Dear expats,

I am doing a stress analysis for a buried pipeline using CAESAR II software. CAESAR II software has two types of soil model type one is by CAESAR II Basic Model which is based on L.C. Peng, published in 1978 and another type is American Lifelines Alliance which is based on "Appendix B: Soil Spring Representation" from the Guidelines for the Design of Buried Steel Pipe by the American Lifelines Alliance.

My question is to follow which type? As my contractor didn’t specify the type.

Will the stiffness, displacement, stress and anchor block loads created by both CAESAR II Basic Model and American Lifelines Alliance will have same result or there will be changes.

I believe you would do better with the American Lifelines Alliance as it is a published document that serves better as a standard.

The two approaches will not produce identical results.

I believe you Mr.Dave that both the results are not identical, then what is the big difference between the CAESAR II Basic Model and American Lifelines Alliance we are not able to get clear idea about both the methods.

If both the methods produce different results then why there is option of both the methods in CAESAR II.

If American Lifelines Alliance is a standard method then why there is option for CAESAR II Basic Model in CAESAR II .

The CAESAR II Basic Soil Modeler (Peng) was part of our original release of the buried pipe modeler. We added the American Lifelines Alliance (ALA) approach much later.
I prefer the ALA method only because it is better documented and formally published. There may be CAESAR II users who use the Peng approach and have adjusted their soil input to produce results that "match" their system's response in the field. For that reason, we usually do not replace existing capability but add new capability instead.

Dear Dave,
I have to model a 8" buried pipeline. In this preliminary phase I have to verify only the pipeline sections at starting and ending point and my battery limit is the first support above ground ( my model starts at this support and I will provide the axial displacement). But, my question Is:

I have a 8" pipeline, buried depth=1,5m and 100°C design temperature; for contain the thermal expansion i placed a buried expansion loop about at 90 meters from the vertical bend(where the pipeline become buried), then I modeled further 250 meters of pipeline up to virtual anchor point. the loop and vertical bends are 40 ND radius (this is a Client requirement, they don't want hot bend) .I have some doubts regard the output results:

1) I modeled the soil by Peng method and ALA Method:
-In the first case the axial displacement at first above support is 88mm
-In the second case the axial displacement at first above support is 70 mm.
As you already said above, maybe ALA method is better than Peng, then I can consider it.

2) I tried also to model a straight pipe section up to virtual anchor point (250 meters). In this case the displacement at first support is 90 mm (axial) and 50mm (vertical); Instead with an expansion loop I have a 70 mm axial displacement. Then, I think that, the expansion loop does not work enough, maybe because the bend radius are so large that the bend is not able to dissipate the thermal expansion (As opposed to have a hot tight bend that works better).
Finally, It's better to insert an expansion loop that absorbs only 20 mm of displacement or it is better to have a straight pipe?

3)Otherwise, What Can I do? I tried to reduce the distance between the vertical AG/UG bend and the loop, I tried to increase the buried depth, I tried to increase the loop width but the results is ever the same.
Please help me, thank you

Is your loop buried?
I doubt that a buried loop, surrounded with soil, will be able to provide the flexibility you want. The loop legs perpendicular to the main run will serve more as axial restraint for the main run rather than strain absorbers.

Yes my loop is buried and it has 40ND Bend radius (Client does not want hot bend); with is axial displacement at first above support are 90 mm.
Maybe it don't give me the flexibility that I want because of the large curvature (If I insert a 5ND radius expansion loop, the displacement decreases up to 50mm).
An idea to decrease the displacement at the first above support could be insert a stop support on it.

Dear,

If I understood as Dave, your loop is buried, so your bends are surrounded by soil. How are those bends going to move as much as you require and absorb expansion?

As Dave said, I also think that those loop pipes will work more as an axial anchor than as a loop. Maybe I am missunderstanding how is the loop placed.

Regards.

Dear Pablo,
Ok I understand that, being my loop buried it is no able to absorb expansion because the bends are not going to move being surrounded by soil.
However, What Can I do in order to reduce the displacements at firt above support? (client does not want anchor block or 5ND bends).

I think that I cannot obtain displacements at first above support less than those already calculated(100mm). then, I will pass this displacements to the piping and they will do their design calculation. (battery limit between me and piping is first above support and my model start at this point).

Have you properly modeled the bend with radius 40ND? How you did it?
And 250 m of straight pipe are enough to simulate virtual anchor point? Do you have displacements at that end- where it is supposed to be that virtual anchor?

Hello mariog,
is the first time for me that I have to model a buried pipeline, then I have a lot of doubts and I don't know if the model is correct.

1) I modeled a bend simply with a piece of pipe,double-clicking on bend check box and appropriate bend radius(8764m). How do you intend to model a bend with radius 40ND?

2) I calculated virtual anchor lenght as per Peng's formula and value is 250 meters. Also CAESAR II Basic Soil Modeler calculated a similar value.
But, virtual anchor lenght is another doubt. At the and of 250 m of pipe I have a 128 mm of axial displacement (in theory after VAL I should have zero axial displacement); I try to model another piece of pipes (300m,400m,etcc) but I have ever a certain value of displacements. How does Caesar show the virtual anchor point if It shows ever a axial displacement?

