基于EWT-MDS的风力机轴承劣化趋势识别及故障诊断

In order to solve the low identification accuracy of deterioration trend of wind turbine bearings and difficult problem of early fault diagnosis, a EWT-MDS method based on empirical wavelet decomposition and multidimensional scaling transform is proposed. Firstly, the method performs the adaptive empirical wavelet decomposition of the whole life cycle vibration signals of the bearings. The variation characteristics of various modal components are quantitatively analyzed by using information entropy as an index; then the multi-dimensional scaling transform algorithm is used to obtain the synergistic variation rule of each degradation component in high-dimensional space, which has a large improvement in detection efficiency and precision compared with the conventional method. The simulation and experimental results show that the proposed method can detect the nodal points of abnormal state of bearings in advance, and can accurately identify the early fault type of bearings through the time domain reconstruction of modal components characterized by bearing degradation and combining the spectrum and the envelope spectrum.