FEA Translation_B31J

Posted by: RB

FEA Translation_B31J - 03/13/22 11:41 PM

Hello everyone,

Sorry for the long post.
I run a trial check for a "32 X 22 RFT" for B31J SIFs and Ks. Report from FEA Translation is attached. I have few points to be clarified. Any leads would be great helpful.


1.Branch SIF location
As per figure C-2-7 in B31J, if d/D > 0.5, SIF shall be applied at the center of the header in the Branch element (Node 20 on element N20-N23 in the attached CII). But, in B31J CII file generated from FEA Translation, SIF values are applied at outer surface of header in the branch element (N24 in element N24-N40). (Refer attached B31J CII file generated)

2. Pipe or Rigid
As per B31J, figure C-2-7 Element from center of header to outer surface of header shall be rigid, but B31J CII from FEA Translation consider it as Pipe element.
PS: Considering SIF at rigid element is not an usual approach. However B31J says it to be.

3. Two branch SIF values
In FEA Translation SIFs are applied at the center of header (N20) as well as outer surface of header at branch element (N24 in element N24-N40), While SIF at N24 is matching with the output report (Refer attached Report from FEA Translation). SIFs applied at N20 is not found anywhere in the Report. How SIF at N20 is calculated and is it necessary?

4. SIF values at Branch element
Inplane & Outplane SIF values are interchanged from the values given in B31J report. Is it because of we didn't give "branch connection type" in the SIF input ?
if No, please provide the methodology of SIF application

Thanks & regards,
RB
Posted by: Richard Ay

Re: FEA Translation_B31J - 03/17/22 02:05 PM

With regard to question #4, yes sometimes in the translation of a model by FEATools it appears the in-plane and out-of-plane SIF values are switched. This done purposely since the translation breaks the tees into three separate elements. When this happens c2 no longer knows the in-plane and out-of-plane directions for the fitting (since there is no longer a fitting). c2 will therefore default to the in-plane and out-of-plane directions for a straight pipe based on the pipe's orientation. This is all explained in the attached white-paper from 2015.
Posted by: vermaccio

Re: FEA Translation_B31J - 03/21/22 03:25 AM

Mr.Richard, where did you get this pdf document? Are there other similar pdf about different arguments?
Posted by: Richard Ay

Re: FEA Translation_B31J - 03/21/22 11:23 AM

Tony Paulin is the author of that White-Paper.
Posted by: Richard Ay

Re: FEA Translation_B31J - 03/24/22 01:46 PM

Here are the first three questions (sorry for the delay).

Q1: In C-2-7 the SIF is identified at the center of the header. More applicable data suggests that it should be applied at the branch/run-surface side too. Bending moments along the stub can vary in either direction. It should not make much difference except in very tight piping systems, and in this case, erring on the conservative side was considered reasonable.

Q2: The stress is defined using the SIF and the section modulus of the matching pipe and moment (more or less). When putting FEATools together we knew that we needed to compute the stress at the intersection point but needed a strong (rigid) element there too. Rigid elements in different pipe stress programs have different meanings. We could call it rigid (but would not get the stress), or we could use a short element which was rigid, adjust the SIF, and get the stress computed appropriately there.

Q3: See the answer for (2) above. The objective is to compute the appropriate stress as (SIF)(M/Z). For the rigid stub from the surface to the centerline an alternate Za is used, and so an alternate SIFa is needed so that SIFa / Za = SIF/Z.