12 replies to this topic

[fcontrer]

I need to size a vertical separator for a reciprocating compressor with a mist eliminator. The inlet presure is around 15 kg/cm2g and the outlet 38 kg/cm2g. The gas have a higher content of hidrogen 92%.
I need to know, What is the maximum droplet diameter recomended to avoid harm in the compressor?

Thanks in advance.

Best regards,

Frank. C.

[mbeychok]

Fcontrer:

I don't quite understand why you need the droplet diameter to size the vapor/liquid separator for your compressor inlet. The size of a separator drum should be dictated by the anticipated flow rate of vapor and liquid from the drum. The following sizing methodology is based on the assumption that those flow rates are known.

Use a vertical pressure vessel with a length-to-diameter ratio of about 3 to 4, and size the vessel to provide about 5 minutes of liquid inventory between the normal liquid level and the bottom of the vessel (with the normal liquid level being at about the vessel's half-full level).

For the maximum vapor velocity (which will set the drum's diameter), use this equation:

Vmax = (k) [ (dL - dV) / dV ] ^ 0.5

where:
Vmax = maximum vapor velocity, ft/sec
dL = liquid density, lb/ft3
dV = vapor density, lb/ft3
k = 0.35 (when the drum includes a demisting section)

The drum should have a vapor outlet at the top, liquid outlet at the bottom, and feed inlet at somewhat above the half-full level. At the vapor outlet, provide a de-entraining mesh section within the drum such that the vapor must pass through that mesh before it can leave the drum. Depending upon how much liquid flow you expect, the liquid outlet line should probably have a level control valve.

As for the mechanical design of the drum (i.e., materials of construction, wall thickness, corrosion allowance, etc.), use the same methodology as for any pressure vessel.

[fcontrer]

Mr. mbeychok:

I need size the complete vessel (with mist eliminator), the droplet size is important to permit the adecuate settling of it. I am working with hidrogen, i have a very little quantity of liquid, if i follow your recomendation of 5 minutes of residence time the vessel could be enormous (probably i will need to work with an on/off valve at discharge). The vessel alone (without demister) could separate droplet around 150 micrometers but i want to know the maximum limit (of droplet size) that permits operate the compressor without risks, for size both parts: demister and vessel.

Best Regards,
Frank. C.

[mbeychok]

Frank C.:

If you don't expect much liquid per se, then I agree that you should disregard my recommendation of 5 minutes of liquid residence volume, and a manually operated drain valve will be adequate. However, I would strongly suggest an alarm to be activated if the liquid level rises too high ... as a precaution against operators forgetting to drain the vessel.

[fcontrer]

Mr. mbeychok:

I am in agreement with you . But , do u have any recomendation about size of the droplet? (i need to especify the mist eliminator too, for that reason i need the droplet size)

Best Regards,

Frank. C.

[mbeychok]

Frank C.:

I can't help you with the droplet size. However, I can suggest:

(1) Ask the manufactirer of your compressor.
(2) When writing the mist eliminator specification, tell the vendors that it is to be used in a compressor suction knock-out drum ... and I am sure they will know what type of eliminator to supply.

For what it is worth, I designed literally dozens of compressor suction knock-out drums without knowing or specifying the required droplet size, and they all worked quite well.

Regards,

[Art Montemayor]

Frank:

Milton is right. The reason I say this is that I've probably designed (as well as built and operated) as many reciprocating compressor suction drums as he has (probably because I'm much older than he is!) - and I did it just as he recommends. The results on every installation revealed no evidence of liquid particles getting into the compressor and causing any harm or upsets. I presume you are relating to a reciprocating compressor and not to a centrifugal machine. If a centrifugal machine is involved, I would be even more confident of the operation because this machine can handle liquid particles without the fear of valve damage - as in recips.

Mist eliminators are great; I know they work because I've used them. But the application for them is where you suspect that the superficial vapor velocity (related to the degree of liquid entrainment) is higher than the maximum recommended velocity by Souders-Brown equation (Milton's equation). You will obtain a conservative velocity value if you use the Souders-Brown and, in my experience, you needn't bother about liquid particle sizes entraining over into the compressor suction valves.

I also presume your suction gas is saturated (with water?) and carries some liquid condensate with it, motivating you to employ a suction separator. This is a good and proper precaution when dealing with such a stream. That means that when you design a separator with the Souders-Brown relationship, you will get a velocity such that your gas will still be saturated - but not entrained with liquid particles... at least not to where you'll be able to identify and even less measure them. I can assure you that it will all be academic. Your compressor will work well and protected by the inlet separator.

