基于关联维数与BP神经网络的滚动轴承故障诊断

Rolling Bearing Failure Diagnosis Based on Correlation Dimension and BP Neural Network

探讨了滚动轴承不同工况下非线性系统关联维数的计算及应用,讨论了滚动轴承在不同工况下的关联维数,以此为工况特征量设计BP神经网络实现故障诊断,并以关联维数作为游隙大小的特征量为例,对故障诊断效果进行分析。

The computation and application of correlation dimension of non - linear system in different working conditions for the rolling bearing are discussed. The correlation dimension for the vibration signal of the rolling bearing is different in different working conditions and can be used as the characteristics for recognizing the rolling bearing failure. According to the working condition characteristics, the BP neural network has been designed to realize the failure diagnosis, taking the correlative dimension as characteristics value of clearance magnitude, and analysis on diagnosis effect has been made.