Centrifugal compressor impeller tip speed.

[rQuestionEngineering]

Why impeller tip speed should normally range between 650 and 900 ft/sec (198 and 274 m/sec) ?
What is the consequences going over range or under range ?

 

[georgeverghese]

Going beyond 800-900fps brings the gas at the impeller closer to Mach 1.0 at low pressures / low specific volume - see the expression for sonic velocity under theoretical 100% isentropic conditions. Approaching or exceeding Mach 1.0 makes for unstable flow and creates shock waves at the impeller tip. Industry working limit for max developed polytropic head per impeller ( which is dependent on tip speed) is approx 10e3ft.lbf/lb mass of gas. And max no of impellers you can cram into a single casing is about 10, which translates to approx 100e3 ft.lbf/lb mass per casing. See the chapter on refrigeration cycles, subsection on centrifugal compressors in Perry Chem Engg Handbook, 6th edition.

Would guess polytropic / isentropic eff drops off rapidly as you decrease speed and becomes nonoptimal from a selection - design point of view.

 

[rotw]

You can easily operate at 300 m/s impeller peripheral speed and beyond while still achieving low peripheral mach numbers.
Example: Hydrogen gas.

So the real reason to answer OP question is :

On the one hand, decreasing the number of impellers makes the bearing span shorter and improve rotordynamics; subsquently casing costs less, marginally but still it is less. But putting less impellers increase the peripheral speed. Having higher peripheral speeds means higher stress. If the gas is corrosive, then the yield strenght performed for a given material is also reduced which is even more constrainful. Ultimately the limit is dictated by the yield stress utilized vs. peformed for the selected impeller material. On a side note: open impellers tend to have better aptitude to tolerate higher stress ;some integrally geared compressors for example could spin at much higher speeds (+400 m/s for titanium material / open type); spining faster in this particular case - considering the impeller diameter fixed - would increase the flow coefficient thus the efficiency.

 

[georgeverghese]

The expression for sonic velocity is also a function of pressure at the upstream side of the impeller. i.e vel = (k.P.v)^0.5, [vel=m/sec, P=N/m2, v=m3/kg] so at high suction pressure, yes, you could be operating at higher tip speeds to keep within 0.8-0.9Mach max. The k value for irreversible compression processes is somewhat higher than ϒ, but may be safer not to take credit for this.

Taking credit for higher suction pressure in max permissible tip velocity is yet another safety critical reason to install a PSLL on the suction side of the compressor.