Time frequency domain distribution and its application to reciprocating machinery analysis

Abstract

Accurate assessment of shipboard machinery condition is essential in increasing the operational capability of naval vessels. Current shipboard machinery vibration monitoring and diagnostic procedures use frequency domain spectra to identify possible faults and problem areas. This method has proven to be inadequate for reciprocating machinery analysis. In reciprocating machinery the vibration signal is no longer a stationary, ergodic process, but is time-dependent and in some cases transient. This study proposes the use of Pseudo Wigner-Ville Distribution and introduces Wavelet analysis as two advanced methods for condition monitoring of non-stationary and transient shipboard machinery. These methods employ a time-frequency domain distribution for the detection of fault location and severity level. To demonstrate the benefits of a time-frequency representation, vibration data from two types of reciprocating air compressors will be processed for analysis. To simulate faulty conditions, constructed artificial fault signals will be introduced into the vibration data and analyzed. It is proposed Wavelet Analysis will play a complementary role to the Pseudo Wigner-Ville Distribution technique.