Abstract
This paper presents the recent application of the scalogram of Continuous Wavelet Transform (CWT) as a vibration monitoring and signal processing tool for a rotor dynamic response under parametric excitation. The experimental test data of coupled lateral-torsional vibrations of a rotor-stator system with transverse crack was obtained through a data acquisition set-up interfaced with Rotor-Kit-4 (RK-4). Analysis was executed on rotor deflection, orbit, frequency and time-frequency spectrum of the RK-4 experimental data. The scalograms of CWT were used experimentally to represent the aperiodic occurrence of rub between the rotor-stator and crack features. Variation in 3-D scalogram peaks in the presence of rub and crack were unique and were used to distinguish quasi-periodic motion from other types of motion. An unbalanced cracked rotor gave a higher frequency amplitude response compared to an unbalanced rotor with rub under the same conditions. Irregularities in orbit orientation near sub-harmonic resonances were observed in the test data. Multiple rebounds inside the orbit loop were unique rub indicators. Conspicuous horizontal components of the higher harmonics were observed near the critical speed when a crack existed. CWT established inherent feature patterns that discriminated unbalance, rub and a crack.
Highlights
Discretisation of multiple faults in rotating machinery is a common problem in signal analysis
An experimental study of coupled lateral-torsional vibrations of a RK-4 system excited by unbalance, rub and a transverse breathing crack force has been presented
Analysis of the peaks of frequency appearance in 2-D and 3-D wavelet scalogram indicated the severity of the vibration in the vicinity of the critical speed
Summary
Discretisation of multiple faults in rotating machinery is a common problem in signal analysis. It is imperative to diagnose and discretise the nonlinear features of the unbalance, rub and crack interaction in rotating systems In this regard, several researchers have investigated the application of sophisticated techniques in order to extract the specific multiple fault features of transient signals [2,3,4,5]. In order to address the aforementioned multiple faults in failure detection, this work focuses on the features extraction of coexisting nonlinear and transient faults, such as unbalance, crack and rotor-stator rub in a high-speed rotor system using the time-domain response in a lateral direction, its orbit and both its FFT spectral and its WT techniques.
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
