A Rolling Bearing Fault Diagnosis Method Based on Rational Spline-LOD and Weighted Derivative Dynamic Time Warping

Due to the susceptibility of rolling bearing failures in wind turbines to noise interference and permissible installation and manufacturing tolerances, this paper proposes a novel rolling bearing fault diagnosis method. This approach combines rational spline local oscillation decomposition (RS-LOD) with weighted derivative time windowing (WDDTW). In this method, firstly, the multi-component vibration signal is adaptively decomposed into several mono-oscillation components (MOCs) by RS-LOD, and then the appropriate MOC components are selected by the proposed frequency integral method, which can improve the signal-to-noise ratio (SNR). Secondly, the WDDTW algorithm is used to obtain the residual vector signal by regulating the selected MOC components, which can reduce the influence of the rolling bearing allowable manufacturing and installation errors. Finally, the residual vector signal is demodulated by Hilbert transform, and extracting the rolling bearing fault characteristics from the Hilbert envelope spectrum. Based on the experimental data in the Case Western Reserve University bearing data center, the feasibility of the proposed method is verified. The analysis results show that compared with the traditional diagnosis method, the proposed method can highlight the rolling bearing fault characteristics more clearly and accurately, which provides a new idea for rolling bearing fault diagnosis.