多分量振动信号时频分析与应用研究

定义了一种新的L．Cohen类时频分布的动态组合核函数，并给出了设计方法。该核函数基于信号特征，在模糊面内核函数是动态和可变的，克服了固定形式的核函数只能分析单一信号的缺点，能同时分析由单频率信号、脉冲信号和多个线性调频信号组成的多分量信号。在机械设备故障诊断中，核函数的参数随机械运动参数的变化而变化，可同时提取旋转机械的谐波分量和脉冲信号，处理机器变速过程中的转速调频、零件冲击和固有频率等的多分量振动信号的识别。仿真和实例分析表明：动态核函数能很好地去除交叉干扰项，提高分辨率，同时又证明了该方法的实用性和有效性。

Time-Frequency Analysis of a Multi-Component Vibration Signal and Its Application

A new kernel function of Cohen's class time-frequency distribution,which can be used to analyze multi-component vibration signal,is defined and its design method is proposed in this paper. This kernel function is additive and changeable in ambiguity plane . A fixed kernel function can only be used to analyze the single-component signals, whereas the new kernel function can process the signals that are composed of the single-frequency, pulse and frequency-modulated components by changing its parameters. In the fault diagnosis of machinery,this new kernel can be employed not only to extract the fundamental frequency and its harmonic components,impact pulse of the rotating machine, but also to analyze the velocity-varied process. The results of the simulation and real examples demonstrate that this new kernel function can remove the interference terms effectively and enhance the resolution of time-frequency plane.