Visualization of rolling bearing fault information based on self-organizing feature maps

The Self-Organizing Feature Maps(SOFM) is an efficient tool for visualization of multidimensional numerical data. It projects input space on prototypes of a low-dimensional regular grid. In this paper a new visualization technique has been validated against U-matrix method based on Euclidean distances between input vectors and neurons weights combined with the distribution of the fixed lattices in the network. SOFM, as an unsupervised neural networks, is applied to condition monitoring and fault detection of rolling bearings. By analyzing and processing of the vibration signals of the rolling bearing, the characteristic parameters which represent operating state of the rolling bearing are extracted to construct characteristic vector and used to train SOFM, The trained results can be visualized in the two-dimensional (2D) space intuitively. Experiment results show that the SFOM can be used to distinguish perceptually between the samples which are classified by fault type. Method proposed could clearly visualize the fault information.