Rolling bearings fault diagnosis based on recurrence complex network

In the case of the non-stationary and non-linear vibration signal of a rolling bearing with faults, a bearings fault diagnosis method based on the recurrence complex network (RCN) is put forward. First, a one dimension time series is extended to high dimension phase space by using the phase space reconstruction method, and a recurrence matrix is built. Then, recurrence quantification analysis (RQA) is discussed. Finally, the recurrence complex network (RCN) method is employed to extract nonlinear characteristic parameters of the vibration signals, which yield the feature vectors. The analysis results of the vibration signals acquired from the bearings with normal, outer track fault, ball fault and inner track fault, respectively, show that the RCN method has better diagnosis effect than the RQA method. ©, 2015, Nanjing University of Aeronautics an Astronautics. All right reserved.