Control of flash separation process

[fd74]

I'm currently designing a small unit to separate ammonia from lubricating oil.
Contaminated ammonia from storage tank is fed to an electric heater and then to a flash vessel. Oil-free vapour is continuously withdrawn from the top of the vessel by means of an alternative compressor (piston-type); lubricating oil accumulates on the bottom of the vessel and is periodically discharged (automatic discharge activated by a level switch).

Control valves are placed (CV1) on the the stream entering the flash drum (FD), (CV2) on the vapour leaving the FD.

The process is designed to be semi-automated (i.e. no automatic control loop BUT control valves are hand-operated).

I REALLY NEED help from any of you experienced on the best control scheme for the flash operation.

To me possible solutions are:
(1) keeping pressure in the FD at design condition by operating CV2 AND keeping the inlet stream flow at design condition by operating CV1;
(2) keeping pressure in the FD at design condition by operating CV1 AND keeping the vapour flow at design condition by operating CV2.

Keep in mind that the whole process is driven by the compressor, i.e. the fluid moves from the storage tank to the FD by pressure differential.

I will really appreciate any suggestion and if you need further details just drop a line.

Thanks in advance,

fd74
(which is not Flash Drum!!)

 

[25362]

Kindly confirm that the subject is a case of an oil mist contaminating an ammonia vapor main stream and should be removed (why?).

Although I am not an expert on ammonia systems, from what I've had the chance to work with I have several questions to which I'd appreciate having your comments:

1. Is it worthwhile to heat up the incoming ammonia oil-laden stream ? If so, why ?

2. If the system is working under the compressor's pressure differentials, and all the gas is sucked by the compressor, why does one need control valves around the KO pot at all? These would, no doubt, reduce the compressor's suction pressure, with its concomitant consequences, wouldn't they ?

3. Have you considered on-line oil/ammonia separators (in the market) with a minimum of frictional pressure drop ?

4. What kind of system is this, is it by chance a refrigeration unit ?

Thanks for your attention.

 

[quark]

One more question, are you using a non lube type compressor? If not what happens to the oil mist added during compression(hmmm...a battery of oil eliminating systems)?

PS: I will never operate a PD compressor with a closed suction either.

 

[MortenA]

I think that the level in the KO drum will be more easily controlled if CV-1 is on the liquid outlet from the KO.

Furthermore if possible it would be better to adjuste the speed of the piston compressor for pressure control in the KO drum. I would also recommend a min flow bypass line - just in case. I dont know the type of piston compressor you intend to use - but i would guess that most have a minimum speed.

Best regards

Morten

 

[fd74]

First, thank you all for your replies.
I understand from your comments I've not been so clear in my description of the system, so I'll try to be more precise and answer to all your questions!

It's not a refrigeration unit, but a small sized application for ammonia recovery. Oil-laden ammonia is stored as LG under pressure in a suitable storage vessel, and from this one fed to the plant as a liquid.

The stream is heated, then undergoes expansion through CV1 and separates in the FD (KO in effect) in two different streams: almost pure ammonia (vapour-top) and oil (liquid-bottom).

Process is semi-continuous; ammonia is withdrawn from the top of the vessel while oil ACCUMULATES on the bottom until its level reaches a pre-set value, and its then discharged automatically from the vessel.

Main question: pressure in the KOD should be the main parameter to keep constant, to achieve desired separation and to maintain constant suction conditions for the compressor. Which are the streams that need to be regulated in order to obtain constant pressure in the KOD?

To MortenA: PD compressor speed-adjustment would be a nice idea and I'll take it in consideration for further improvement of this unit. Do you have any suggestion on how to obtain desidered conditions with the present configuration?

To quark: Nice kick...anyway it's non-lube PD compressor

Best regard to all,

fd74

 

[25362]

As far as I could grasp from your explanation the ammonia is stored as a supernatant liquid on top of an oil (lube/seal oil ?) layer. If so, why not install an automatic or manual oil drainer over there, saving the efforts and money needed for vaporizing, flashing, compressing, condensing, etc. ? What is the required ppm level of oil in ammonia ?

There is probably a satisfactory explanation for those questions, fd74 is kindly asked to refer to.

Still I don't fully grasp the whole situation. How is the tank pressurized in the first place ? Please explain.

You plan to use heating, vaporizing, and re-compressing the oil-free heated ammonia vapors... where to ?

As I said, there are highly engineered coalescing elements in the market that would provide lubricant separation from ammonia vapors down to 0.1 ppm with little friction drop.

And activated carbon filters are available that could bring the oil-in-ammonia content down to 0.005 ppm, which could be considered effectively oil-free. Of course, in this case friction drops should be taken into account.

The KO drum must be so designed (with a demister?) to avoid the carrying over of oil droplets as a mist by the high velocity ammonia vapors.

Anyway, considering the suggested setup, if the oil-free ammonia gas is to be PD re-pressurized at constant suction/discharge pressures, I think one way to do it is by flow controlling the liquid ammonia fed to the vaporizer, reasonably assuming that the (limited) solubility of oil in liquid ammonia is practically constant under the given conditions.

There are ways for unloading ie, changing the volumetric efficiency of PD compressors, but I feel this is not one of those cases.

fd74, any comments please ?

 

[quark]

There are some hickups in your process. You have to create enough pressure drop across CV1 to get a phase change and also create reuired pressure in the flash drum by the compressor (or any other means)to avoid any back pressure on the evaporating liquid.

This requires your CV1 to be operated at a single fixed position for a fixed upstream and downstream pressures.

Secondly, draining of the accumulated oil becomes difficult during compressor operation as the flas drum is maintained at a negative pressure.

Still if you want to go for it, then my choice will be to control the compressor speed with respect to the flash drum pressure.

Regards,

 

[25362]

To fd74.

You say: the stream is heated, then it undergoes expansion through CV1 and separates in the FD (KO in effect) in two different streams: almost pure ammonia (vapor top) and oil (liquid-bottom).

Sorry, but this will not work at all.

As you explained it, ie, heating the liquid and only then flashing it to a lower pressure, would mean part of the ammonia will stay as a (cooler) liquid in the KOD together with the oil.

Try to make an enthalpy balance or use a Mollier chart and you'll see what I mean.

 

[StoneCold]

25362
Is right. It will be like high pressure condensate being let down to atmospheric pressure. A small percentage will flash but the rest will remain a liquid.

 

[fd74]

I would like to thank all of you for your tips, project definition is at a very advanced phase so I'll take in consideration all of them trying to make things work.

Anyway, to 25632, you are very active in the thread but not being familiar with ammonia and LG technology lead you to wrong assumptions.
Two things in particular: (1)ammonia is stored under pressure as saturated liquid (starting point for the process on Mollier chart); (2)it's not possible to withdraw ammonia vapor directly from storage tank because that would cause quick heat absorption from the surrounding by the evaporating liquid. As ultimate consequence you would experience freezing of the mass.

Hope this makes things clearer.

Thanks again to everybody,

fd74

 

[25362]

To fd74. I feel you are offended, and I ask you to forgive me. I've never intended to do that.

I never suggested to remove vapors from the tank. I did indeed refer to the compressor's discharge vapors, ahead of the condenser.

I also suggested to drain the tank if the lube oil is insoluble in liquid ammonia, as it usually is.

As for the design, as described, please explain how do you plan to keep saturated liquid ammonia in the liquid state by heating it up, if that is your plan, ahead of the expansion valve V1.

I'd be grateful to learn from your explanations.

Kindly refer to my comments above.