Multi phase flow in pipelines

[polariskid]

I am trying to find a rule of thumb for the minimum gas velocity to move water in a pipeline. I have a low pressure system and I am concerned that if I have too large diameter pipe that I will accumulate fluid.

 

[KernOily]

As far as I know there is no rule of thumb. (Somebody please feel free to correct me here...) The whole thing is strongly dependent on the densities, viscosites, and fluid velocity. Best thing to use is a two-phase flow software package and do the calcs iteratively. Been there, done that, bunch of times. Thanks!
Pete
pjchandl@prou.com

 

[ggordonstewart]

There are some commercially available software programs that help you to estimate pressure drops, liquid hold-up and surge/slug volumes such as Hysys and PipeFlo.
This is not an exact science, so err on the side of caution when designing a piping system and eliminate any potential cause of surging/slug formation that is possible to do so. G. Gordon Stewart, P.Eng.
Gas & Oil Process Engineering Consultant

http://www.ggordonstewart.com/

ggstewar@telusplanet.net

 

[zdas04]

For the last several years, I've used 36 ft/sec as a minimum design velocity in a "dry gas" gathering system. This velocity has proven pretty effective at reducing accumulations of condensation--not perfect, but I don't get liquid slugs very often anymore.

 

[KernOily]

zdas - what is your basis for this velocity? Did you do some testing? What liqud fraction had you been dealing with? What is the gas? I'd love to hear more about this. Thanks! Thanks!
Pete

 

[zdas04]

I started with the vertical flow correlations (i.e., Turner, Grey, etc.). When you cut through all the pretty pictures, the minimum velocity for lifting water always ends up around 32-38 ft/sec with 36 ft/sec showing up often. All of the correlations are based on observations of empirical data and some pretty creative curve fitting. I made the huge leap of faith and decided that if we could unload a vertical pipe at 36 ft/sec then horizontal and inclined pipe should certainly unload at that velocity.

The gathering system I'm dealing with is pretty low pressure (around 40 psig) and all the wells have decent separation equipment so the resulting stream is mostly gas and water vapor. The problem is that at wellhead pressures approaching vacuum the 100%-humidity water content is 6-10 BBL water/MMCF so there is a bunch of water in the lines. The system has 250 miles of pipe serving 114 wells and we pig about 500 bbl/week that had condensed in the gathering system. All that the minimum velocity does for us is to reduce the incidence of collecting enough water to slug.

 

[a1j2]

zdas04 could you provide a little more detail in the "dry gas" term..please.... I deal with a retrogade gas condensation fluid in a 200 km 16 inches pipeline and I will like to use your 36 ft/sec term... On the other hand the 36 ft/sec is for the gas speed on the free vapor space of the pipeline (full diameter MINUS the space taken by the liquid condensation) or on the area based on the full diameter of the pipe??? Does this rule works for gathering and transmission lines or only in one of them?? Thanks

 

[Fausto]

Hi,

There is the "Turner Limit" or "Turner Velocity", but
that concept is applicable to vertical flows only (for
example, in gas wells producing some water). I don't think
you can extend the principle to horizontal pipes, though. I
would suggest you to use a computational package to do such
calculation.

Usually, low G-L rates in a pipe yield stratified
flows. Accumulation of liquid in this case happens,
chiefly, in the lowest parts ("valeys&quot of the pipe. You
might try to investigate the instability of a solitary
wave as a criterion to determine whether or not the water
will flow.

---Fausto

 

[zdas04]

Fausto,
You're both right and wrong.

The vertical-flow correlations turn out to actually have nothing to do with horizontal/inclined flow. The authors of those correlations never intended them to be extrapolated outside their experimental conditions which included a vertical flow-orientation.

The pipeline models generally do a poor job of modling water mobility - they all assume a steady-state multi-phase flow regime which is a contridiction in terms. All the flow visualization work that has been done in the last few years shows that every possible flow regime exists for very brief time and then the energy level changes and the flow is different.

While doing research for a paper at the SPE, my co-author found some new flow-visualization work that was done in Holland that shows velocities above 4 m/s (13 ft/sec) will keep liquids mobile in any size pipe. He (my co-author) said that the video of the test was pretty compelling, but I haven't seen it. He is usually reliable on flow stuff (he's the Petroleum Engineeing department head at a prominant west-Texas university) so I've changed my design conditions to use 13 ft/sec as a minimum instead of the 36 ft/sec that my intrepretation of Turner suggested.

David

 

[KernOily]

Hi David:

You said: "The pipeline models generally do a poor job of modling water mobility - they all assume a steady-state multi-phase flow regime which is a contridiction in terms. All the flow visualization work that has been done in the last few years shows that every possible flow regime exists for very brief time and then the energy level changes and the flow is different."

I'm not sure I understand your point here. When you say that every possible flow regime exists for a brief moment and then things change, is that at the pipe entrance, at a disturbance, etc.? My day job with 2Ø steam systems, and the flow visualization videos I've seen, both show that the flow regime is stable over pipe runs of practical length as long as nothing else changes.

"While doing research for a paper at the SPE, my co-author found some new flow-visualization work that was done in Holland that shows velocities above 4 m/s (13 ft/sec) will keep liquids mobile in any size pipe. He (my co-author) said that the video of the test was pretty compelling, but I haven't seen it. He is usually reliable on flow stuff (he's the Petroleum Engineeing department head at a prominant west-Texas university) so I've changed my design conditions to use 13 ft/sec as a minimum instead of the 36 ft/sec that my intrepretation of Turner suggested."

I don't get this either. Please correct me if I'm wrong, but... this is all a function of liquid fraction. 13 ft/s, which I think you mean is vsg, or the vapor slip velocity, is not enough to guarantee a particular flow regime unless you know what the liquid fraction is - no?
Thanks!
Pete

 

[Fausto]

Hi Zdas,

You are 100% correct, I didn't pay attention to the fact
that Polariskid was talking about *horizontal* flows in a
pipeline, whilst the Turner Limit is for *vertical* pipes.

My sincerest apologies to you all for the inappropriate
and misleading post. I will be more careful next time.

---Fausto

 

[zdas04]

74Elsinore,
I don't know how I missed your April post to me, I wasn't purposly ignoring you.

Multi-phase flow is very energy intensive, and each type of multi-phase flow exists in a different total-energy envelope. As the gas does work draging water drops along in a slug flow, the kinetic energy of the gas is transfered to the liquid. When the energy of the gas is low enough, it can no longer support the liquid and you degrade from mist flow to wavy flow (for example). Making whitecaps on the surface of the wavy liquid takes even more energy so the KE of the gas drops further. The liquid becomes less mobile and accumulates into a slug. As the slug moves, it tends to compress or "re-charge" the gas enough to support mist flow again. I've seen many visulaziations that show this cycle happening with great frequency. All of the visualizations I've seen are air/water (at ambient temp) or natural-gas/produced water at nearly ambinent temp. I've never tried to do the math with steam because all of my steam experience has been at very high pressures and any liquid accumulation would have been disasterous to the HP turbine.

Your second question is a very good one. The way I interepret the explaination I got of the Dutch work is that below 13 ft/s the gas can no longer support mist flow and the wavy-flow/slug-flow cycle starts. The corolary is that above 13 ft/s the mist-flow can be sustained. This velocity is superficial gas velocity (assuming that the entire line is available for gas flow). Of course there is some liquid cut above which there is simply no way for the liquid to fit in the line as a mist. At that point you usually abandon liquid-cut gas correlations and go to gas-cut liquid arithmetic. I've found that point to be a function of pressure and temperature.

David