Early fault feature extraction method of vibration signal of mine-used motor

In view of problems of parameter setting, frequency aliasing and signal distortion existing in current fault feature extraction methods of vibration signal of mine-used motor, an early fault feature extraction method of vibration signal of mine-used motor based on dual-tree complex wavelet transform was proposed. Firstly, collected vibration signal of mine-used motor is decomposed by using dual-tree complex wavelet transform, so as to obtain dual-tree complex wavelet coefficients of each layer. Then soft threshold filtering is used to filter the dual-tree complex wavelet coefficients of each layer. At last, denoising signal is obtained by reconstruction of the filtered dual-tree complex wavelet coefficients. The application results show that the method can effectively remove noise in the motor vibration signal, and extracted early fault feature can reflect actual operating condition of motor, which provides an effective basis for early fault diagnosis of motor.