PolSOM based approach for bearing fault diagnosis

This article presents a polar self-organizing map (PolSOM) based approach for bearing faults classification. PolSOM is a new type of SOM. It is trained to produce a low-dimensional representation of high-dimensional data by using unsupervised learning. In this article, high dimensional feature data based on vibration signal are calculated to represent the different health conditions of bearings, and then PolSOM is introduced to visualize the data for fault classification. Synthetic data and bearing data are employed to test the proposed method. Results show that our proposed approach has good performance in bearing fault diagnosis.