liquid carry over to compressor

[upm]

upm
Gents,
i know the risk of carrying over liquid to comressors and what damage does it cause (and even explosion). from technical point, the compresor will act as pump and try to cpmpress liquid, but any body can explain how does this take place? is it due to heavier liquid weight thatn gas where the load on the bearing increases excessivliy?any body aware of technical paper on that matter?
pls. explain
upm

 

[phex]

two things coming to my mind immediately:

a) incompressibility of liquids

b) compressors are rotating equipment. if a droplet of liquid enters the compressor, due to high velocities inside, it might be doing damage to all moving parts. this is in line with your remark concerning liquid weight (density, i'd say)

hth,
chris

 

[pmover]

assuming centrifugal compressor...
the situation depends upon the amount of liquid (i.e. a slug of liquid vs. a trickle of liquid) , liquid type (light hydrocarbon vs heavier liquid), and process fluid velocities. liquid slugs will likely shutdown the unit due to excessive vibration levels and may even cause seal damage (depending upon type, etc.), bearing damage, and/or rotor damage.
in terms of an explosion, oxygen or an oxidizer will need to be present. so any combustion taking place will likely occur after the compressor is damaged.
check with the compressor mfg and obtain their analysis.
the best course is to design system to prevent any liquid from entering compressor!
good luck!
-pmover

 

[Montemayor]

upm:

I dearly hope you're refering to a centrifugal compressor rather than a reciprocating one. Otherwise, you've got a major safety and operability problem on your hands.

Neither of the two types of compressors should be made to tolerate this obvious bad operating procedure. Reciprocating machines are not forgiving and will rupture (not "explode" - although, it will seem as such) under the excessive hydraulic pressure instantaneously created by trying to "compress" a relatively un-compressible fluid.

A recip will try to "pump" the liquid through the discharge valves, but first it has to clear the compression chamber that is being squeezed by a positive displacement piston. This machine has not been designed for this type of action. A gas recip simply will not perform as a positive displacement pump because the valve action is not fast enough to evacuate the cylinder in time to prevent a tremendous instantaneous pressure. A centrifugal has a much easier chore in getting the liquid to pass through it - it doesn't have to tolerate valves on the suction or discharge.

In the case of a recip, the effect also puts a tremendous load on the crankshaft - many times breaking it! Of course, the bearing load is astronomical - but who cares at that moment about the effect on the bearings? This is a catastrophic effect that is to be avoided at all costs.

Art Montemayor
Spring, TX

 

[chicopee]

Liquids are nearly incompressible and will damage pistons and cylinder heads as there is very little clearance at top head.

 

[upm]

thannks gents for the help.
In fact i missed adding that it is a centrif. compressor...pls. forgive me.
We do have a high level cut out in the suction drum, the problem was the operator has bypassed the trip logic to line up the process during start up phase.
we are in the process to cancel this feature (ops. access to bypass safetu critical) and i am in the process to evaluate technically what could of happened (theoratical simulation) to be honest this is the first time in my life to do such a calculations and will be very interesting.
liquid carry over to pd compressors are night mare and not even compared to centrif. comp. as most of you mentioned.regards
upm

 

[JeffHofacre]

Liquid carryover is potentially very damaging to any compressor. A reciprocating gas compressor is typically designed with the suction bottle on the top and the discharge bottle on the bottom. This way if a little liquid carries over then the liquid will just pass through. If the bottles were reversed a large quantity of liquid could burp into the cylinder surely causing damage.

The above design is considered a belt and suspenders approach. Liquid separators are installed prior to each stage of compression.

You can contact Ariel's website for more info.

http://www.arielcorp.com

 

[hacksaw]

sounds like you need a HAZOP review of your interlocks.

startup bypasses can be essential to the operator. they can be manually reset to arm the trip logic or self-reseting depending on what the machine requires.

 

[PSucher]

Be careful when turning a centrifugal comrpessor into a pump. We all know they are not designed for such service.

The liquid is a larger mas flow and as such you could greating affect your power draw.

Erosion will always occur as the liquid partical in a centrifugal compressor will never follow ideal tangential path. I believe the few two-phase machines (flow with both gas and liquid) have reduced velocities to minimize this erosion.

If you want gas and liquid machine, get exper companies to sell you one (or redesign yours). It is likley that a new machine will cost you less in the long run.

For papers on this, you may want to seach the TurboMachinery conferences (hosted by Texas A&M) There will likely be a few papers (you can buy) there...

 

[KiwiBob]

Slugging a centrifugal compressor with liquid will clearly do damage - my own experience is knowing of one machine in an ethylene plant in France that had neither high level cutouts nor vibration protection. In that case a gearbox had the teeth stripped (the configuration was ST-Comp-Gear-Comp), one rotor shaft sheared behind the coupling, the casing splitline opened and released a lot of gas (that did not ignite). Conversely I also have experience of liquid ethylene arriving in droplet form at the compressor (evidenced by sudden casing frosting), and doing no damage. My conclusion is that if you're lucky and the liquid rate is low enough, you might get away with it - but not normally. At my (olefins) plant we treat the vibration protection as Safety Critical, and the drum level trips as lower priority in terms of routine checks. The argument is that tightly-controlled vibration protection will act fast enough to trip the compressor as soon as the first serious-sized drops arrive. That is our internal analysis based on 40+ years of operating experience, you have to act based on your own situation.

 

[upm]

Gents,
thanks for the info.
in our olefins plant, we do have high level cut out and vibration trip cut out as well and both are safety critical inst. the problem is when opeartions decided to bypass any of those to stablize the process (during s/u) and to avoid triping the machine once the level get higher. in that regards i am trying (technically) to simulate what will happen in such case.
upm