Hello all. I am in the process of developing and analyzing a model of a pipeline - in particular, a pipeline expansion analysis. 24", 3km long pipeline lying on the seabed. I am attemping to determine the expansion (displacement) on the ends of the pipeline to be used in other analyses.

A few questions that I've come across along the way...

I used the buried pipe modeler to input the soil data that has been provided to me. after you convert the model, why are all the densities set to zero? I haven't done much soil analyses but I figured it'd be best to use that tool for determining restraint(vertical) stiffness - even though my pipeline is not buried.

Also, when Caesar calculates the virtual anchor length (or when I look at the displacement output report), it calculates a much smaller number than the number I get when checking by hand calculation.

any guidance is appreciated.

Have a look in to CAESAR II User Guide, there is explain the procedure for this. In buried converting procedure Caesar will set the density to 0 and will replace this with restrain.
Have a look in all the post regarding underground pipe and check the Peng paper or American Lifelines Alliance.

The reason the densities are reset to zero is because you have continuous support, and there will be no weight stresses.

Due to the complex interaction of subsea pipelines with the seabed/hydrodynamics, it is difficult to realistically set-up the "soil springs".

And I'm afraid there is a risk modeling 24", 3km long pipeline lying on the seabed, by using Caesar buried pipe modeler. Operational loads can cause subsea pipeline to buckle- vertical (upheaval buckling) or lateral movements (lateral buckling) or axial ( pipe walking). This can lead to very high pipeline stresses and/or relieving stress in other sectors. Caesar modeler cannot evaluate this behavior.

As software SIMULIA/Abaqus would be a solution for a big company.

Now, keeping in mind you haven't the proper tool, I understand you are rather interested to have some "numbers" as displacements at ends.
Your question is focused on the "accuracy" of the virtual anchor length.
I think you find some guidance in Mr. Diehl's posts in
http://65.57.255.42/ubbthreads/ubbthreads.php?ubb=showflat&Main=6625&Number=30861#Post30861
and in other threads in this forum.

My best regards.

mariong,

thanks for your input and reference VAL thread.

re: Caesar as the proper tool for this case - I have a question. Is it not accurate to use Caesar for determing pipe end displacements of a pipeline for only the effects of pressure, temperature and soil springs? I am not designing the pipeline but trying to determine the max displacements from these effects.

My model is simple consisting of a straight length of pipe with certain design pressure, temp. I've used the Caesar soil modeler with our project soil data to convert (bury) the model so I can extract an approximation on the vertical stiffness of the soil - representing the pipeline laid on seabed, and use this vertical stiffness in restraint definition for my model. How is Caesar not suitable for this scenario? I thought Caesar would be a proper tool?

My goal is to determine the displacement of the pipeline from these effects to be used in the adjacent tie-in spool analysis.

thanks again,
jc

For a buried pipe, I think it is- presuming soil springs are proper set-up. For pipeline on seabed- I would say "no": pipeline can buckle and this effect cannot be taken into account by Caesar.


Ok, IMO the only "risk" is to get numbers unrealistic high because buckling is ignored. In the same time, being unable to calculate buckling I cannot see other documented procedure.



I'm not sure I understand your procedure. Caesar considers as reference the American Lifelines Alliance document "Guidelines for the Design of Buried Steel Pipe " Appendix B: Soil Spring Representation.
See http://www.americanlifelinesalliance.org/pdf/Update061305.pdf
I would think that being confidently able to set-up springs in Caesar, the same info (vertical bearing soil springs) should be enough for other software?
Maybe I'm not able to understand your issue.

My best regards.

mariog,

thanks again for your insight.

further clarification on my case... I am trying to develop a model (call it Model A) of a pipeline using Caesar that will give me pipeline end displacements which I can use in another Caesar model (Model B) for the tie-in spool analysis.

the pipeline model (Model A) that I hoped to develop will provide an indication on the displacements based on an operating case (W+P+T1). the only inputs evaluated here in this file would be pressure, weight and soil parameters.

