Bearing state characterization method based on composite multi-scale generalized weighted permutation entropy

The invention belongs to the technical field of fault diagnosis, and particularly relates to a bearing state characterization method based on composite multi-scale generalized weighted permutation entropy, which comprises the following steps: acquiring the composite multi-scale generalized weighted permutation entropy; obtaining composite multi-scale generalized weighted permutation entropies of the bearing in different states, and forming a high-dimensional vector; and bearing state characterization is carried out through separation of the obtained main components. Compared with other improved permutation entropy values, the composite multi-scale generalized weighted permutation entropy disclosed by the invention expands the selection range of the weighted permutation entropy in the weighted value calculation process, and has important significance for improving the fault characterization capability of the rolling bearing and monitoring the running state. According to the technical scheme, starting from the