No abstract

Hybrid bearings represent an attractive alternative to ball bearings for use in highspeed cryogenic turbopumps. However, the internally-developed cross-coupled forces can generate instabilities responsible for a speed limitation of the machine. To reduce these forces and raise the onset speed of instability, the use of deliberately-roughened stators, already successfully tested for liquid "damper" seals, is investigated. Rotordynamic results are presented for a five-pocket orifice-compensated hole-patternland hybrid bearing tested with water at high speed and high pressure. Experimental data show a good prediction of leakage flow rate and direct damping but a significant improvement in stability compared to a conventional smooth-land hybrid bearing, resulting in an elevation of the onset speed of instability. Comparisons between measurements and predictions from a code developed by San Andres (1994) shows good predictions for flowrate and direct damping but an over prediction for the direct and cross-coupled stiffness coefficients by about 30 and 50 percent, respectively. The use of the Moody friction-factor model is thought to be mainly responsible for the poorer predictions of stiffness coefficients. Journal of Tribology

No abstract

Test results are reviewed for two annular liquid seals (L = 34.9 mm; D = 76.5 mm) at two clearances (.1 and .12 mm). The seal stators use hole-pattern-roughened stators that are identical except for hole depths of .28 and 2.0 mm. Tests are conducted at three speeds out to 24,600 rpm and three pressures out to 68 bars. Test data consist of leakage rates and rotordynamic coefficients at centered and eccentric positions with static eccentricity ratios out to 0.5. Test results are consistent with expectations in regard to the reduction of cross-coupled stiffness coefficients due to stator roughness. However, the measured direct stiffness coefficients were unexpectedly low. A partial explanation for these results is provided by measured friction factor data which show an increase in the friction factors for pressure-driven flow with an increase in clearance. A prediction model for rotordynamic coefficients, incorporating the friction-factor data, predicted a substantial loss in direct stiffness but could not explain the very low (or negative) values that were measured. The model did explain the measured drop in cross coupled stiffness (k) and provides an alternative explanation to observed reductions in k values; specifically, an increase in the friction factor with increasing clearance causes a reduction in k irrespective of any parallel reduction in the average circumferential velocity.