Fault diagnosis of rolling bearings based on Marginal Fisher analysis

Feature extraction plays an important role in fault diagnosis. It is critical to extract the representative features for improving the classification performance. An intelligent fault diagnosis method based on Marginal Fisher analysis (MFA) is put forward and applied to rolling bearings. The high-dimensional features in time-domain, frequency-domain and wavelet-domain are extracted from the raw vibration signals to obtain rich faulty information. Subsequently, MFA excavates the underlying low-dimensional fault characteristics embedded in the high-dimensional feature space by preserving local manifold structure. Thus, the optimal low-dimensional features are obtained to characterize the various fault conditions of rolling bearings and finally fed into the simplest k-nearest neighbor classifier to recognize different fault categories. The diagnosis results validate the feasibility and effectiveness of the proposed fault diagnosis method, compared with the other three similar approaches.