Intelligent monitoring method of bearings based on PCA-SOM

To address the issues of excessive maintenance and untimely maintenance of bearings, this paper proposes a performance evaluation method for bearing condition monitoring based on the combination of Principal Component Analysis (PCA) and Self-Organizing Map (SOM) networks. This method utilizes PCA to reduce the dimensionality of dual-domain features and conducts clustering analysis using the SOM network, constructing a minimal quantization error as a degradation indicator to assess the bearing degradation state. This is an unsupervised approach that does not require the setting of training labels for the network. Experiments were conducted on the IMS dataset, and the results indicate that this method can effectively identify early bearing faults and reduce reliance on human expertise, thereby enhancing both overall efficiency and the intelligence level of assessments, ultimately achieving intelligent bearing fault recognition.