Mechanical Fault Diagnosis Using Numerical Hilbert Analysis

Dynamic instability induced by the initiation and development of mechanical faults in a rotary element is known to have a large negative impact on the reliability and operation safety of an entire system. This type of nonlinear system response is generally perturbed by shock impulses of extremely short time scale and amplitude. Thus difficulty presents itself in identifying and isolating features indicative of the presence and progression of faults possibly leading to mechanical deterioration. The perturbed and deteriorated states of a bearing-shaft system subjected to the actions of various types of commonly seen mechanical faults are investigated using the Numerical Hilbert Transform. The presented approach characterizes and realizes temporal events of both short and long time scales as instantaneous frequencies in the joint time-frequency domain. Examples are given to demonstrate the feasibility of applying the approach to the characterization of various deteriorating bearing states and the identification of parameters associated with several failure modes.