Many people think Vibration Analysis is just a term engineering school geeks use. Any good mechanic, or electrician, knows vibration analysis can tell the trained technician how well a piece of rotating equipment is operating. College trained engineering students understand vibration analysis to be a very valuable tool in the preventative maintenance and troubleshooting of rotating equipment.

When your machine starts making more noise, you understand something must be wrong. Vibration analysis can determine whether it is system noise: cavitation, air flow, piping, or resonance; or machine related: bearing wear, imbalance, looseness, misalignment, etc... But it is not a lot of hocus pocus. All you need to do is understand what type of vibration data to collect, which may be different for each type of equipment, and observe the amplitude & trending.

Since my expertise is with blowers, motors, pumps & gearboxes, my vibration data collection centers primarily on the bearings. We measure in 3 planes: vertical , horizontal, and axial to the shaft. If the vertical and horizontal amplitudes(values) are nearly equal, there is very little, or no imbalance. Imbalance almost always exhibits itself in the horizontal plane. In fact, the difference between the vertical and horizontal readings can be a measure of the amount, or severity, of rotor imbalance.

Axial vibration levels should be less than half of radial(vertical and horizontal) levels. If axials are high, suspect shaft misalignment. With detailed analysis of the vibration spectrums, a trained technician could tell you whether the misalignment is radial, or angular.   Some vibration spectrums can tell you if there is looseness, v-belt wear, bearing wear, twisting of the baseplate, or resonance.

Most good analyzers now have a measurement for bearing condition. It is often called an envelope spectrum that is a snapshot of the high bearing frequencies that is extrapolated, "computerized" with an FFT calculation, and then given a severity value. With my software, 0 - .05 g's/sec is considered measurement of a new bearing, .10 - .30 g's/sec. is a good running bearing, and .75 g's/sec is considered a worn, or even bad, bearing.

However, true vibration analysis must incorporate trending. Measure the vibration now, then in a week from now, and then maybe in a month from now, to see if the vibration is changing in amplitude or frequency. A vibration amplitude that increases from one day to the next, confirms the problem is getting worse. By the same token, if you measure a relatively high reading, and it doesn't change over the course of a week, month, or even a year, this could be "normal" for a particular point on a machine.

Likewise with a bearing condition reading. I could measure .50 g's/sec on a bearing and think it may be going bad, but after 6 months it doesn't change, than I may be picking up other high frequency noise. Maybe from other machines in the room, or the floor above or below, or even air noise through the butterfly valve. I've opened the inlet b-fly valve on an air blower and "improved" the bearing condition.

Vibration analysis can be considered expensive, and not everyone can be trained to understand spectrum analysis. But as a tool in your own PM program, it will prove itself and save you many $$ in repair costs. Invest in PM, it will save you money!