基于形态梯度谱熵的滚动轴承退化特征提取

Degradation feature extraction based on morphological gradient spectrum entropy

提出一种基于形态梯度谱熵(MGSE)的结构复杂度分析方法,以此对滚动轴承的退化程度进行表征。首先对信号进行数学形态梯度运算,利用数学形态梯度算子在运算速度以及冲击成分提取中的特点,得到信号的冲击谱分布,结合信息熵理论,定量刻画信号的结构复杂度。将该方法应用到滚动轴承性能衰退因子的分析中,以MGSE定量描述性能退化过程中的非线性演化趋势。采用仿真信号和辛辛那提大学智能维护系统(IMS)滚动轴承全寿命数据进行分析,对比结构复杂度中的SE谱熵和C0复杂度,并对影响因素进行分析。结果表明:提出的形态梯度谱熵方法能够定量表征滚动轴承的性能退化程度,具有运算速度快、影响参数少、结果稳定的优势。

Astructural complexity method based on mathematical morphological gradient spectral entropy(MGSE)is proposed. Firstly,mathematical morphological gradient operation is performed on the signal,the character of the morphological gradient operator in calculation speed and impact component extraction is introduced and obtaining the impact spectrum distribution of the signal is obtained. Combining with the information entropy theory,MGSE is able to characterize the structural complexity of the signal quantitatively. The method is applied to the analysis of rolling bearing performance degradation factor and MGSE is used to describe nonlinear evolution trend in performance degradation process quantitatively. The simulation signal and University of Cincinnati intelligent maintenance system(IMS)rolling bearing life data were used to analyze the SE spectrum entropy and C0 complexity in the structural complexity,and the influencing factors were analyzed.The results show that the proposed mathematical morphological gradient spectral entropy method can quantitatively characterize the degree of performance degradation of rolling bearings,and has the advantages of fast calculation speed,less influence parameters and stable results.