Quantizing the Health State of Railway Axle Bearing via Signal-to-noise ratio defined by EEMD and SVD

Abstract

State quantized characterization is of great demand with the development of automation and intelligent machineries. It is helpful for monitoring the operation state for machines in real time, and is critical for detecting the defects and potential faults. State characterization is also a base rock to implement a state-based maintenance strategy for avoiding the unexpected accidents and significant economic losses for multifarious machines. Axle bearing is one of the most important component in vehicle, and its health state will affect the stability and safety of the whole railway system. In this paper, a promising state quantized characterization method based on signal-to-noise ratio is proposed to monitor and detect the railway axle bearing in a high-speed train. The novelty signal-to-noise ratio is defined via ensemble empirical mode decomposition and singular value decomposition, and a dimensionless health indicator based on the signal-to-noise ratio is proposed to characterize quantitatively the health state of the axle bearing. Experimental results showed that the proposed health indicator is effective in distinguishing three types of railway axle bearing defects. Moreover, guideline thresholds for classification of normal and abnormal bearing health statuses and metrological boundaries of the health indicator for identification of three different types of railway axle bearing defects are provided.