论文部分内容阅读
在许多场合下,组成时域信号诸分量不是单纯地线性相加,而是具有卷积或相乘的关系,使信号的功能率谱出现混迭现象,不易识別。时序倒频谱可将功率谱上叠加的干扰周期分量简化成清晰易辨的谱线,可有效地把功率谱上识別不出的信号分量提取出来。本文介绍了时序倒频谱,给出了在机床噪声分析中识別回声和诊断低频杂音的应用实例。与传统的FFT倒频谱应用结果作了对比。
In many cases, the components of the time-domain signal do not simply add linearly, but rather have convolution or multiplication relations, resulting in aliasing phenomenon of signal power spectrum, which is not easy to identify. The time-series cepstrum can simplify the superposed periodic components of the power spectrum into clear and easy-to-distinguish spectral lines, which can effectively extract the signal components that can not be identified in the power spectrum. This article introduces the timing cepstrum and gives examples of how to identify echoes and diagnose low frequency noise in machine noise analysis. Compared with the traditional FFT cepstrum application results.