An experimental investigation on the influence of an assembly preload on the hysteresis, the drag torque, the lift-off speed and the thermal behavior of three-pad air foil journal bearings

Abstract

Abstract Rotors supported by air foil journal bearings often show distinctive subharmonic oscillations arising from the nonlinear force-displacement relationship of the elastic foil structure. Also, air film induced instabilities may be observed in such bearings. A straightforward possibility to decrease the amplitude of subharmonic vibrations and to increase the onset speed of rotor instability is to introduce a mechanical assembly preload to the bearing. The resultant rise in stiffness and damping and the improved rotor dynamic behavior goes along with a larger drag torque. This leads to an increase in the bearing power loss and temperature with the risk of a thermal runaway, which may cause bearing failure. In this manuscript, the influence of an assembly preload on the bearing hysteresis, the startup and shutdown torque and the bearing temperature at different bearing loads is investigated experimentally. The experiments are accomplished on an automated high-speed test rig which can run at speeds up to 60,000 rpm . Specifying the external bearing load and the rotor speed, this test rig allows for a measurement of the bearing drag torque, the horizontal and vertical displacement of the bearing and of different temperatures of the rotor and the bearing. It is observed that the lift-off speed changes with the bearing load. The experimental results also show that the shutdown torque increases with the bearing load and the assembly preload, whereas the bearing coefficient of friction and the bearing temperature does not significantly change with the assembly preload. The measured coefficient of friction is compared with results from literature which were obtained from pin-on-disk tests at different speeds and loads. It is shown that the discrepancies between the measured coefficient of friction and corresponding values from pin-on-disk tests reported in literature can be traced back to the high dependency of the friction coefficient on the sliding speed.