Nondestructive Testing of Train Rolling Bearings Using Improved Teager Energy Operator and Minimum Entropy Deconvolution

Abstract

In order to solve the problem that signal-to-noise ratio of vibration signals for early failure of rolling bearings in trains is low and fault features are difficult to extract. A new method for feature extraction combined with minimum entropy deconvolution (MED) and an improved Teager Energy Operator (TEO) was proposed to detect rolling bearing failure. In view of the excellent performance of MED in extracting the impact of the signal, MED was used on the noisy bearing vibration signal to reduce the noise interference and enhance the impact component in the signal, and then the improved TEO was used to demodulate the noise-reduced signal and extract the instantaneous impact components. The Fourier transform was performed on the demodulated Teager energy signal to obtain the Teager energy spectrum of the signal. The fault condition could be diagnosed by analyzing the main frequency components in the Teager energy spectrum. The proposed method was applied to the rolling bearing simulation data and the fault diagnosis examples of outer rings and rolling elements of train rolling bearings. The experimental results demonstrated that the proposed method can effectively reduce the noise of the signal and enhance the impact component of the signal to effectively diagnose the rolling bearing faults of the train, and have a certain application value.