Bearing Fault Feature Enhancement and Diagnosis Based on Savitzky–Golay Filtering Gramian Angular Field

In actual engineering production, bearings typically operate in harsh environments. The fault features of bearing vibration signals are often submerged by background noise, making it difficult to extract the fault signal features and impacting the accuracy of fault diagnosis. To address this problem, this paper proposes a bearing fault diagnosis method based on the Savitzky-Golay Gramian Angular Field (GAF) with fault feature enhancement combined with ResNet. First, the acquired vibration signals are segmented, and the segmented signals are subjected to Butterworth high-pass filtering to obtain the high-frequency components of the signals that contain fault information. Secondly, the extracted high-frequency components are boosted by the S-enhancement algorithm for fault features. The boosted signals are then filtered by Savitzky-Golay to achieve data smoothing aggregation enhancement. Subsequently, the feature-enhanced GAF graphs are obtained using the transformation method. Finally, bearing fault diagnosis is performed using the Glamian Angle field diagram as input to the ResNet18 model. To verify the feasibility of the proposed method, experiments were conducted using Case Western Reserve University (CWRU) bearing fault dataset and bearing fault dataset of laboratory experimental platform. The experimental results showed that the fault diagnosis accuracy were 99.28% and 100%, respectively. The results validated the feasibility of the proposed method. Through comparative experiments with the Symmetric Dot Pattern (SDP) method, the traditional GAF method and the Recurrence Plots (RP) method, the results demonstrate that the proposed method has high diagnostic accuracy, proved the effectiveness of the method.