IMO, you need fine meshing of that 40ND buried bend, let's say as it is made up of multiple bends.

Again IMO, the pipeline end which is supposed to be fully restrained has a displacement because there is a "cap effect" considered by FEA method for your truncated model. Try to identify a point with about zero displacement along the modeled pipeline and check the distance from that point to the bend is greater than the theoretical VAL. I would prefer a fictitious explicit anchor placed at a distance a bit greater than the theoretical VAL, however I've seen that in practice such anchor has the same engineering effect as the matador's red cape....

Regarding VAL, after the expansion loop (node 140) I tried to model 850 meters of pipeline(node 990). I checked the axial dispacements and they ranging from -0.9mm at node 420 to +0.9mm at node 510; then for about 90 meters of pipeline the axial displacements are about zero. Node 420 is located at about 270 meters from my expansion loop, then I think that my virtual anchor lenght is correct. Are you agree?

Regarding fine meshing for 40 ND buried bend, I don't understand very well how to obtain it. Maybe I have to model the bend using short straight segments? How can I model a fine mesh?

Thank you

Bends in buried segments will be finely meshed automatically.

In this case, you are not longer recommending the "undocumented" procedure described in Mechanical Engineering News June 2000?

I took a look at that newsletter.
Now I am unsure of aleps_86's question. Is his bend not meshed at all or is the standard bend mesh insufficient?
I assumed the former and mariog assumes the latter.
(That undocumented "feature" is to "rebury" the buried section where extra nodes are desired.)

aleps_86,

When performing the judgement of the two alternatives you've mentioned, I think the first step is to be sure about the model accuracy. In both alternatives, the truncated models have to count properly the displacement from VAL, and I think you've found a conclusion on this subject.
The modeled buried bends must simulate as accurate is possible the complex interaction between the toroidal geometry of bend and soil. That June 2000 Newsletter, available on site, may be the basis of an improved model.
Maybe you have to pay attention also to the pipeline geometry on the "vertical" plane to the above ground end of pipeline.
And, of course, the soil must be accurately considered.

After rechecking the model, maybe you'll see an improvement of the calculated displacements. As you can see there are lots of "maybe" because I cannot guess what is there without having the model.

Dear mariog,
Yes, regarding VAL I think that I have found a conclusion but the last doubt about it is: At the point where I suppose to have a virtual anchor, I need to insert an anchor restraint or not? because If I put an anchor restraint the displacements increase a little but I think that the reply is No. Is not necessary insert an anchor restraint. Are you agree?


More complex is the topic about the modeled buried bends. I read the June 2000 Newsletter, and as Dave said,if my understanding is correct, I need to break the bends into small bends, bury only the bend and then bury the entire model. But I don't know if I understand very well it. However, If I upload here my Caesar model and if you have time, you can quickly check the model in order to verify if my modeling and output displacements are correct?

Thank you

About your first question: the important point is that you need to have a modeling procedure that assures the pipeline end shall receive a displacement corresponding to the length VAL-to-end. I understand that your results confirm that, with a VAL estimated as 250 m, truncating the model to 250m does not assure this goal. How to solve this problem is after you. Of course placing an anchor beyond VAL, even makes sense, is not only ugly but also shall rise a lot of question marks from others. In the end, just keep in mind that the soil cannot be modeled exactly and software cannot simulate the continuous contact pipeline-soil, so small differences in results are not significant.

About the second point, I'll try to make a compilation from Mr. Diehl's article.

If you are running large radius (50*OD) bends (or any other radius for that matter) through the buried pipe modeler, you can add a node at the start and end of each bend. "Bury" only the bends (by specifying a soil model number for these segments) by click on "bury the system".
Once the program listed "Model conversion complete", instead of clicking on either the "OK" or "Cancel" buttons at the bottom of the window, click on the Close (X) button at the top right corner of the window. After this first step, the restraints shall be not added and control shall be returned to the Underground Pipe Generator and yo can find that the model now has extra soil model nodes. The effect is that node density is automatically increased for bends. The modeler shall maintain the bend designation through these back-to-back "partial" bends so each extra node is a change in direction with a bend.
All you need to do is to continue in the buried pipe modeler and bury this modified model once again, this time bury all sections that should be buried.

The short version is exactly what Mr. Diehl said "That undocumented "feature" is to "rebury" the buried section where extra nodes are desired."

I would add that in the last time I haven't used this procedure so I cannot say how is functioning in the last versions.

Sorry Mariog, but I don't understand what do you mean when you say:




Please clarify.

However I tried also to model 850 meters length of pipeline, both with anchor restraint and without it, and the displacements are ever the same (the difference is only +-7 mm ).

I tried to convince you that you don't need to investigate more this point!
The problem appears when you model too short the straight part of pipeline.
After solving this point, it makes no difference when you model 800 m or 8000 m.

