6 replies to this topic

[Said Salim]

Hi All

I have digitized all our centrifugal compressors performance curves and obtained correlations between polytropic efficiencies vs Inlet flow for all stages at different rpm. However, those curves are valid for design inlet conditions ( Pressure, Temperature and Composition). I am intending to improve those correlations to address all changes in the inlet conditions of each stage. Can any share supportive documents or references that address impact the off-design conditions on the obtained value of polytropic efficiencies.

Said Salim

[Bobby Strain]

I would start with the curves you generated. Then calculate polytropic efficiency by taking measurements. See if it fits. As I recall, the curves should be reasonably close if the inlet conditions are not greatly different from design.

 

Bobby

[Pilesar]

Multistage centrifugal compressors are tough to extrapolate too far away from measurements. There are simulators that will model the compressor 'by the wheel' instead of 'by the stage' but the usual steady-state simulators do pretty well adjusting performance based on changing conditions. Compressor vendor curves have a tolerance that I think is too broad. Measured values give better accuracy. Use the Fan Laws to estimate how performance would change with different inlet conditions if you do not have a simulator do this for you. Assume the measured efficiency is fixed for the same fouling condition -- that is about the best you can do. 

[breizh]

Hi,

You may find tips in the document attached .

Breizh

[Said Salim]

Bobby
There is minor changes in pressure and composition. However, the difference is significant in inlet temperature especially when you know thay our cooling system is dependent on sea water temperature which is varying during the day itself. Also, for air compressors inlet temperature is affected by ambient teamperature which is changing between summer and winter by more than 7 degrees C.

[Bobby Strain]

Get the data. That's the only way you can confirm anything. Of course, the data must be accurate.

 

Good luck,

Bobby

[Said Salim]

Bobby
I am getting the actual data to calculate the efficiencies. However, my idea is to take design data to the new conditions and see how design values will change and finally compare those values to the actual values.