I've done this application with electrolytic Hydrogen, compressing the saturated gas in a Norwalk 5-stage compressor to 3,000 psig and into adsorbent dryers. No problem. My separator had about an hour of design residence time for the water and as I remember, the size was not large at all. You don't have to use that residence time, you know. You only drain on level demand, so just design for a liberal inventory of water to be safe and have a manual drain capability in case of automatic drain failure.

I honestly believe that you have no need to use a demister or any fancy internals inside a properly designed separator using the Souder-Brown relationship.

I hope this experience helps.

Art Montemayor
Spring, TX

[fcontrer]

Mr. Montemayor:

Many thanks for your comments!

Frank C.

[Art Montemayor]

Doug:

Thanks for giving me the opportunity to write about one of my favorite subjects. I suspect Milton Beychok also has a penchant for the Souders-Brown relationship.

First, let me state that I agree with your basic statement. I have designed liquid vapor separators (both vertical and horizontal orientation) using the Souders-Brown equation as well as the gravity separation method which applies the terminal velocity of a particle settling under the action of gravity. I have preferred the Souders-Brown because it hasn’t failed me yet – with or without internal baffles, plates, or wire mesh. I prefer the design of a “pure” Souders-Brown separator nowadays because it eliminates the need for inspecting and repairing internals. The less stuff I put inside vessels, the better off my clients are.

In my opinion, you are absolutely correct in believing that the Souders-Brown equation is based on a base liquid droplet size. By the way, not many people know who George Granger Brown and Mott Souders were – except perhaps for Milton and me! They were both my heroes and idols when I was struggling through Chemical engineering at Texas A&M. Our department head, Dr. Lindsay, studied and worked with Dr. Brown at U of Michigan and he had marvelous stories to tell. Dr. Brown, of course, wrote the Unit Operations book we studied: “Unit Operations”. And in it he describes the derivation of the Souders-Brown equation. I understand that they developed it for determining the correct distillation column diameter and the vapor capacity. This was in the days of bubble cap technology and pre-WWII. To find out the real facts about the relationship, read: Souders, M., and G. G. Brown, Industrial Engineering Chemistry; #26; volume 98; (1934).

For an interesting bio on Mott Souders (by no lesser person than Max Peters), go to: http://books.nap.edu...71.html#pagetop

These two guys were the Right Stuff when it came to practical and profitable engineering. If we only had more of them around today!

You are correct again when you say we left out the mention of other “K” values. My fellow Texan, Ernie Ludwig in his “Applied Process Design for Chemical & Petrochemical Plants, Volume 1, Chapter 4” gives a nice table of “K” values for knitted Mesh. I have my own values and prefer not to use mesh. But I have found his values to be conservative and sound.

One thing I have never resolved, since I designed and built my first separator more than 40 years ago, is: How does one believe the various K’s given by some authors for different droplet sizes? Is there a magical droplet fairy that they dispense, on command, to accurately measure the size of the liquid droplets formed or passed over in separators? And another thing: how fast does the fairy have to be to measure the diameter (although the droplet obviously can’t exist as a sphere under the influence of gravity) of the droplets before they merge or co-join together upon contact with each other? Or do they use a secret, solid-state liquid droplet meter that measures the droplets as they pass in front of the instrument while held in suspension within the vapor stream? I wonder……

Humor and facetiousness aside, I believe that as serious engineers we can appreciate the fact that the design has to consider Stokes Law and the effect gravity obviously has on any liquid particles included in a vapor stream. However, I’ve found that it is next to impossible to discuss the size of the particles without a tongue-in-cheek smirk on my face. Who is going to prove otherwise when one asserts that an equation holds for a certain liquid particle size only -and for how long? In my fondest esteem for their work, I secretly believe that Mott Souders and George Brown identified this technical dilemma and using their engineering brains they took us out of it by formulating an equation that has worked for industrial applications during the past 60+ years. Now, that’s REAL engineering!

Like I said, thanks for the opportunity of sharing the memory.
Oh, and if Milton reads this, I want to apologize for thinking he was younger than I. I fear that in trying to build his ego, I tarnished his pride. Yesterday was my 67th birthday and I'd like to think I'm gaining on him - the company would be great.

Regards
Art Montemayor
Spring, TX

[pleckner]

And from me Art, a very happy birthday and I hope to see you around for many, many more!!