双谱分析在往复泵故障诊断中的应用研究

Application of Bispectral Analysis to the Fault Diagnosis of Reciprocating Pump

针对当机械系统发生故障时 ,非线性特征更为突出的特点 ,提出了基于双谱分析的往复泵故障的振动诊断方法 ,研究了利用阀箱上振动信号的双谱 ,尤其是双谱沿 f1 =f2 的切片谱识别泵阀故障。诊断实例表明 ,采用双谱分析可有效抑制振动信号中的噪声 ,得出比传统功率谱分析更明显的泵阀故障特征 ,可有效识别往复泵泵阀的初期故障。由于该方法仅利用阀箱上振动响应信号的双谱分析即可完成对泵阀不同类型故障的正确识别 。

When some faults appear in a machinery system, nonlinear features become much more prominent. This paper puts forwards a new fault diagnosis method based on bispectral analysis for reciprocating pump. With bispectral analysis, especially with the f1=f2 diagonal slice spectrum of the bispectrum on valve case, the identification of valve faults is car out. The diagnosis samples of pump valve shows that the bispectral analysis can effectively suppress noise so that it is much more sensitive to the fault features than traditional FFT power spectrum analysis, and the incipient faults of pump valve can be easily detected. As this method can distinguish different faults with the aid of bispectrum of vibration signal on valve cases, so it could be easily extended.