Rolling bearing fault diagnosis based on EEMD and sparse decomposition

Abstract

Rolling bearings, as the core components of wind turbine, are prone to failure due to the influence of complex working condition and harsh environment. However, the bearing defect-induced impulse features are always submerged by strong noise and harmonic interference, thus increasing the difficulty in detecting rolling bearing fault. Therefore, aiming at this problem, a new fault diagnosis method based on ensemble empirical mode decomposition (EEMD) and sparse decomposition theory is proposed to improve the performance of rolling bearing fault diagnosis. First, the EEMD method is applied to adaptively decompose the signal of rolling bearing into multiple intrinsic mode functions (IMFs) components; second, calculate the Hurst exponent of each IMF component to eliminate the harmonic components; and then, use the residual IMFs to reconstruct the signal which is used as the input of sparse decomposition method, and the orthogonal matching pursuit (OMP) algorithm is adopted to extract the impulse components from the constructed signal; finally, through the envelope demodulation analysis based on Teager energy operator, we can achieve the accurate diagnosis of rolling bearing fault. Simulation and engineering application are performed to verify the effectiveness and superiority of the proposed method.