A lightweighted planetary gearbox fault diagnosis method based on multisensor fusion

Most research methods of planetary gearbox fault diagnosis with multisensor have some shortcomings such as large parameters and unconducive for deployment. To address these issues, a lightweighted planetary gearbox fault diagnosis method based on multisensor fusion is proposed. Firstly, the collected planetary gearbox signals are preprocessed by tunable-Q wavelet transform (TQWT) and Gramian angular fields (GAF) image encoding method. Secondly, a multibranch feature fusion module (MBFM) is constructed for multisensor fusing, and a multiscale residual attention network (MRAN) model is built to extract fault features with multi-scales. The Mixture-of-Depths (MoD) is adopted to lightweight the model. The results show that the parameters and floating point operations (FLOPs) of the proposed fault diagnosis model are 3.5M and 452.6M, respectively. The diagnostic accuracy reaches 97.2%, and it has good anti-noise performance and stability.