The techniques for determining bearing damage are based on the detection of strong impacts or pulses happening during the shaft rotation. We will distinguish two major groups as follows.

Techniques based on the analysis of simplified parameters

• Overall RMS or peak velocity vibration value: usually ISO 10816 is applied.
• Crest factor: is the ratio between the peak value and the RMS value.
• Kurtosis: is a statistically calculated parameter from the time waveform that informs of the amount of peaks present in the waveform.
• Skewness: is also a statistical parameter calculated from the time waveform, that gives information on how the signal is symmetrical with respect to its mean value.
• Variable high frequency band: typically defined between 1 kHz and 20 kHz.
• HFD: high frequency band between 5 and 30 kHz in acceleration.
• Spike-energy: high frequency band between 5 and 60 kHz.
• SPM or Shock pulses: measures high frequency vibration between 30 and 40 kHz.
• Acoustic crack detection / IFD: band between 80 and 120 kHz.
• Acoustic emission: band with a range higher than 200 kHz.

The main limitation of these techniques is their inability to discern between bearing problems and other machine problems that also generate high frequency vibration such as: gear issues, cavitation, electrical frequencies, etc. Other limitation, no less important, is the limited frequency response ranges of the sensors they use and the mounting method used to attach them to the machine.

Techniques based on spectral and time waveform analysis

Spectral and time waveform analysis is decisive in diagnosing the damage of a rolling element bearing, allowing us, by comparison of the evolution of the spectral amplitudes, to predict the future degree of deterioration and plan its replacement well in advance. In this chapter we will go deeper into the use of these techniques.