Compressor head curve

[essenwood]

Hi all,

This is my first posting, please bear with me, tx.

I am a process engineer and has this question on the manufacturer supplied performance curve.

The curve that is supplied is based on a set of base conditions such as suction pressure, temp, specific gravity (Molecular weight) of gas, ratio of specific heats.

If the real actual conditions are different, this performace curve is no longer accurate of valid.

How do I make the correction to the actual conditions.
I have not done so in my previous works and would like to know how to do it (ie. generate a new curve based upon the manufacturer supplied one).


Thanks so much.

Essenwood.

 

[pmover]

essenwood,
it depends upon the curve(s) you've obtained from the mfg. however, u should be able to predict performance of compressor based on provided curves.
for example, if the compressor curves are in head (polytropic or adiabatic) vs flow (acfm, cfm, mmscf/d) with varying speed lines (assuming variable speed driver), u can predict compressor outlet conditions if inlet conditions are known and having the compressor curves.
simply determine inlet conditions (pressure, temperature, and flow- a must!). ensure that flow units match those of mfg perf curves. now assume that the outlet conditions will be somewhere within the operating region of compressor. if constant speed compressor, determine head point at corresponding flow and then compute discharge pressure (knowing head generated and inlet conditions). if variable speed unit, select a speed line, obtain head at flow point, and then compute discharge pressure.
if a mfg provides you curves and u have inlet conditions, outlet conditions can be predicted. the reverse is true if you know outlet conditions and want to determine inlet conditions; however, inlet conditions are so very important to a centrifugal compressor.
have fun!
if need be, i can now scan and forward a technical document relating the subject matter that will likely assist you.
have fun!
-pmover

 

[pmureiko]

I'll assume this is a centrifugal compressor since you have a performance curve. The overall equation is:

H (head) = Zavg * R * T1 / [MW * np *(k-1)/k] * [(P2/P1)^(np*(k-1)/k)]

Where Zavg is the average compressibility factor and np is the polytropic efficiency.

I'd calculate the head requirement based on this equation and go to the curve to see the flow capacity at these conditions to create a new curve. You will probably want to create a family of curves to show the affect of varying the various parameters.

Compressor performance is sensitive to suction conditions as shown in the above equation. The head requirement is inversely proportional to MW. MW variation can be a big problem if it is a light gas and varies alot (presence of hydrogen). Temperature is absolute and would take large variations to have a significant effect on head.

 

[essenwood]

Thanks so much Pmover and Pmureiko,

Your inputs are very valuable to me. It has given me some directions. If it is not too much trouble; I would be very grateful if you could send the scanned document.

Tell me of anything that I can repay your kindness; I have maybe some interesting engineering reading material that you might want.

Thanks

My email

essenwood1967@hotmail.com