谱相关密度分析在轴承点蚀故障诊断中的研究

Spectral correlation density analysis on bearing point defect diagnosis

利用谱相关密度提取轴承故障特征时需要在循环频率域和频率域上同时兼顾高分析带宽和高分辨率,从而使得该方法的计算量庞大,难以达到较高的分析精度。鉴于此,首次在循环平稳分析中引入解析的思想,利用解析形式的谱相关密度在循环频率域不存在高频特征的特点,提出运用时域选抽技术,在保证分辨率的同时降低分析带宽,减少计算量,从而得到更好的分析效果。本文以一般调幅信号解析形式的谱相关密度分析为基础,对滚动轴承点蚀故障模型进行了分析,推导了其谱相关密度分析的理论结果,给出具体的算法实现。仿真调幅模型和实际轴承故障信号,证实了理论分析的正确性和算法的可行性,同时也验证了谱相关密度分析对调幅特征的提取能力。

Spectral Correlation Density （SCD） analysis should simultaneously insure wide frequency analysis bandwidth and high frequency resolution at cyclic frequency domain and general frequency domain when dealing with amplitude modulation （AM） phenomena in rotating machinery. Huge computation of the method brings obstacles to high analysis precision. So, this paper introduces analytic principle into cyclostationary analysis to resolve the flaw of the method. Theoretical derivation shows that the analytic SCD of AM signal only includes low-frequency characters at cyclic frequency domain, so time decimation technique can minify frequency analysis bandwidth and reduce computation to great extent on the premise of high resolution at low frequency band. Based on AM signal analysis, a model of vibration signal from point defect in rolling element bearing is mainly talked about. The SCD characters of it are summarized and an effective arithmetic aiming to AM phenomena are presented. Simulation and experimental results verify the correctness of the theoretical analysis, and indicate that the analytic SCD is a robust method in AM character identification.