Fault diagnosis of satellite flywheel bearing based on convolutional neural network

The bearing is one of the core components of the flywheel, providing a stable slewing support for the flywheel, and its operating state often directly affects the flywheel and even the entire spacecraft's normal operation. In view of the problem of automatic and accurate identification of the bearing faults, this paper uses convolutional neural network (CNN) to develop a satellite flywheel bearing fault intelligent diagnosis method. First, the vibration signal characteristics of satellite flywheel bearing under different faults are studied. Second, the time-domain signal graphs are constructed by combining vibration signals under multiple rotational speeds and used as feature input maps. Finally, the bearing fault intelligent diagnosis method is presented based on the excellent image recognition characteristics of CNN and the constructed feature maps. The experimental verification shows that the proposed method can achieve better diagnostic results.