Rotordynamic Fault Diagnostics Using Phase Coherent Filtering

Fault diagnostics of rotordynamic vibro-acoustic signals require extracting diagnostic features from the vibration signals to determine the operational condition of vital components within a gearbox. However, fault diagnostics’ efficacy is reduced by the poor signal-to-noise ratio (SNR) in the vibration signal, which impairs the extraction of the diagnostic features from the vibration signal. A method used successfully in improving SNR is classical adaptive noise cancellation (CANC), which requires a primary signal and a reference signal. This reference signal is comprised of either a noise signal or a diagnostic signal that is correlated with the primary signal. However, under most fielded equipment operational conditions a reference signal is either not available or is time/state variant. Therefore, a Phase Coherent Filtering (PCF) method was developed to extract the correlated noise reference signal from the primary signal. This PCF result can then be employed by CANC to least square estimate the diagnostic signal portion of the primary signal, which dramatically improves the SNR as well as the extraction of diagnostic features, thereby verifying PCF’s ability to optimally extract either the diagnostic signal or reference signal noise from the primary signal.