摘要：In order to improve the reliability of fault diagnosis, total energy growth rate of rolling bearing during run-up is defined and applied to bearing detection. Firstly, an approach of merging short-time Fourier transform (STFT), linear fitting and median filtering is developed to extract the total energy growth rate. Secondly, the relationship between the total energy growth rate and different running conditions is discussed. Thirdly, the total energy growth rate is adopted to diagnose faults as input vector of radial basis function (RBF) neural network. Experiment results show that the total energy growth rate is an effective failure symbol for fault diagnosis of rolling bearing during run-up.
