hi,

I have to model a vertical long thin pipe with length of 2000m and diameter of 15 cm. Gambit can draw and export the whole geometry but fluent seems to have difficulty in such large meshihng. How do I use fluent to model this long thin pipe?

Why do you have to model the entire thing? Your pipe is over 13000 diameters long. You should have a fully developed flow profile long before you reach the end of the pipe. Perhaps you could reduce the length to a much shorter value (<100m?) and evaluate the exit flow profile.

hi,

actually i am modelling the classic gas lift technology in petroleum production.the well is 2000m deep and nearly 15cm of pipe diameter.It would be great if you can advise me on reducing the length of this model.Because the pipe is directed vertically into the ground, reducing the length will change the bottom hole pressure, won't it?

What you have to do is work from the top end of the pipe backwards. If you have knowledge of the top end conditions, maybe you can use some continuum-type equations for compressible gas pipeline flow to estimate the pressure near the bottom of the well. Look at Crane or similar for such formulae. Then you model the pipe from that point down to the bottom using CFD and the appropriate pressure at the exit of the CFD model. You might even have to iterate a few times, but at least it should run.

hi,

Thank you for your advise.Assume that i have all the B.C. parameters from case study.Do you mean i can shorten the geometric length of pipe by changing the pressure values at inlet and outlet? Will this conform a fully developed flow regardless the length of vertical flow? What is the assumption made?

I am currently working on a 200m (shortened) length and 15 cm diameter pipe. I am using VOF model and injected upward with fuel-pressure inlet of 10psig, CO2 gas injected horizontally into the long thin pipe at 30 meter above from down end oil injetion.gas injection took 5 psig of pressure inlet.

when i set both (oil and CO2)as pressure inlet, iteration at 1s diverged. I managed to set both oil and CO2 inlet as velocity inlet and it can converged.How do i explain this phenomena? as i increase the number of time step, some time step would meet convergence but some do not. Why is this happenning? I appreciate your advise,thank you.

I think I understand the problem a little better now. You're using CO2 to help drive oil up the well. Here's what I was suggesting. (1) Assume that the oil and CO2 are completely mixed at 50 meters from the bottom of the well. (2) Use analytical methods for evaluating pressure drop from 50 meters. (3) Build 50 meter long CFD model. (4) Specify CFD model exit boundary pressure as well top pressure + pressure drop from step 2. (5) Solve CFD model.

If the CFD model shows that the oil and CO2 are well mixed at the model exit, you are done. If they are not well mixed, repeat from Step 1 with > 50 meters.

This approach is OK if the oil and CO2 are well mixed at some point. I think your convergence troubles may be from having too severe aspect ratios on your cells, caused by having such a deep well.

Thanks for your suggestion.It really helps very much.

If my simulation involved design problem where i wish to determine the among few variables:

1)pipe diameter,3 parameters 2)gas injection depth,3 parameters 3)gas injection pressure,3 parameters

and

testing of a kind of material(properties justified) to reduce shear stress during up flow.

i expect to obtain my answer through the outflow velocity thus obtain the volume flow rate to perform the comparison above.

By specifying the outflow as pressure outlet, will it limit the result? besides that,with regards to the analytical method used to analyze the pressure drop, do i assume my oil(liquid) hold up? is it the same portion as how i define the gas velocity inlet and oil velocity inlet down hole?

please advise,thank you in advance