Condition Monitoring of Axle Box Bearing Based on Improved Safety Region

To improve the operating reliability of the axle box bearing of high-speed trains, the safety region theory was introduced into condition monitoring of the axle box bearing. The traditional assessment of the safety region was transformed into determining the boundary values, aiming to avoid the influence of complex model parameters on the evaluation process of the safety region. The normalized energies of intrinsic mode functions were used to construct the eigenvector of bearing operating state, and the correlation function was used to establish the evaluation model of the safety region boundary values, where the particle swarm optimization algorithm was adopted to get optimal solution. On the basis of estimation results of boundary values, correlation function was utilized to assess the operating state of bearing quantitatively. The effectiveness was verified by the fatigue test of rolling bearing, and the method was used for condition monitoring of the axle box bearing. The results show that the detection rate and classification rate of bearing operating state of fatigue test are 0.951 and 0.939, respectively; the classification rate of operating state of the axle box bearing is 0.935, indicating that the axle box bearing is running normally, which is consistent with its actual operating state.