Research on Vibration Characteristics of Cylindrical Roller Bearings Based on Convolutional Neural Networks and C-Attention Network

Cylindrical roller bearings, as key components of industrial machinery, has vibration characteristics that significantly impact on the service life. This paper presents the vibration characteristics of cylindrical roller bearings operating under lubricated conditions. Firstly, several coordinate systems are established based on the motion characteristics of the bearing. Using the generalized Reynolds equation, a lubrication model is constructed to analyze the motion relationship of cylindrical bearing. The dynamics model of the roller is then established. On this basis, Ansys is used to calculate the equivalent force and temperature variations of the bearing under load, which helps evaluate its lubrication condition. Finally, the relevant parameters are defined based on the kinetic model, and the vibration signals of the bearing within 0.06s to 0.1s time frame are extracted for in-depth analysis of its vibration characteristics. This study provides a feasible approach to analyzing the roller bearings.