A Deep Fuzzy Clustering Network for Gearbox Health Assessment

T Conducting gearbox health status evaluation can effectively obtain the operating status of the equipment so as to develop maintenance plans. A gearbox health condition evaluation method based on deep fuzzy clustering neural networks (DFCNN) is proposed. The proposed DFCNN incorporates the high-level feature extraction layer of Generalized Supervised Deep Autoencoders (GSDAE) into an improved fuzzy C-Means (KMDFCM) clustering algorithm. GSDAE is used to eliminate the variation of feature parameters due to differences in operating conditions, and then KMDFCM is used for unsupervised condition assessment. Taking a planetary gearbox as an example, it is shown that the proposed method provides a new idea for solving the health condition assessment of gearboxes.