Dear Mr. Ay,

I have some remarks on your points.

1) For purposes of design, two "restraint conditions" of the pipelines can be recognized, restrained and unrestrained.
For example:
- "Unrestrained" end of a pipeline means that the pipe is free to displace axially;
- They are "Restrained" sections of buried pipelines, where the axial strain is zero, because enough friction (as axial resistance) has been developed from the "unrestrained" end, changing progressively the thermal growth and the pressure elongation.
There are Codes (ISO for example) and Books that give a formula of the "virtual anchor distance" that counts exactly the scenario above mentioned and the result of such formulas means that you expect that measuring "VA" meters from the unrestrained end, you expect to find the begging of the restrained section of the pipeline.
I would say that the Virtual anchor is simply the "border" of the "restrained" zone.

2) The virtual anchor, as a border of the restrained zone is a real point (and the sense of "real" is that it is a point you can walk up to it) but is "fictitious" in the sense there is no "physical anchor" there. Moreover, VA is dependent on the operating parameters so the Virtual anchor, as a border of the restrained zone, moves when operating parameters change.
Anyway when you dig the pipeline in that point, you not only find anything special there, but also risk to modify the equilibrium of the "axial strain" and probably the "virtual anchor" will move somewhere.

I needed to say this because the description of the virtual anchor as a length from a variable reference point seems for me to be a kind of Zen. IMO, makes sense when you replace the "reference point" with "free (unrestrained) end".

Also, I do not second the idea to stop modeling at VA.
It is true that, in real word, the points past the virtual anchor length have no effect on piping on the "near side" of the unrestrained end.
But when VA is 100 meters (from the unrestrained end) and you finish the model to 101 meters, it is you that introduce an effect of the model ended @101 meters to the piping on the "near side" of the unrestrained end. This is a "cap effect" and means that software cannot understand that in that point pressure has no effect as modifying the stress and strain.
Instead software counts the truncating point having effects as a cap was placed at pipeline truncating point. Or, continuing with Zen Mind, I would say that we evolve from a virtual anchor to a virtual cap.

BTW, it would be possible to construct a special FEA element that inhibits the cap effect? Something called "pipeline truncating element"?

Best regards,
M