An intelligent fault identification method of rolling bearings based on LSSVM optimized by improved PSO

This paper presents an intelligent fault identification method of rolling bearings based on least squares support vector machine optimized by improved particle swarm optimization (IPSO-LSSVM). The method adopts a modified PSO algorithm to optimize the parameters of LSSVM, and then the optimized model could be established to identify the different fault patterns of rolling bearings. Firstly, original fault vibration signals are decomposed into some stationary intrinsic mode functions (IMFs) by empirical mode decomposition (EMD) method and the energy feature indexes extraction based on IMF energy entropy is analyzed in detail. Secondly, the extracted energy indexes serve as the fault feature vectors to be input to the IPSO-LSSVM classifier for identifying different fault patterns. Finally, a case study on rolling bearing fault identification demonstrates that the method can effectively enhance identification accuracy and convergence rate.

► An intelligent diagnosis method for bearing faults is proposed based on least squares support vector machine optimized by improved particle swarm optimization algorithm.
► Original fault signals are decomposed into some stationary intrinsic mode functions (IMFs); and energy feature indexes are established based on IMF energy entropies.
► Hyperparameters of LSSVM are optimized with IPSO algorithm.
► Fault patterns are identified sequentially and automatically.