旋转机械故障诊断的11/2维循环谱分析方法

Analysis method of rotation machinery fault diagnosis using 11/2 dimension cyclic spectrum

首先介绍了谱相干及高阶循环累积量谱的定义 ,说明了高阶循环累积量谱具有的抑噪和相位保留的能力 .但由于高阶循环累积量谱分析需要更长的时间序列 ,因此本文从循环谱相干定义出发 ,提出了利用 1 1 /2维循环谱区分线性与非线性特征的方法 .指出 1 1 /2维循环谱分析方法 ,不仅保留了高阶循环累积量谱的抑制噪声、提取非线性耦合特征和分离循环平稳分量与平稳分量的能力 ,而且减少了所需时间序列长度 ,提高了监测的实时性 ,最后以非线性平方相位耦合信号和轴承故障信号为例进行了仿真验证 .

This paper introduces high-order cyclic cumulation and spectrum coherence definition, and explains the high-order cyclic cumulation features which suppress noises and hold phase. Based on cyclic spectrum coherence definition, a method using 11/2 dimension cyclic spectrum to distinguish linear or nonlinear characteristic is presented since longer time series is required to analyze high-order cyclic cumulation. This method holds the high-order cyclic cumulation features in suppressing noises, extracting nonlinear coupling characteristic and separating cyclostationary component from stationary component, and decreases the analysis length of time series and increases the real time degree of inspection. The nonlinear square phase coupling signal and the bearing fault signal are used to simulate and testify the validity of this method.