I used a separate pipeline model (Model C) to use the soil data for our project and converted (buried) it to get the soil stiffnesses - of which I am extracing the vertical stiffnesses to be used in Model A analyses.

You say the risk here is to get unrealistic high values. I can`t argue since the values that I am getting do not seem to be inline with a similar design.

Re: buckling; a separate design is being completed by my co-workers on the pipeline - global buckling, stability analysis.

I guess my issue is whether my methodolgy and tool (Caesar) is suitable for the evaluation.

Hope this clears it up. I invite others opinion on the manner I am attempting to evaluate.

regards,

JCSC,

What I don't understand is the necessity to make model C.
You said "to use the soil data for our project and converted (buried) it to get the soil stiffnesses - of which I am extracing the vertical stiffnesses".

So-if I understand well- you have a pipeline laid on seabed, you have some soil properties/description and you try to use Caesar as a kind of "convertor" by simulating a buried "straight length of pipe" in that soil .

For me, it is unclear why the goal is to extract "the vertical stiffness" and to use it as input for Model A with the ultimate goal to evaluate "pipeline end displacements".

In fact, what kind of model will be model A- a buried or unburied one?

Let me explain a little.
IMO, performing a buried model you are going to do an evaluation of "pipeline end displacements" presuming the behavior is as one for a buried "strength length of pipeline".
That means you expect will be a transition from "fully restrained" status (which is a condition of zero longitudinal strain) to "unrestrained" status and the transition length will define the virtual anchor length.

First question would be the accuracy of this model for your application. If you take it as acceptable, you can keep in mind it is conservative because- in reality- it is likely that will be buckling in the pipeline before to have such behavior/end displacements. That was the sense of "unrealistic high values" I've mentioned.

Having so simple model (and in my opinion unrealistic), I don't know why you really need an elaborate "buried pipeline FEA" model to simulate it. IMO, equally you can consider the old Schnackenberg's article "How to calculate stress in above/below ground transition" (a reference article included in Pipeline rules of thumb handbook).
You can see there that "the longitudinal resistance of the soil needs to be known" rather than "a vertical stiffness" evaluation.
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 move which is counteracted by constant and opposite longitudinal soil force.

Returning to your procedure, I would say that IMO the key point is to correct evaluate the longitudinal soil resistance.
For this purpose, you can use Caesar but it cannot due more that the algorithm implemented.

Presuming you are going to consider American Lifelines Alliance approach, you can look to their "axial soil force per unit" evaluation and I guess it's after you to evaluate if the algorithm presented is well fitted with your application.
For details, please see page 11-12 of Caesar User Guide (Buried Pipe Modeling) or ALA document appendix B.

My best regards.

If your Model A is a straight run, you could estimate the normal load on the sea bed and estimate the axial friction based on the ALA calculation. With that number and the axial pipe loads (T+P), you can estimate the displacements by hand. If there are bends, then bearing comes into play and hand calculations will be insufficient for your conditions as would, IMO, a CAESAR II analysis.

Just one more comment on ALA axial soil force formula.
Their analysis work assumes that the pipeline is buried and that means that their estimation of the maximum axial forces at the soil-pipeline interface depends essentially on the soil pressure applied on that interface.

Maybe ALA formula appears to be a little bit unfamiliar because the soil pressure is not constant over the interface, so it must be performed a math integration of the elementary friction force (dF = friction_coef*p*dA) over the lateral surface of the cylinder-interface.

The result is ALA (O’Rourke) estimation of the maximum forces per unit length at the soil-pipeline interface as equal to the multiplication of the coefficient of friction (between the surrounding soil and pipeline) as well as the product of the circumference and the average of the vertical and horizontal soil pressure on the pipe [which is p_med=0.5(γ*H+k0*γ*H]
One more term is added as soil adhesion term.

IMO, this model is specific for buried pipelines and it is hard to be "adapted" to the case of a pipeline lying on soil/ seabed.

Best regards.

good point

Very good description.

Thanks

Water Head is also putting weight on pipeline. I understand that force also need to add..

Please advice. Same case I have to deal with.

Nitesh