Bearing Faults Classification Based on Variational Mode Decomposition and Artificial Neural Network

Abstract

Bearing fault is the most causes of machine breakdowns. Consequently, the monitoring of this component is a key point to increase the reliability, security and avoiding serious damage in machine. Vibration signal is widely used for diagnosis which is considered as a powerful tool for detecting mechanical defects. In this paper, a rolling bearing fault-diagnosis method based on variational mode decomposition (VMD) and artificial neural network (ANN) is proposed. First, the processing methodology of bearing diagnosis starts with the decomposition of the vibration signal by VMD technique into a set of intrinsic mode functions (IMFs). According to the aim of fault diagnosis, the selected fault indicator is calculated from the energy related to the most sensitive IMFs to the bearing defect. Second, the extracted feature is then used as input to the ANN. the proposed approach is then validated using data from the bearing data center of Case Western Reserve University. The results prove the efficient of this method which is able to discriminating from four conditions of rolling bearing, namely, normal bearing and three different types of defected bearings: outer race, inner race, and ball.