An improved EMD method based on the multi-objective optimization and its application to fault feature extraction of rolling bearing

An improved Empirical Mode Decomposition (EMD) method based on the multi-objective optimization is proposed in this paper and is applied to extract the fault feature of rolling bearing with inner and outer race fault. Firstly, a new rational spline interpolation which has a shape controlling parameter is introduced based on cubic spline interpolation. Secondly, the selection criterion concluding time domain objective and frequency objective are considered simultaneously rather than separately. Then, the Particle Swarm Optimization (PSO) is used to find out the optimal IMF and determine the optimal shape controlling parameter. The procedure of the improved EMD based on multi-objective optimization (MO-EMD) is described in this paper. The effectiveness of MO-EMD is validated by the simulation signal and the robustness of MO-EMD to noise is also investigated. At last, the MO-EMD is employed to process the vibration signals of rolling bearing with fault. By comparison with original EMD, EEMD and improved CEEMDAN, it can be found that the MO-EMD has better decomposition capability and can restrain the mode mixing apparently. Combined with envelope spectrum analysis, more fault feature information can be extracted by MO-EMD.