What specific type of analysis is used to diagnose rolling element bearing failures in large turbine-generator sets based on captured vibration signatures?

The specific type of analysis used to diagnose rolling element bearing failures in large turbine-generator sets based on captured vibration signatures is vibration spectrum analysis, often employing Fast Fourier Transform (FFT). Rolling element bearings, such as ball or roller bearings, exhibit characteristic vibration frequencies when they develop faults. These frequencies are related to the bearing's geometry (number of balls/rollers, pitch diameter, contact angle) and the rotational speed of the shaft. Vibration spectrum analysis uses FFT to transform the time-domain vibration signal (amplitude versus time) into a frequency-domain spectrum (amplitude versus frequency). This spectrum reveals the presence of discrete frequencies corresponding to specific bearing defects, such as ball defects, roller defects, inner race defects, and outer race defects. By comparing the frequencies observed in the spectrum to the calculated theoretical defect frequencies for the bearing, analysts can identify the type and location of the fault. Furthermore, the amplitude of the defect frequencies can indicate the severity of the damage. This type of analysis allows for early detection of bearing failures, enabling timely maintenance and preventing catastrophic equipment damage.