Rotordynamic Faults: Recent Advances in Diagnosis and Prognosis

Table 1

Rotordynamic “fault table.”

Faults	Sensors	Fault identification techniques	Localisation of fault	Prognosis	Work of note in fault modelling
Unbalance	Accelerometer Velocity transducer Proximity transducer	Operating deflection shapes; Time-frequency analysis and many others	Neural-network modelling and time-frequency analysis	FCM-Markov model	Unbalance in full-engine FEA model,; Early detection in aeroengines
High radial loads and misalignment	Accelerometer Thermocouple Displacement transducer; Angular sensor	Bearing temperature observation advanced Software-driven filtering methods	Various vibration analysis techniques	Prognostic modelling using simulated vibration data	Modelling the relationship between misalignment and other faults; Harmonic balance method
Rub and looseness	Accelerometer Velocity transducer Optical sensor	Single value decomposition Time-frequency analysis	Least-squares fitting approach	Adaptive wavelet denoising for prognostics	Rob and looseness using FEA models Whole engine modelling techniques
Fluid-induced instability	Proximity probes Keyphase transducer	Full spectrum plots Large amplitude subsynchronous vibration; Detection	Blade tip clearance and loading distribution	Cross-coupled stiffness and direct-damping studies	Nonlinear modelling accuracy of model based; Diagnostics numerical modelling
Bearing failure	Accelerometer Velocity transducer Displacement transducer; Proximity probes	Postfailure analysis for future failure diagnosis Dynamic time-frequency analysis	Statistical change in the bispectral domain	Numerical modelling of aeroengine bearings Model-based prediction of remaining useful life On-board prognostics	High and low fidelity dynamic system modelling; Comparisons of algorithms to understand bearing vibration phenomena
Shaft cracks	Keyphase transducer Proximity probes	Higher order spectra Neural networks; Grid support vectors	Forced response measurements	Nonintrusive torsion vibration monitoring	Statistical analysis FEA modelling
Blade cracks	Keyphase transducer Proximity probes	Hilbert-Huang transformation Model-based diagnostics; Morlet wavelet/quantum; techniques	Novel bearing structures	Active magnetic actuator	Global and local asymmetry modelling Finite element modelling
Rotor Bow	Keyphase transducer Proximity probes Seismic transducer	Examination of “slow roll” data; Model-based diagnostics	Partial mathematical modelling	Statistical symptoms analysis	Advantages of fault modelling for identification; Nonlinear modelling techniques