On-Line Health Monitoring Index of Rotating Machinery

A wavelet packet method is utilized to decompose vibration signal collected from gearbox online monitoring systems.Signal reconstruction is implemented based on optimal selection of selected frequency domain. A health monitoring index is proposed,which can reflect health condition of gearbox during whole life cycle.In order to realize health warning,a dynamic threshold is proposed to detect the existence of early gear failure.Three sets of whole lifetime vibration data collected from gearboxes are used to validate this method.The analysis results illustrate that the proposed method has good performance in detecting early gear failure and owns great potential in engineering applications.