2025 International Conference on Equipment Intelligent Operation and Maintenance (ICEIOM2025)

26 / 2025-05-08 15:05:37

An Intelligent Detection Method for Equipment Based on SVDD

rolling bearing,fault diagnosis,feature enhancement,anomaly detection,intelligent detection

Special Sessions > SSL 05 Condition monitoring, fault diagnosis and intelligent maintenance of mechanical equipment

To address the spectral discrepancies in vibration signals caused by rolling bearing faults, an intelligent detection method based on SVDD is proposed. The method utilizes spectrum data from normal operating conditions to train a high-dimensional hypersphere model, determining the radius and center. Anomalies are identified by computing the distance between test spectrum and the hypersphere center. For the test spectrum classified as anomalous, a dimension-wise contribution analysis in the high-dimensional space is performed to adaptively generate a weighting vector, enhancing fault-related frequency components while suppressing normal vibrations and noise. The proposed method requires only healthy state data for both anomaly detection and feature enhancement, and demonstrates effective diagnostic performance and strong application potential in both simulation signals and experimental gearbox bearing signals.