Application of MED-TET to Feature Extraction of Vibration Signals

Vibration signal reflects the operation status of the equipment. It is widely used in the field of mechanical fault diagnosis. However, the weak fault impact signal will be masked by the vibration noise, which makes it difficult to extract the fault features of the raw vibration signal. Aiming at this feature extraction problem, a hybrid method based on minimum entropy deconvolution (MED) and transient-extracting transform (TET) is proposed. First, the original signal is pre-processed by the MED method, which effectively reduces the interference of noise on the signal and enhances the impact component. Then, TET is used to extract the transient features of the pre-processed signal. Finally, the extracted transient information is used for fault diagnosis of rolling bearing. The validation of the method is carried out on simulated signals and Case Western Reserve University (CWRU) bearing data. Also, the proposed method is compared with other feature extraction methods. Those results show that the method can effectively extract the impact components in the vibration signal under strong background noise, which the effectiveness of the method is verified.