Application of Energy Projection Method Based on LMD in Fault Feature Extraction of Bearing Type

Aiming at the non-stationary and multi-scale characteristics of the bearing fault signal, a method of extracting the fault feature of bearing based on Local Mean Decomposition (LMD) energy projection is proposed. Firstly, the LMD method is utilized to decompose the bearing fault signal, and the feature frequency components of the signal are separated at the same time. Secondly, the decomposition of the main product function component weighted optimization and eliminate false composition, feature reconstruction PF components and to enhance the signal characteristic information. At last, the energy of the acquired PF is projected to a special frequency area, based on the energy distribution of structure bearing so avoidance signal feature vector to obtain low dimension feature vector. Experimental results show that the proposed approach is feasible and effective.