Dealing with Multi-phase 'Slug' Flow

[Raspberry]

Having carried out a few calculations involving multi-phase flow and checked the flow regime; it is beginning to seem inevitable that the flow will be slug flow when following the Gregory-Aziz flow pattern (Superficial Gas velocity 3-80m/s and superficial liquid velocity of 0.3-10m/s).
I had always wanted to avoid slug flow where possible, but is it such a bad condition? what are the issues to consider when dealing with it?
I also wondered if, once in multiphase flow, the two-phases could be considered to be travelling at approximately the same velocity?

Thanks for your responses in advance

 

[BigInch]

The worst situation is slug flow for three reasons,
1.) High dynamic piping loads generated from unsteady high mass "lumps" flying around in the system can do nasty things to your pipe supports and pipe bends. It's not unheard of to knock an elbow right off the turn.
2.) You have to build large equipment that is sized to take those big liquid slugs over what is usually a relatively short slug arrival time, whereas if you had a steady state gas and steady state liquid flow, you could design for the "average" flowrate of liquid, maybe handle it with a small knockout pot and not have to catch it all at once in a large slug catcher.
3.) The pressures and flowrates are greater in absolute value in slug flow and any variations are much more pronounced than ain a steadier flow regime, which makes things just that much more expensive to build and considerably more difficult to control.

 

[SJones]

If you have an inclined pipe or riser, the gas flow can be cut off for substantial periods until the accumulated liquid has been mobilised.

If internal corrosion control involves chemical inhibition, slug flow can create difficulties with chemical delivery and maintenance of a film owing to high wall shear.

Steve Jones
Materials & Corrosion Engineer

http://www.linkedin.com/pub/8/83b/b04

 

[MikeHalloran]

I was on a yacht where the lift tubes in the vertical wet mufflers went into slug flow at any useful speed. The resulting vibration shook the entire ~200 ft. boat, as the water chamber at the lift tube bottom entry alternately filled with water and then emptied itself abruptly (of tens of gallons of water) and burped out some gas and then repeated at ~2...5 Hz.

The oscillation reflected upstream as large variations in exhaust gas backpressure that acted on the cantilevered exhaust pipes made by my then employer, rocking the huge engines on their flexible mounts. ... visibly.

Smaller lift tubes would shift the flow regime to annular mist, where the mufflers were intended to operate. Out of the 450-ish boats I've worked on in some way, only that one acted up because of an exhaust component that was too big.


To answer your question, the superficial velocities are real, _average_ velocities, and are not guaranteed or likely to be equal, not least because they are averaged both in time and across the pipe area. They are useful in guesstimating what the flow regime might be in a pipe that you can't open, and thence in figuring out which tables, graphs, and equations to use.




Mike Halloran
Pembroke Pines, FL, USA

 

[zdas04]

To expand on Mike's post, "steady state multi-phase flow" is a contradiction in terms. The Aziz map is useful for explaining the concepts, but useless for any practical matter. The problem is that the flow regime changes many times per second. It is always trying to get to the lowest potential energy, but you keep adding mass and energy to the process by flowing fluids into and out of each section of pipe.

At any point in a real system, the flow regime will change hundreds of times per second. At any given time in any multi-phase line you'll see every possible combination of entries from the map. An annular flow will collapse into a wavy flow, which morphs into a stratified flow, which launches a slug a few feet and you have single phase gas for a few milliseconds before you're back to annular flow. Every now and then one of those micro slugs will gain enough momentum to top a hill and start accelerating as a slug. I've seen those chain reactions start slugs that contained hundreds of barrels of water moving at very high velocity. The force of 1000 bbl (175 US Tons) of water hitting an elbow, inlet plenum, or filter pot at 60 ft/sec (40 mph) is much like a train wreck.

The only way I know of to insure against this happening is to regularly pig every line on a schedule that precludes slugs from getting big enough to do damage when they launch. Otherwise you need to design for the odd freight train coming down the line.

David

 

[Raspberry]

Thanks everyone for all this valuable input. I hope i'm a bit better equipped to tackle multi-phase issues now )

 

[btrueblood]

Just for yuks, my anecdote is on the opposite side of the size spectrum: I used to help build small thrusters for spacecraft. A user fouled up the filling procedures for the propellant tank on a particular spacecraft, and nobody caught the error until after launch. The result was a mix of helium pressurant gas bubbles and liquid fuel slugs in the 1/4" o.d. propellant lines feeding the thruster. The question was could the thruster still work under the mixed-flow conditions. We did some testing that suggested it might survive, but a catalyst bed (ceramic nodules) on the downstream end of the pipe (which reacted the fuel to a hot gas) definitely took a hammering from slugs of helium followed by being slammed by slugs of fuel. A very large hollow area opened up within that bed after a series of tests, and the thruster performance suffered. For the end user, the choice was to use the thruster regardless, or junk an otherwise usable spacecraft, so they used it.

 

[davefitz]

Some other issues not discussed above:
- sizing pumps or valves to reliably control or supply flow when the flow vs DP characteristics are unpredictably is problematic
- predicting the process dynamics ( heat transfer, reactions) when a primary flow stream has dynamically varying or unpredictiable flow characteristics

There may be proven methods to avoid slug flow issues, such as using internal ribbon turbulators , or using multiple smaller parallel flow channels to deliver the flow .