Research on Rolling Bearing Fault Diagnosis Based on Continuous Wavelet Transform and Convolutional Neural Network

Fault diagnosis of rolling bearings is crucial for ensuring the safe operation of mechanical equipment. Traditional methods rely on manual feature extraction, such as spectrum analysis and wavelet packet decomposition, but they suffer from low efficiency and poor generalization under complex working conditions [1-3]. With the development of deep learning technology, convolutional neural networks (CNNs) have been widely applied due to their powerful ability for automatic feature extraction. This paper proposes an intelligent diagnosis framework that integrates continuous wavelet transform (CWT) and convolutional neural network (CNN). By extracting time-frequency features of vibration signals using CWT and generating high-resolution time-frequency maps as inputs to the CNN, end-to-end fault classification is achieved. Experiments were conducted using the Case Western Reserve University bearing dataset, covering four fault categories: normal, inner race, rolling element, and outer race. The results show that the proposed method achieved 100% classification accuracy on the test set. TSNE visualization revealed the hierarchical nature of CNN feature learning. This study provides theoretical support for the intelligent operation and maintenance of industrial equipment.