Greetings all.

We are looking at modelling a buried FRP pipeline, however the pipeline will be segmented, ie the pipeline lengths will be joined via rubber couplings.

At this stage we are brainstorming for ways to model this using CAESAR II, one suggestion is to model the entire pipeline with all the coupling joints modelled as rubber bellows, which I hope to experiment with in CAESAR II later this week.

Has anyone else used a similar approach or have any other suggestions?

Do a search under "coupling" and you will find loads of very usefull instructions on how to model them.
Ranka

Cnodes with anchor stiffness

???

You could do this with the expansion joint bellows element where you input all 4 stiffnesses. You can leave the length of this element zero if necessary. You can also do this with restraints between node and Cnode and input the stiffness there. The trick here is to determine the stiffness of the coupling in all six degrees of freedom so you can model it properly.

What are you expecting a Caesar analysis to tell you about this system?

Hi Edward

You are quite right - that is exactly the question I have been asking others within the company when I have asked to look at performing this analysis, however I have not received an answer.

The only useful thing I can think of is looking at the displacements of the piping at the joints and making sure that the displacements are within the allowable range for the joints specified, beyond that I don't believe there is much more I can comment on.

Realistically I know I will not get the relevant stiffness values for the O'rings in the joints etc so I will be having to guess and run a sensativity analysis on these values.

Given I have 3000m of piping, with joints approx every 12m, and both ends of the joint need to be modelled, that is approx 600 joints to be modelled in the buried model.

When I bury the model I lose all my restraint information, hence I would need to add all the joint information (600 of them) into the buried system. If there are any changes to the system it would mean starting all over again. To me it is a big job that I don't think will tell us very much.

In the end what can I do? I don't think we should be carrying out a caesar analysis, but if they want a model, then I have to give them a model, although what I say in the report to go with it I have no idea.

The only bright light at the end of my tunnel - if the powers that be want me to build this model then they have to QA it....

Check the June 1998 Newsletter article on modeling Victaulic couplings for more background.

What you might want to do to speed modeling is to build a small model of the coupling and "Include" it into your buried model. If you choose your node numbering correctly you can probably stick a bunch of these in one "model" and, using the "Piping File Include Increment", drop them in in big batches. You will have to fix the piping node numbers to hook up to the back end of each coupling but that's easier than typing all that stuff in again. Also, be sure to make these couplings weightless for the buried model.

An excellent question... "What are you expecting a Caesar analysis to tell you about this system?"

But these days ignorant people don't want judgement they want some computer output.... not realizing garbage in=garbage out....

A client I have worked for insists on computer output for a certain un-named hazardous fluid. The analysis in my opnion is not necessary for thermal displacements, sustained, or occasional loads. But they demand computer output...

Instead of wasting $$$ on my superfluos analysis via the computer the money would be better spent hiring a decent fabricator/installer, or making sure the operators they hire are are paid enough so they can hire people on the basis of skills needed more so than the lowest hourly wage!

itchy,

If you must model these couplings, I would advise against the restraint method, first because as you have noted Caesar II will remove the restraints when it converts the file, and second because you won't have the room on your spreadsheets to add them after it is buried since there will be so many soil restraints. Instead, model these couplings as zero-length expansion joints as they will not be removed when you bury the model. Once you model one such expansion joint you can carefully use the duplicate command in the list editor with the proper node increment to create the others.

I guess there is more than one way to skin this cat.

Thanks Loren, We'll have a look at all of the possible methods and decide what would be best.

I had built a trial model with just a few joints in it and to get over the issue of not having enough room in CAESAR II to define all the retraints I placed extra nodes near the joint and then deleted the support information at my joint node to enable me to then define my joint.

However I think as suggested it would be much more appropriate to define the joints prior to burying, as then we can copy blocks of elements etc to help build the model.