Abstract:Envelope order analysis is one of the vibration analysis methods used for the gearbox vibration analysis during the varying speed period of rotating machinery. However, its performance may be severely deteriorated due to the existence of disturbances. The Independent Component Analysis (ICA) is a recently developed method for sources separation from mixed signals. In ICA, it is usually assumed that the number of sensor channels should be not less than the number of independence vibration sources. However, it is difficult to prior determine the number of independent vibration sources. Then, the applications of ICA may be limited in practice. An alternative multiple vibration sources separation approach based on the ICA and the envelope order analysis is proposed in the paper. The envelope extraction is first utilized to reduce the dimension of vibration source in original data sets, and then the even-angle sample of order track is applied for the envelope waveform. Subsequently, the ICA is employed to separate the even-angle sample signals according to the independent of vibration sources, which allows obtaining the fault features. Simulations and experimental tests are included to verify the validity of the method.