simply any change in angle up to 3 degrees (with a single cut).

    For these geometries, they can simply be modelled as kinks in the pipe.

    Answer Verified By: MWan 

  • Thank you for your answer.

    While we understood the codes, we did not know if it can be modeled simply as kinks in AutoPIPE or having to model it as a widely spaced miter bend to be correct. We had performed the analyses based on these 2 methods and found that the stresses were equivalent, and your answer gave us the affirmation that kinks can be used as well.

Reply Children
No Data