Crack detection in rotating shafts using combined wavelet analysis

There are many rotating mechanical systems on board ships that comprise the main shaft transmission element. Proper functioning of shafts depends on many factors: alignment system, connections with other elements - couplings, gearboxes, bearing elements with ball bearings or bushings, operated regime, and al. All these factors can lead to fatigue of the shaft over time and to appearance of internal/external cracks which can determine a total damage of the shaft at some point in time. To avoid these disasters the operation of these rotating systems is monitored based on vibration analysis of recorded signals (FFT and DFT analysis). Due to the transient modes and the presence of the noise, the processing of the recorded signals must be performed with methods that present a good resolution simultaneously in the time and frequency domain. In the paper, measurements are made on a stand with a rotating system and it is determined the influence of some factors (measuring points and directions, speed, operated mode, crack size, presence of misalignment, noise) on the quality of the Wavelet Transform results. In order to improve the clarity of amplitudes and frequencies changes at the initiation of shaft cracks, the authors are proposing the combined use of Continuous Wavelet Transform (CWT), Discrete Wavelet Transform (DWT) and Wavelet Packet Transform (WPT). The authors propose using power spectral density (PSD) as it is a method with good performance in detecting faults in rotating machines.