Second-order transient-extracting transform for the analysis of impulsive-like signals

Abstract In this paper, a high resolution time–frequency analysis method, termed as second-order transient-extracting transform (STET) is proposed in the analysis of highly non-stationary signals with strong impulse components for machine fault diagnosis. The limitation of two recently published post-processing techniques are discussed and highlighted based on a theoretical analysis of the first and the second-order frequency-varying models. A STET technique is then proposed to overcome this limitation. The discrete implementation of STET is also presented in the study and the performance of the technique is examined by a number of simulated numerical data and three sets of experimental data. The results confirm the effectiveness of the proposed technique in the analysis of noise contaminated signals as well as bearing defect signals acquired under varying speed condition.