Mathematical and Experimental Modeling of Bearing Outer Race Creep at High Speed.

摘要

As the need for super high speed components (pumps, motors, etc.) continue to grow rapidly, so does the need to make measurements at speeds higher than ever before. Bearings are a major component in any rotating system. With continually increasing speeds, bearing failure modes take new unconventional forms that often are not understood. Such measurements are impossible if bearings fail to perform.;One such problem which is becoming increasingly prominent in high speed bearing applications is the rotation of bearing outer races also referred to as "creep". Once a bearing race starts to creep; failure of the application is eminent.;Race creep accounts for problems as simple as induced vibrations or heat generation, to as catastrophic as downing an aircraft or destroying expensive turbo-machinery, hence the importance of this study. Literature review of the topic reveals obvious lack of research.;The topic of interest in this work is the dynamics of rolling elements at high speeds. Centrifugal forces and gyroscopic moments induced by bearing high speed rotation of rolling elements, and their very likely effect on race creep is little discussed in literature. Focus will be on bearing "outer" race rotation of an angular contact ball bearing.;In this work, a mathematical model for the dynamics of rolling elements in a high speed bearing application is derived. An expression for the preload values counterbalancing the torque driving outer race to rotate is derived.;Two experimental apparatuses were developed to laboratory reproduce the phenomena and verify governing equations predictions.;Two specially designed Strain Gauge Transducers to measure creep torque exhibited on bearing outer race, and bearing preload were deployed in both apparatuses. Results obtained from equation predictions for creep torque and preload were compared with measurements from the two transducers.;To our knowledge, this is a novel approach; hence, a reasonable error budget between theoretical and experimental values was relatively high; compliance between the trends of the theoretical and experimental curves gave more confidence in the model developed. Trending usually weighs heavier in evaluating experimental work than just scalar error values. Moreover, calculated preload values from model proposed the almost tenfold that recommended by bearing manufacturer. Manufacturer values were insufficient to prevent creep, further supporting the inability to reliably predict errors.;The unique design of the experimental apparatus posed a challenge to some of the theories proposed in literature as causes of race creep. Two theories explaining creep as either the effect of bearing applied radial load, or as a travelling strain wave exerted by housing on race are found in literature. The deployment of the test bearing outside the housing in the second apparatus clearly discredits both accounts, yet race creep was still present. Another interesting finding was that race consistently crept in opposite direction to shaft rotation at medium speeds, then started to change directions as speed continued to increase, first with low torques then with more pronounced torques at higher speeds. It was observed that this change in creep direction took place near the natural frequency vibration of the outer race.;Governing equations obtained from model proposed would help engineers, design-in critical parameters necessary to eliminate race creep at the early stage.;Experimental measurements from this approach are deemed to be the most accurate and are expected to result in a state of the art semi-empirical, semi-theoretical model for creep elimination for Angular Contact Bearings.