基于尺度优化形态学滤波的滚动轴承故障诊断

Fault Diagnosis for Rolling Bearings Based on Scale Optimization Morphological Filtering

为高效提取轴承振动信号中的冲击故障特征,提出了一种改进形态学帽乘积算子(IMHPO);同时,提出了一种新的结构元素长度选择策略,根据振动信号的极值点确定结构元素的长度范围并采用Hoyer测度评估形态滤波信号中的故障特征信息来选择最优结构元素长度。此外,将对角切片谱(DSS)应用于形态学帽乘积算子的滤波信号,以解决宽带噪声污染并进一步消除故障无关分量。仿真分析和轴箱轴承试验结果表明,所提出的IMHPO-DSS方法能够增强故障相关冲击特征并有效诊断轴承故障。

To efficiently extract the shock fault features from bearing vibration signal, an improved morphological hat product operator(IMHPO) is proposed. At the same time, a new strategy is proposed for selecting the length of structural element. The length range of structural element is determined according to extreme points of vibration signal, and Hoyer measure is used to select the optimal length of structural element by evaluating the fault feature information in morphological filtering signal. In addition, diagonal slice spectrum(DSS) is applied to filtered signal of IMHPO to solve the broadband noise pollution and further eliminate the fault-unrelated components. The simulation analysis and axle box bearing test results show that the proposed IMHPO-DSS method can enhance the fault-related shock features and effectively diagnose the bearing faults.