I have a system which is split into two parts, one is above ground connecting to the other which is underground by means of an extruded tee. The system is analyzed to ASME B31.8 considering the underground part as "restrained".
When checking the stresses in the operating case, the code stress at the underground SIF node was much lower than the immediately adjacent straight pipe section. I was actually expecting the opposite as the SIF by definition should intensify the stresses at the respective node.
Attached is a snapshot showing the stress plot at node in query. Does any one know why this happened?

B31.8 provides two sets of stress calculations: 1) standard stress calcs that focus more on bending and 2) simplified, quick checks for pipe that is fully "locked in" by the soil that surrounds it.
My guess is that by calling this pipe "restrained" (I do not believe it is) you are falling under those simple calcs that do not develop stresses similar to the above ground piping.

You have a large number of supports concentrated in the area. It looks like some of the supports are offloading the loads from the pipe in the center, perhaps.

If you add some extra breaks and re-bury the pipe, does the problem go away?

Are your node numbers prior to burial of the pipe clean? I.E. minimum node spacing of 10 in the region, but also not less than say, 1-2 node numbers available per linear foot?

Dave,
In B31.8 I can see that there are 2 sets of equations, one for "restrained" pipes and one for "unrestrained", the load case assessment is also different, and this is reflected in CII calculations. Are these the same 2 sets that you mean?

Moreover, the SIF (tee) node as you can see in the attached snapshot still falls in the "restrained" part of the model, and I assume that it should follow the same set of calculation equations as the adjacent pipe.

I have also tried something else, I have identified a SIF in the mid of one of the adjacent pipes (without modelling any branches), and the same thing repeats, the program calculates and lists the SIFs in the stresses report, but the code stress falls significantly at the SIF location (see attached picture).
I still do not understand why this happens, and whether it is correct or not. Your interpretation is much appreciated.

Thanks

The supports are automatically created by the software depending on the identified meshing zone, and I don't think this will be a node naming/spacing problem, the stress calculations should not be affected by the node numbers/spaces.

Yes, CAESAR II has two pipe conditions which will set the stress checks - restrained and not fully restrained.

For restrained pipe, CAESAR II will evaluate stress in categories (OPE), (SUS) & (OCC). For pipe that is not fully restrained, CAESAR II will check stress fro (SUS), (OCC) & (EXP).

Additionally, B31.8 has two stress checks for restrained pipe - SL must be less than 0.9ST and the combination of longitudinal stress & hoop stress (long duration) must also be less than 0.9ST. But this combined check only applies to plain pipe - not tees or bends.

There are several instances where a Code offers more than one stress check for a load case. CAESAR II only reports the stress that is closest to it's limit (%). Here, with the same limit applied to both stress checks, it is not obvious which check is being reported.

Needless to say, your "simple" stress check is now quite complicated.

I suggest you submit your question and model to SmartSupport for a more detailed review and comment.

Dave,

Thanks for your fruitful response.
I think this interpretation is persuasive, especially for my case where I am pretty sure that the biaxial stresses are higher than the longitudinal ones. This pipeline actually follows a strain based design in its buried sections (apart from the tees), and as a result the used wall thickness will be too low for our stress checks, resulting in high hoop stresses.

I can see now that CAESAR II is fully compliant with the code, but in principle, don't you think that biaxial stresses should be assessed for pipe fittings as well, and not only limited to straight pipe sections?

Thanks

I can only guess...
That biaxial stress calculation is applied ONLY to straight pipe, not because we are most concerned about straight pipe, but because there is not a similar, simple stress formulation for pipe that is not straight.