Bearing Life Prediction Method Based on MFO-Grid-LSSVM

In order to improve the accuracy of least square support vector machine(LSSVM) for the prediction of bearing remaining life, an optimization method based on MFO-Grid-LSSVM is proposed. The method uses empirical mode decomposition (EMD) algorithm to decompose bearing data. IMF's strongly correlated component be reconstructed. Principal component analysis (PCA) was used to reduce the dimensionality of the reconstructed signal in time-frequency domain, to obtain degradation state of bearings during the full life cycle. Parameters of LSSVM model were optimized by combining the moth fighting with the grid algorithm, and the principal component satisfying conditions is used as the model input to predict the remaining life of the bearing. Using IMS bearing data to verify, simulation results show that the method has high prediction accuracy and smaller prediction error.