“Introduction:

For over 75 years machinery condition has been judged from the data obtained from shaft vibration measurements of one type or another. During this time measurement and analysis methods have been refined and improved. However the issue of accessibility to important areas of the machine still exists. In the future this problem will be solved using embedded sensors.

Currently there are four ways to assess machine condition based on shaft vibrations using vibration measurement (1): relative shaft displacement, (2): absolute shaft displacement, (3): bearing pedestal displacement, and (4): Babbitt strength. Because of the indirect nature of the bearing pedestal measurement, this method does not normally lend itself to an accurate evaluation — especially in machines with heavy casings, fluid filled bearings, and high speed. However, it will be shown that some generators with stiff fluid film bearings are the exception. Examples will be provided to show the disparity between direct shaft measurements and pedestal measurements. An absolute shaft vibration measurement provides a measure of shaft stresses whereas the relative shaft measurement relates the amount of journal bearing clearance used by vibration and the stress on the Babbitt of the bearing liner. Thus it could be concluded that absolute shaft vibration is about the rotor whereas relative shaft vibration is about the bearing. As in all situations exceptions exist where pedestal movement is the issue.

“History:

Direct shaft vibration measurements were made in the laboratory as early as the 1920’s by researchers at General Electric Company. This technique involved a shaft stick sometimes called a “fish tail” or guided rod and a seismic sensor (Figure 1). This method continued until recent years when safety experts deemed this method too dangerous. While early industrial vibration experts measured bearing pedestal vibration the US Navy published MIL STD 167 in the 1960’s that involved shaft vibration. Overall vibration levels were measured but the vibration usually was predominantly once-per-rev. In fact the author used this method into the 1980’s because the client would not believe the data gathered from modern sensors. Another method used by the author in the early 1960’s involved strain gauged spring steel cantilevered beams (Figure 2). During this time large turbomachinery was protected by shaft riders that used a seismic sensor on a spring loaded rod (Figure 3). These sensors measured absolute shaft vibration but were subject to noise from rubbing induced friction. Finally in the mid 1960’s Don Bently came up with the solution to measure relative shaft vibration with an eddy current based non-contacting displacement sensor (Figure 4).”