Performance of two 10-lb/sec centrifugal compressors with different blade and shroud thicknesses operating over a range of Reynolds numbers

SUMMARY A 10-lb/sec, 8:l pressure ratio, centrifugal compressor was tested at the NASA Lewis Research Center as part of a program to study the effect of scaling centrifugal compressors to small sizes. It was scaled down to 2 lb/sec where adjustments were made to increase the blade thickness and fillet radii. The \L) adjustments were required because of structural and manufacturing considera- W, tions. To determine the effect of the modifications, the 2-lb/sec design was ~r: directly scaled back up to 10 lb/sec and a second compressor with thicker rotor blades and diffuser vanes was built. Both 10-lb/sec compressors were tested over a range of speed, mass flow, Reynolds number, and tip clearance. Impeller shrouds with two different wall thicknesses were tested to determine the effect on performance of heat transfer from the compressor exit to inlet. C \c I lest results at design speed, with the design tip clearance of 0.009 in., showed that the thin compressor produced a pressure ratio of 7.69, with a total to total adiabatic efficiency of 0.792 at the design mass flow of 10 lb/sec. At the same mass flow and operating condition, the thick compressor produced a pressure ratio of 7.64 with an efficiency of 0.777. The efficiency loss that occurred as Reynolds number was reduced was found to be dependent on the thickness of the shroud that was used. The exponent commonly used in a relation between loss and Reynolds number was found to be equal to -0.095 for both compressors when tested with the thin shroud, and -0.183 when the thick shroud was used. The tip clearance of the thin compressor was increased in 0.001 in. Incre- ments at design speed with an inlet total pressure of 14 psia. The efficiency was found to decrease about 0.25 points for every 1 percent of blade height increase in tip clearance.