Detection and diagnosis of changes in the time-scale eigenstructure for vibrating systems

Focuses on techniques used in monitoring machine condition and diagnosing mechanical faults in vibrating mechanical equipment which has varied operational modes and whose dynamic signals are nonstationary. Because of the nonstationary nature of the vibrating system, one has to apply the time-scale transform to capture the nonstationary modes. The autoregressive moving average modeling captures the modal signatures, in particular, the autoregressive components (natural frequencies and damping in various modes) of the vibrating system. We consider the problem of detecting and diagnosing changes in the eigenstructure of the state transition matrix of the state-space model. A chi-square test is performed at different level of resolution in the time-scale space. The dynamic system that is considered in this research is a vibrating gear system.