Intelligent fault diagnosis method for bearings based on WSVM-CNN

Monitoring rolling bearings is crucial for ensuring the safe and efficient operation of mechanical equipment. Data-driven fault diagnosis algorithms can effectively identify faults, but the lack of industrial fault data poses challenges for modelling. This paper proposes a small-sample fault diagnosis method by integrating a weighted support vector machine (WSVM) into a convolutional neural network (CNN)-based bearing fault diagnosis algorithm, introducing a new fault diagnosis method called WSVM-CNN. First, after filtering out faulty samples and ambiguous samples, pre-classification of fault states is performed; then, classification of faulty samples is carried out. The proposed method was tested on a public dataset and achieved effective accuracy, showing significant improvements in diagnosis speed and accuracy compared to single algorithms.