Compressor Performance Problems 2

[gregq8]

Currently the feed rate to our FCC plant is limited by the wet gas compressor performance (2-stage variable speed centrifugal compressor).

We have conducted a survey on this compressor and have plotted the polytropic head (corrected to a standard speed) versus actual flow. The polytropic head is currently significantly lower than previous surveys conducted as well as the manufacturers curve however the calculated polytropic efficiency is in the range of previous survey results. The molecular weight of the feed gas is significantly lower than the manufacturers rated case.

Please would you explain if the drop in head on the polytropic head vs flow curve is indicative of a loss in compressor efficiency or is it due to the drop in feed molecular weight?

Is there a better plot that can be used for centrifugal compressors to track the efficiency that is independant of feed molecular weight?

Thanks

Greg

 

[Homayun]

I am not a compressor expert, but here is my opinion.

I have learned that for compressing lower MW gases, you need larger compressors. And this is easy explainable by the fact that lower MW results in lower density hence, lower pressure. So as you know a centrifugal compressor brings up the pressure with a certain ratio, so you would then get a lower discharge pressure as a result of lower MW as well.

So you either would have the same amount of volumetric gas, but at a lesser discharge pressure, or you would have lesser amount of gas at the same pressure. I would guess.

So this of course translates itself into the compressor performance.

 

[zdas04]

Centrifugal compressors are "dynamic" machines (i.e., they modify the flow characteristics by imparting velocity effects as opposed to a positive-displacement machine that modifies the flow through changing volume).

Do an experiment. Grab a handful of small rocks and throw them as far as you can. Now grab a handfull of leaves (or feathers) and throw them with the same force--notice any difference in the distance traveled?. Drag, bouyancy, and interference have a huge influence on the performance of a dynamic compressor. If you lower molecular weight (and therefore lower density) a very small amount and you'll impact performance significantly.

If your polytropic effeciency is in an acceptable range, then the "problem" is you are pumping feathers when the curves you are working with are expecting something more substantial. There is nothing you can do except generate new curves and change your expectation.

David Simpson, PE
MuleShoe Engineering

http://www.muleshoe-eng.com

Please see FAQ731-376 for tips on how to make the best use of Eng-Tips Fora.

The harder I work, the luckier I seem

 

[pmover]

good analogy David!

from memory . . .

the poly head equation for the centrifugal compressor is:
Hp = R T1 Zavg n /[(n-1)MW] [(P2/P1)^(1-1/n)-1]

for a given geometry & configuration compressor, inlet conditions, and a given speed, the head developed will be the same; the outlet pressure will be lower due to the lower MW.

the compressor mfg should be able to generate a set of performance curves based on the actual MW being compressed. provide the mfg an analysis and the curves can be generated.

options are to retrofit the compressor (i.e. larger diameter wheels - if possible) or replace (rather extreme). are there any process conditions that can be changed to maintain production/operation? highly recommend contacting the mfg and consult with them.

fyi, if requested, i've a few articles written regarding centrifugal compressor behavior/performance and retrofit. please post.

good luck!
-pmover

 

[gregq8]

Thanks for the help David & pmover.

The compressor feed molecular weight changes with reactor temperature and fresh feed properties. These conditions (and some others) are continually changing which is why it would be helpful if we could track compressor performance irrespective of feed mole weight.

The ratio of Hp / Qs^2 is independant of feed mole weight and therefore I would like to know if a plot of Hp/Qs^2 versus Qs has been used by anyone to evaluate compressor performance instead of plotting Hp versus Qs?

Greg

 

[zdas04]

gregg8,
Could you define your terms? If Qs is volume flow rate (at standard conditions) then it is absolutely not independent of molecular weight, a given molecular weight would have the same volume flow rate (at standard conditions) regardless of pressures, but it is very dependent on the makeup of the gas.

David

 

[gregq8]

David,

Qs is actual volume flow rate at suction conditions which as you mentioned is not independant of feed Mw. Maybe this needs to be plotted against compressor speed then?

Greg

 

[jeyx]

Folks,

very interesting discussion! From the above mentioned can we say that the polytrophic head is back proportional to the gas molecular weight? Or this would be something more complex for the real compressors? Our plant FGC underperforming with about 27 % less discharge pressure than its designed for. With that the MW is only 15 % less than as per design intent.

Thanks,
Jeyx

 

[pmover]

ok & yes, good discussion,

there are two other parameters that change in relation to moleweight as the gas composition changes, they are specific heat ratio (k or n) and the compressibility of the gas. variations in the specific heat ratio have minimal effect on the outcome of the head developed, but there is some. the compressibility will change as well, but most likely not to the extent in which the results you've observed. a suggestion is to confirm gas composition and other parameters (p & t) to ensure results.

remember the previous postings about centrifugal compressors and the head developed. for a given compressor geometry, physical configuration, etc. the head developed per wheel will be the same. the outlet conditions are determined by the inlet conditions of the gas.

hope this helps.
good luck!
-pmover

 

[pmover]

correction on previous posting . . .

"variations in specific heat have minimal effect on the outcome of the pressure developed" rather than "head developed".

-pmover

 

[jwalker2]

I think you may want to have a look at ASME PTC 10.

My basic understanding is that if you want to compress a differing gas to the design MW then the performance will change not just because of density but also because of the aerodynamic speed change. The aerodynamic speed is mechanical speed over Mach number I think.

This change means the compressor acts as though the rotors are moving more slowly than the real mechanical speed.

I'm interested in finding out what parameters the vendors use to convert their test data for much lower molecular weights. I am not sure why the general engineering community can't be able to do this themselves.

It usually takes quite a while to get responses these days.

 

[tgmcg]

Using the fan law relationships is a good place to start. More accurate for low volume ratio machines, less accurate for high volume ratio machines. I use these for all compressor re-rate applications and save tons of time vs going to the vendor. The vendors basically confirm my results. And, as you say, the vendors are too busy these days to help on such studies.

Besides, some vendors will slant their analyses to try to sell you a new machine. If you can't develop a reasonable estimate of your own, you're a sitting duck for an unscrupulus vendor.