Fault Diagnosis of Rolling Bearings Based on Acoustics and Vibration Engineering

Currently, the fault diagnosis and maintenance of rolling bearings have become an urgent problem to be solved. A fault diagnosis method based on feature extraction and word bag model was designed based on the theories of acoustics and vibration engineering science. At the same time, the traditional word bag model was optimized, and a rolling bearing fault diagnosis method based on the adaptive extended word bag model was designed. This method mainly expands the word bag model into a 3-layer structure and constructs codebooks for the feature vectors of each layer. The results indicate that the fault diagnosis method for rolling bearings designed in the study has high diagnostic accuracy and stability, providing reliable technical support for the normal operation and safe maintenance of mechanical equipment. © 2013 IEEE.