齿轮箱起动信号的倒阶次谱分析方法

Processing Gearbox Launching Signal with Order Cepstrum Analysis

研究旋转机械非稳态信号的分析方法.对等时间间隔采样的齿轮箱振动信号,利用插值算法实现角域重采样,得到等角度采样的插值点,分别对原始信号和重采样信号进行传统的频谱分析和阶次分析.结果显示出阶次分析法在处理转速变化信号时的优越性:阶次谱分析能够有效地避免传统频谱方法所无法解决的“频率模糊”现象.将倒频分析引入阶次分析中,成功的识别了齿根裂纹故障.

The processing method of the non-stationary rotary machine vibration signal is studied. The angle domain signal is resampled from the time domain signal with interpolation algorithm. The order spectrum of the angle domain signal, as well as the spectrum of the time domain one, is worked out. It shows that the order analysis method can avoid effectively the “frequency smear” phenomenon which is beyond the traditional frequency spectrum analysis. The cepstrum algorithm is led-in the order analysis, and the gear tooth crack fault is diagnosed by order cepstrum successfully.