A novel amplitude-independent crack identification method for rotating shaft

• Published in: Proceedings of the Institution of Mechanical Engineers. Part C, Journal of mechanical engineering science Journal of mechanical engineering science, 2018-11, Vol.232 (22), p.4098-4112
• Main Authors: Yang, Laihao, Chen, Xuefeng, Wang, Shibin
• Format: Article
• Language: English
• Subjects: Amplitudes; Catastrophic failure analysis; Dependence; Feature extraction; Identification; Identification methods; Malfunctions; Matching; Rotating machinery; Rotating machines; Rotating shafts; Rotation; Vibration analysis
• ISSN: 0954-4062; 2041-2983
• DOI: 10.1177/0954406217748686

The shaft crack is one of the most common and serious malfunctions in rotating machines and may lead to catastrophic failure if undetected in time. However, the conventional crack identification methods are amplitude-dependent and thus can be only applied to the crack identification under some specific conditions. In this paper, a novel amplitude-independent crack identification method (AiCIM) is significantly proposed to eliminate the amplitude-dependent property and promote the effectiveness of the crack identification. First and foremost, a fast time-varying vibration phenomenon of the cracked-rotor system is newly found. Through the theoretical analysis, the fast time-varying vibration mechanism of the cracked-rotor system is revealed for the first time. It is indicated that the vibration signal of the cracked-rotor system is modulated by the fast-oscillated instantaneous frequency, which is independent of the amplitude of the vibration signal. AiCIM is then put forward on the basis of the fast time-varying vibration mechanism and matching time–frequency analysis theory. Specially, the amplitude-independent instantaneous frequency of the vibration signal is extracted via the matching time–frequency analysis theory, and the time–frequency representation energy-concentration is enhanced along the instantaneous frequency trajectory. Since instantaneous frequency of the vibration signal carrying the critical fault information is employed to identify the shaft crack, AiCIM is only relevant to the phase of the vibration signal, i.e. amplitude independent. As a result, AiCIM successfully eliminates the dependence on the signal amplitude and is more sensitive to the weak crack. Both the numerical and experimental results demonstrate that AiCIM behaves best to extract the fast-oscillated feature of the fast time-varying vibration induced by the shaft crack in comparison with other time–frequency analysis methods, and AiCIM effectively suppress the effect of noises on the instantaneous frequency estimation because of its amplitude-independent property. Influences of the crack parameters on the nonlinear instantaneous frequency are finally discussed with AiCIM. This study provides a potential way to the online crack identification.