Dear Sir,

1) Kindly explain any one what is Virtual anchor length ?

2) How to correlate this VAL with the actual piping model / iso. Why virtual anchor length, say 94522 mm, is appearing in the Caesar result against all the elements. What does this mean or signifies.

3)Also I want to understand that if such virtual anchors (where axial displacements becomes zero)are many for very long buried gas pipeline model like 200-300 km or even more, then whether we need to consider any thrust block for pipe stability at such virtual anchors (null points)for high seismic zones? what is the recommendation from a stress engineer expected other than stress qualification of the such long buried pipeline.

Kindly explain

Warm Regards
Sharun Suresh

1) The virtual anchor length is a distance. This distance represents the length of pipe (from some reference point) that is needed to build up enough friction resistance to restrain the thermal growth. Points past the virtual anchor length have no effect on piping on the "near side" of the reference point.

2) The virtual anchor length is a fictitious point, you can't walk up to it. For example, assume you're standing next to a location where the pipe is buried, and the VA is 94.5 M down the line. If you walk down the line 50 M, the VA is still 94.5 M away, not 44.5 M. Why - because you moved the reference point.

The VA is useful in that it is an indicator of where you can stop modeling the pipe system. For example, assume your pipe leaves the area of interest and runs for several kilometers. If your VA from the last point of interest is 100M, then you can stop modeling at any point after that 100M - you don't need to model several kilometers of pipe.

3) No, the virtual anchor point doesn't exist, it is not real. The VA is just a distance past which you aren't concerned with (compared to your area of interest).

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

Thank you very much for the valuable reply and efforts to explain things.

On point no. 2, I find Mr. Mariog's thought convincing in the sense that VAL could be measured and reached so long as the transition point on AG (preferably with an anchorage) is defined as a reference point which is fixed. With other design conditions, VAL could be worked out manually or by Caesar.

I shall highly appreciate if the below point is clarified :

If the UG pipeline takes turns, changes in elevations and such major directional changes occurs as always expected in pipeline running several hundred KM, then still complete pipeline modelling is not required ? Please clarify.
In such long distance pipeline, Suppose pipe comes up AG after few KM run and further goes UG and this happens at quite some places for pressure boosting or other control requirements, if any, than do we foresee complete pipeline modelling and along with any thrust block requirements ?

Sharu,

First, the theory and what software like Caesar can do is to count axial stress and strain. No buckling is considered. However, in reality, the pipeline in a soft soil can have a behaviour different than we assumed to be, because the energy necessary to buckle is less that the energy necessary to produce the counted stress-strain axially. But this is only an warning in our discussion.

In case you have a buried bend somewhere (as a part of a Z "offset"), let's say 300m from the pig trap, and VAL is 200 m, you may evaluate the possibility of that bend to move or not. I mean in this case it is not so important that the distance is greater than VAL, it is likely that in those 300 meters you haven't 100 m restrained because in the end of the segment with length of 300 m (where the bend is), soil cannot restrict axial movement by a force as calculated for a VA. However, in case you have a strong block at that bend (in practice, in our days- no, you haven't such block) you'll have indeed 100 m restrained there.

What I try to say is that you may apply such judgement case by case, looking to both ends of a straight segment of the pipeline and decide what is to be included in your model. And in case you decide to eliminate the middle part of a long straight segment, the problem of the "cap effect" (at truncating points of the model) is a supplementary issue that you need to solve, case by case.

Dear Mariog,

Thanks for the explanation.

May request you to kindly reply little more specific to my query as I am still not very clear.

Also please tell me more about the "cap effect".

Regards

Additional Comments on Mariog's 1st post above:

1) Yes this is all true. But caution is in order since the restrained / unrestrained state for a particular pipe element can vary with a change in temperature. So in one load case with T1 the restrained/unrestrained state of an element can be different from another load case with T2.

2) I hesitate to say the VA is a real point because that can lead to trouble. A few years ago we had a client with a buried line and information on the VA (from some reference point). They wanted to add a valve in the line, so the went down the line some distance past the VA and started digging. They were surprised when the pipe jumped out of the ground, stating "but we were past the VA". So yes, your statement "but also risk to modify the equilibrium of the "axial strain" and probably the "virtual anchor" will move somewhere" is very true.

3) I also agree that you just don't stop modeling at the VA. There are too many uncertainties involved (soil data from the Geotech group, the theory employed, construction/installation issues, etc, etc). For this reason, as a rule of thumb, we suggest going out to 2*VA to stop the model. This is not a hard and fast rule.

Thanking you sir for simplifying some of the basic concepts.

Meanwhile, being a beginner in buried pipeline analysis, I was doing some study for a while and waiting to hear more from Mr. Mariog regarding some unanswered part of my query, I further would like to present a stock of problems I am facing as below:

1.What are the various methods to carry out Seismic analysis of Buried Pipeline (running hundreds of KMs)? Can we somehow simulate in Caesar?
2. Which method gives the most promising results of such pipeline?
3. What are the various inputs required to proceed further (additional inputs apart from conventional inputs as required in Caesar)?
4. What would be the possible causes of failure and how to mitigate?
5. How to simulate underground faults crossing?, what input expected from civil ?
6. Generally what are the special stress recommendations in such cases? Any guidelines for placing thrust blocks in such cases ?

Anyone, please share your ideas on the above which would be helpful for others also who may be facing similar issues.

Thanks and Regards

Dear Experts,

I have gone going through the earlier posts regarding seismic analysis of buried pipeline. But due to the mixed opinion on this regard, ambiguity still exists.

So request you all to share the latest ideas and on-job experiences regarding the queries posted in my previous post.

Thank You...