A new methodology for fault detection in rolling element bearings using singular spectrum analysis

In this study, a new methodology for fault detection in rolling element bearings is proposed, which is based on singular spectrum analysis (SSA). The main idea of the methodology is to build a baseline space from the feature vectors corresponding to the healthy bearing condition. This baseline space is made from the directions of the first three principal components, which are obtained from the decomposition stage of the singular spectrum analysis. Then, the lagged version of any new signal corresponding to a measured (possibly damaged) condition is projected onto this baseline space in order to assess its similarity to the baseline condition. The Euclidean norms of these projections are used to form three-dimensional feature vectors. The category of a new signal vector is determined on the basis of the Mahalanobis distance (MD) of its feature vector to the baseline ones. The methodology is validated using datasets acquired from two different test-rigs. From the results obtained for the correct classification rate, it is shown that this methodology performs very well. The suggested methodology also has simple steps and is easy to apply.