Liquid Ring Compressors and Molecular Weight

[lsuche]

Can anyone describe why liquid ring compressors vary in capacity (ACFM) for the same discharge pressure requirment as molecular weight changes? These compressors are recovering flare gas.

Our vendor has provided different ACFM vs. Discharge pressure curves for each molecular weight gas described in our design basis.

Is this due to leakage at lower molecular weights?

Thanks.

 

[lsuche]

Can you describe why liquid ring compressors (rotary type) vary in capacity (ACFM) for the same differential pressure requirment as molecular weight changes? These compressors are recovering flare gas that varies in molecular weight.

Our vendor has provided different ACFM vs. Discharge pressure curves for each molecular weight gas described in our design basis.

Is this due to leakage (recycle in the machine) at lower molecular weights?

Thanks

 

[zdas04]

I don't work in ACF. Ever. On occasion I will need to convert mass flow rate or its surrogate SCF into velocity and I go through ACF to get there, but I don't make note of the number.

This is especially true of compressor calculations since the ACF on the inlet will always be very different from the ACF on the outlet which on its face seems to violate the continuity principle (it doesn't, but why waste time proving that?).

I know that the liquid ring manufacturers use that stupid terminology, and I always ask myself "if they are this clueless of appropriate units, what else are they doing stupid?".

Liquid ring compressors are positive displacement machines. That says the hp utilization is a function of mass flow rate (which includes molecular weight), suction pressure/temperature, and discharge pressure/temperature. So if you hold hp and suction pressure/temperature constant and vary molecular weight (and therefore mass flow rate) then the only thing left to change is the discharge pressure. You didn't post the graphs, but my guess is somewhere on there they say that the curves were calculated at constant hp.

David

 

[lsuche]

David,

Thanks for the response. You are correct, the curves were calculated at constant hp.

The design basis for our machines was 3000 SCFM of gas with a molecular weight of 6-59. The mass flow is less at a lower molecular weight, but the head requirement is higher -correct? That is why the hp stays the same for a particular SCFM or ACFM rate.

ACFM is ACFM...regardless of the MW. Why does the machine care?

Thanks a bunch for your help.

-Scott

 

[zdas04]

You have hit upon the reason that I don't use ACF. If you were moving empty volume then sure, and ACF is an ACF. You are not. You are moving MASS. That mass (of gas) is a function of pressure, temperature, and molecular weight. Change any one of them and you have change the amount of work that the machine has to do.

"Head requirement" is one of those made-up insider terms that are defined so inconsistently from person to person that without a definition it is best not to use. We talk about pump head in liquids and when I've suggested that it would be more effective to talk in terms of multiples of suction pressure people have been outraged. For compression we talk in terms of compression ratios.

David

 

[lsuche]

I apologize for the confusion...the ACFM vs Discharge pressure curve also has a hp curve associated with it. I will post a curve for further discussion this weekend.

Thanks again.

 

[lsuche]

I apologize for the confusion...the ACFM vs Discharge pressure curves provided by our vendor also have associated hp curves on the same plot. I will post a few of them for discussion this weekend.

Thanks again.

 

[lsuche]

Curves are attached for discussion. You will notice that capacity increases as MW increases...

 

[btrueblood]

Well, ok. Gamma (ratio of specific heats) for H2 is 1.67 (ideal) vs. 1.4 for air. That plays a role here too.

But, something is still screwy with your chart, or the vendor's calcs, or both - the shaft power should not be the same at the same conditions of speed for 100% H2 and air...

 

[MortenA]

Look at zdas first post and then at the curve in the legend at the upper right corner it says (2. line) "Estimated flow...." and the HP is constant for all curves.

The curves are "back calculated" and not real. Note that as zdas says - liquid ring compressors are PD compressors.

Best regards

Morten

 

[ccfowler]

My experience with liquid ring compressors is modest, but I can't help wondering if the solubility of the gas in the ring liquid could be playing a role here. Am I asking a silly question?

Valuable advice from a professor many years ago: First, design for graceful failure. Everything we build will eventually fail, so we must strive to avoid injuries or secondary damage when that failure occurs. Only then can practicality and economics be properly considered.

 

[zdas04]

Solubility is not zero, but it tends to be negligible. If you can find Henry's Law constants for the gas and the seal liquid (often very difficult to find both), you see that most of the time the seal liquid gets satuated in a few seconds and then doesn't take any more.

David

 

[ccfowler]

zdas04,

Thanks for the information!

I had not been thinking of the ring liquid as being cooled and recirculated. The liquid ring compressors (mainly vacuum pumps) in my experience all operated with the ring liquid (water) being continuously replenished for make-up and cooling purposes. The warmed discharged water flow was simply discharged along with the gases (water being dropped out in a separator and drain system in the cases of compressor duty).

I have always been a bit fascinated by these devices for their near-isothermal compression characteristics.

Valuable advice from a professor many years ago: First, design for graceful failure. Everything we build will eventually fail, so we must strive to avoid injuries or secondary damage when that failure occurs. Only then can practicality and economics be properly considered.

 

[zdas04]

Interesting, I don't see many liquid rings with water seals, and any kind of oil is too expensive to dump after one cycle. Even the couple of them that I've seen using water, recirculate the water (probably because in my industry, when people see water going onto the ground they start yelling "SPILL" and it gets to be a pain.

David

 

[ccfowler]

zdas04,

In my experience base, most liquid ring vacuum pumps are used for evacuating non-condensible gases from steam power cycle condensers. In some cases, they are just used as "hoggers" to draw out non-condensibles at start-up, and some cases they are used in continuous duty to remove the continuing small in-leakage air flow from the low-pressure turbine shaft seals, other minor leakage at condenser system gaskets, and dissolved air in the power cycle make-up water flow. (Commonly, make-up demineralized water for the power cycle is introduced as a spray into the condenser for rapid removal of the dissolved gases. Additional non-consensible gas removal is accomplished at the de-aerating feedwater heater--usually the lowest feedwater heater stage that always operates above atmospheric pressure.)

