The mixed fault detection method for rolling bearings based on ICA-Teager

The vibration signal of rolling bearings is always the mixed fault signal which is interfered by noise or mixed aliasing by different fault source, and analysis and fault feature extraction of the signal is difficult problem. Therefore, a mixed fault detection method based independent component analysis and Teager Energy Operator (TEO) demodulation (ICA-Teager) is proposed. Firstly, make the vibration signal acquisition of the ICA processing, to extract the independent component ICs; secondly, TEO demodulate and achieve the amplitude envelope of the ICs using TEO; finally, implement Fourier transform of the amplitude envelope, and detect faults by matching the dominant frequency spectrum with the fault feature of theoretical calculation. The experimental results show the ICA-Teager can be isolated the mixed signal and extracted the dominant frequency characteristics of each IC to detect faults.