An Experimental Study on Empirical Wavelet Transform

Abstract

Rotating machinery is one of the key components in mechanical equipment. Once the rotating machinery failed, it would lead to economic losses, or even a serious safety accident. Therefore, rotating machinery fault diagnosis has attracted a great deal of attention. Since the fault vibration signal of the rotating machinery is a mixture of different frequency components, signal decomposition methods have become powerful tools to extract the local information of the fault signal. Empirical wavelet transform (EWT) is an effective and self-adaptive signal decomposition method which has been widely studied. However, seldom of them attempt to conduct a systematic study on the performance of the EWT in terms of signal decomposition. In this paper, we perform the systematic simulations on how the retrieval performance of EWT is affected by the two-dimensional parameters, amplitude ratio and frequency ratio. By comparing the decomposition components with the original ones, we obtain the results which could guide people to evaluate the EWT method. Finally, the applicable region of the EWT in the fault diagnosis on rotating machinery are discussed.