| A new branch of the rotating machinery vibration monitoring and diagnostics field was recently given birth at Case Western Reserve University (CWRU). Namely, the use of Chaos structures and other nonlinear dynamic characteristics in analyzing the vibration signatures of rotating machinery. The research and results presented in this thesis consist of the first laboratory experimental investigation to confirm and further explore earlier computer simulation results related to oil-whip and rub-impact. For this purpose a new test-setup was designed, built, and debugged for the versatile CWRU rotordynamic research rig. Through the use of Spectrum Analysis, and Chaos tracking techniques (Poincare' maps), the nonlinear dynamic characteristics of both phenomena were explored, and computer simulation models were developed. It was shown that oil-whip can cause quasi-periodic vibrations rotating machinery, whereas rub-impact is capable of developing and sustaining chaotic vibrational behavior. The nonlinear hysteresis loop of rotordynamic systems was examined and the results presented here confirmed phenomena uncovered in the earlier computer simulations. A Sommerfelt number-consistent "instability threshold load" was experimentally observed, along with its corresponding hysteresis loop which is similar to the classic instability speed hysteresis loop. In addition, new phenomena, not previously approached or encountered were discovered and are here reported and analyzed. The main new phenomenon encountered was a second (i.e., at higher speed than the classical hysteresis loop) Hopf bifurcation/saddle node instability hysteresis loop, separated by a region of stable-in-the-small operation (i.e., only one solution; no stable limit cycle). A second equally important result was the discovery that spectrum analysis of speed-up or coast-down rotor vibration can be used to locate the saddle node speed. This last result is of high practical importance, as described herein. |
