Monitoring Point Configuration of Gearbox Based on EEMD and VPRS

Abstract

To gain the sensitive fault information, proper sampling point configuration is essential and important. A method based on ensemble empirical mode decomposition and variable precision rough set is proposed to optimize the monitoring points in the gearbox fault diagnosis system. First, the vibration signal was processed by ensemble empirical mode decomposition, the energetic characteristic vector can be got by the intrinsic mode functions. Samples in different conditions constructed the decision table, each monitoring point corresponded to a decision table. Second, the approximate dependence values of condition attributes to decision attributes of every monitoring point were got by variable precision rough set. Finally, the importance of the sampling points was achieved by the approximate dependence value. The experimental results show the method is effective and feasible for the monitoring points configuration, and it can minimize the impact of noise as well as improve the efficiency and accuracy of the fault diagnosis system.