The liquid ring water is simply clean water, and no nasty gases are being drawn from the condenser. The amount of water used by the liquid ring vacuum pump(s) is trivial in the context of the overall water useage of the power plant. In some cases, the water discharged from the liquid ring vacuum pump(s) is re-used for other purposes such as make-up water for ash conveying systems at coal fired plants or make-up water for cooling towers.

In some cases, the liquid ring vacuum pump is a second-stage device with a steam jet ejector being used to draw the deepest vacuum. In this arrangement, there is still no risk of nasties being introduced to the water being discharged from the liquid ring vacuum pump.

Many power plants do not use any liquid ring vacuum pumps to evacuate the condenser. The more traditional vacuum pump arrangement is a multiple stage system of steam jet ejectors with each ejector discharge flow passing through its own condenser to remove all condensible water vapor. (The condensate from each of these condensers is returned to the feedwater flow of the power cycle.)

Valuable advice from a professor many years ago: First, design for graceful failure. Everything we build will eventually fail, so we must strive to avoid injuries or secondary damage when that failure occurs. Only then can practicality and economics be properly considered.

 

[zdas04]

I work in Oil & Gas and nothing we produce can be discharged onto the ground. When I worked in steam plants (nuclear, mid-1970s), the air ejectors were all two stage gas jet ejectors. Never saw a liquid ring until about 8 years ago.

David

 

[rmw]

Steam jets are very sensitive to inlet steam pressure/temperature while in power plants, steam pressures and temperatures along the expansion line of the turbine vary as unit load and heat sink conditions (temperatures) change.

Therefore it is difficult to pick a "point" at which the steam pressure and temperature will be constant enough to be the "right point" to operate the jets successfully. To operate the jets, then a higher pressure/temperature point must be picked and then the pressure reduced steam to the jets pressure controlled at a designed point and as is often necessary, desuperheated as well. Jets like saturated steam, can operate on SH steam - if it is constant SH, but hate wet steam. Another operating cunumdrum - getting operating steam to the jet nozzles "just right". Not doing so varies their performance and life significantly. Not too hard to do in an Industrial Plant which has fixed steam header pressure/temperatures.

Liquid rings bypass all that fun and are affected mostly only by the temperature of the heat sink for the LRVP seal water. Most LRVP's in my experience are isolated from the raw cooling water by a Hx, typically a PHE in order to get the closer approaches. In such cases, the recirculated water in the LRVP's is just that, recirculated water, made up as needed with fairly good quality plant water to account for evaporation losses. I'd shudder at the thought of putting any of the raw cooling water that I am familiar with into the LRVP's I have operated.

But, in the case of water based LR's, inadequate cooling of the LR cooling water will result in the LRVP 'consuming' much of its own cooling water as part of its load due to the heat addition of the driver HP.

Much different world than the oil cooled LRVP world.

LRVP designers will tell you that they don't pump SCFM, they pump what is actually there, ACFM. Not defending them, just saying.....

And in the water world, the ACFM consists of non condensables and water vapor, all of which varies as CCFowler has noted with respect to a wide variety of non-condensable gases coming off the make up water, and coming from the boiler itself as well as the water vapor content which varies with the condenser cooling water temperature (the colder the CW, the more water vapor is 'wrung' out of the non-condensable stream). Lots of partial pressure relationships going on all at once in a stream to a LRVP in a power plant.

rmw

 

[ccfowler]

Well said, rmw! Ideal steam power plant load and cycle conditions seem to happen about 0.001% of the operating time. I've not had the "pleasure" of trying to deal with a unit operating with sliding pressure. Thankfully, I presume.

In air compression or vacuum pump duties (of my experience base), liquid ring compressors/vacuum pumps seem to run fairly consistently with the exit gas flow conditions being saturated (water vapor) at most about 5F above the ring water discharge temperature. These exit conditions seem to be the controlling parameters. Inlet gas/vapor conditions seem to matter mainly for their burden on heating the ring water. Too much or too little make-up flow to the ring water seems to be a good way to mess up these machines, too, but varying the make-up flow within the tolerable range for the particular machine seems to be a useful performance control parameter when needed. At first glance, liquid ring machines seem to be much simpler and more rugged than they really are, but when operated prudently, they can seem to be almost indestructible.

I suspect that liquid ring machines operating on something other than water probably have some comparable operating characteristics where the ring liquid vapor pressure and discharge temperature are the dominant parameters affecting performance. Differing ring liquid viscosities would seem to be quite important, but lacking any personal study of the matter, I'm not at all sure of the balancing of ring liquid characteristics vs. performance characteristics.

I would not want to deal with trying to run any kind of corrosive or abrasive materials through a liquid ring machine. I've not seen any of these machines operating with water of any worse quality than something comparable to normal potable fresh water. No nasties of any kind.

Abrasives would scare me under any circumstances, but corrosives may not be a big deal if the machine's materials are properly selected for the duty. Happily, I've not had the pleasure of dealing with any such application.

Fortunately, I've not yet needed to convince anyone that these machines are NOT good candidates for the application of any magical VFD's. I'm not sure how long it would take for me to stop laughing if "seriously" faced with such a question.

Valuable advice from a professor many years ago: First, design for graceful failure. Everything we build will eventually fail, so we must strive to avoid injuries or secondary damage when that failure occurs. Only then can practicality and economics be properly considered.