Mecway and ABAQUS Compared to ASME PTB-3

Sergio

Hi, I'm working on some under pressure parts, and in order to validate the results, took as reference this problem found in the PVEng site:

https://pveng.com/wp-content/uploads/2016/06/PVEfea-9128-1.0-Abacus.pdf

First results looks very close to the references values (probably the difference is due to different geometry in the drawing than in the model used in Abaqus report, see the plane surface between SCL#3 and 4).

@Victor , is possible to compute the "primary membrane stress + primary bending stress + second equivalent stress" (Pl+Pb+Q) with the stress linearization results???

Regards

Victor

I don't know a lot about the ASME code but are you asking for it to do stress classification to identify secondary stress Q then sum it with Pl and Pb? I think that might be difficult for general purpose software because it depends on quantities like distance from a discontinuity that are hard to determine automatically. Is it possible to obtain Pl+Pb+Q in a "dumb" way from the other stresses? Can you do it manually by solving again with different loads, such as maybe thermal stress in isolation from mechanical loads?

MartinB

Sergio
'Pl+Pb+Q' equivalent stress calculation - is that not what has been calculated in your linearised stress results? For instance at SCL4, Mecway value for max bending plus membrane = 32723, compared to 32698 listed against Pl+Pb+Q in the reference document.
As per ASME VIII Div 2, Q does include the effect of primary loads, its just a localised stress at a structural discontinuity.

disla

Good afternoon, Sergio,

It's been a long time since this post but I would kindly ask you if you remember and want to share it how did you manage to solve the axisymmetric problem and extract stress linearization at the same time.
It is only available for solid elements and any attempt to move the problem to solid 3D increases dramatically my node number.
Yours seems to have some thickness but I don't see any cyclic symmetry on the tree.

Thanks

Sergio

Hi @disla , as far as I remember I have used frictionless support for taking in count the axial simmetry

Sergio

These are the project files in case someone one to try

https://drive.google.com/file/d/1qlSoiOQNtQe6HhbdEVFaHbAQojy6SB7s/view?usp=sharing

Still I have to finish to understand why Pveng report shows Pl+Pb+Q mostly and not Pl+Pb, this makes hard to compare with Mecway results.

disla

Wow, thanks.
That’s much more than expected. I'm working on this example too and will tell you If I find an explanation for Pl+Pb+Q mostly and not Pl+Pb.

Sergio

There is a difference in this FEA model against the P-Veng, look in the plane zone bellow the pipe, that is present in the results but not in the drawing. I will fix as soon as I have a little of time, saddly I don't have the editable geometry, but is easier to fix even importing the step. Definitely I will redrawn in Alibre to have the whole example editable.

JohnM

@Sergio, thanks for yet another great validation of the solver!