But tell me please, if you model 850m and 250m, both without that anchor, you got the same results?

the cases and axial displacements are the following:

1) 250 meter length: with anchor -94mm; without -85mm
2) 500 meter length: with anchor -92mm; without -90mm
3) 850 meter length: with anchor -95mm; without -94mm

In the end, I have chosen to model 500 meters of pipeline because I think that it represents better my VAL.

OK, I see that for 850m in model, you have a difference of 1mm (with anchor -95mm; without -94mm), not 7 mm as you said.

But as I said, is important to have a good procedure to calculate the displacement received in the end of pipeline (in the hot bend, if there is one\ in front of your battery limit if you haven't that loop/offset/leg). Anyway the actual figure that software returns depends on the soil modeled and this is not an exact science, isn't it? Again, important is to be sure that with your modeled soil your calculation returns a correct value of the displacement.

Yes, the difference is 1 mm, sorry but I made an error error of digitizing at the above post.


If I'd like to verify that the modeled soil returns a correct value of the displacement, Is it possible calculate it by hand? I think that is not a easy job to calculate displacements at first above pipe support,but If a approximate calculation is available will be surely useful in order to verify that my values are reasonable ( I would not that I will provide a value of 95 mm but the actual value will be 50/60mm for example). I hope to be clear

You can consider the old Schnackenberg's article "How to calculate stress in above/below ground transition".
The key parameter considered in article is "the longitudinal resistance of the soil that needs to be known".
In fact, Schnackenberg's model is a simple linear one, where the stress difference between "fully restrained" and "unrestrained" points (multiplied by pipeline metal sectional area) generates a tendency to "elongate" to the ends of pipeline (typically toward pig stations).
Fortunately, this reference article is still included in "Pipeline rules of thumb" handbook.

For your case, the results based on this article are only a rough approximation.
Actually, your model shall return a correct value of the displacement when VAL is established in a correct position; this may be checked by performing few trials that would convince you that points with zero displacements are indeed where they are supposed to be.

Ok I will read the article above mentioned by you.

In the end, Once tha VAL is established in a correct position the Caesar model should provide a correct value of displacements. I understood that maybe, truncate the model at the exact length of VAL is not a good modeling procedure. Then is ever necessary to model a length of pipeline greater than effective VAL, and this length of pipeline is reached after have performed few trials (with different lengths) where we can demonstrate that
the pipeline length does not affect the displacements at bends/transition AG-UG zone,etc.

Hey guys, new user here. I am working on modeling PIG traps for work and am having a few problems understanding whether what I am doing is accurate. I'm modeling a 42" Pig trap and working on the buried section. I have already buried it several times to test the VAL and came up with a value of 393 feet. I am using twice the VAL for conservative results. I did not, however, place a hard anchor at the first node. Is this a good standard practice?

Secondly, I was getting ridiculously high bi-linear restraint values when I would bury the pipe (stiffnesses over 2 million psi, and a yield load of 3 million lbs). I was suspicious of this and read through the Caesar and Autopipe manuals to get a better idea of soil restraint properties and calculations, since I'm mechanical, not civil. I saw that there are equations for distances between soil restraints and when I checked the model I realized that were only 2 spans of well over 300 feet in the buried model. So I went in and added nodes so that the model will better match the equations provided in the manual. This decreased the bi-linear restraints, which seems more realistic, but when I compared the analysis results, the stresses and loads increased quite a bit. Is this simply a more accurate result or did I mess up Caesar's modeling?

Ok, I now have a better understanding on how CAESAR assigns values to the bi-linear restraint stiffnesses, but still not sure about yield loads. Also, when I calculated the VAL by hand, I got a value well over the estimated VAL by CAESAR. I did not take into account the up bend of the pipe, but all other variables were the same for the soil. Is my calculation too simplified?

I think when calculating VAL by hand you should have the same friction pipeline skin as when software calculate it. This is difficult since software uses axial soil springs as defined by ALA in "Guidelines for the design of buried steel pipe", however you may try to make an equivalence.

ZDawg,

I cannot say the anchor you mentioned it is a good standard practice; rather is an "annoying" device to solve a hidden issue.

As logic, truncating the model of pipeline in a zone where the pipeline is supposed to be fully restrained will make pipeline "subject to an end cap pressure force" i.e. will make pipeline as unrestrained in that point and beyond. The hidden issue is that introducing in model a fictitious unrestrained part of pipeline, the model can influence -in calculation- the true unrestrained part of pipeline in case that part is nearby.

Again as logic, placing an anchor where pipeline is supposed to be fully restrained cannot influence the true unrestrained part of pipeline, however that anchor will be seriously loaded because tries to maintained (in model) the restrained status for the zone where it is placed.
And more than this, in case you use it, that anchor should be placed where pipeline is fully restrained, and you need to experiment the place to be sure about this point.
And that's all, it is after you to use it or not!

when it is making fine mesh, in my case CII errors is showing for bend angle less than 5 degrees. How to resolve that??

Go to the Configuration Module, and in the Geometry pane, change the minimum allowed bend angle.