为了提高故障诊断正确率，提出了一种基于复双谱的机械故障诊断方法．这种方法根据复双谱的不同耦合方式进行故障判别．由于耦合方式不同，复双谱包含的信号间的耦合信息也不相同，这些信息可以用来进行故障诊断．在减压阀故障诊断实验中，对特定的故障信号和正常信号同时采用相同的耦合方式，根据二维小波可以有效提取信号中频率信息的特点，利用二维小波提取复双谱中的这些耦合信息的特征值，并将这些特征值输入支持向量机进行故障判别．实验结果表明，随着耦合方式的不同，同一故障的诊断正确率也不相同．%To improve the correct rate of fault diagnosis, a method of mechanical fault diagnosis based on complex bispectrum was proposed. This method diagnosed fault according to the different coupling modes of complex bispectrum. Because of different coupling modes, the complex bispectrum contained different coupling information of signals, and the information could be used to diagnose fault. In the fault diagnosis experiment of pressure reducing valve, the same coupling mode was applied to both fault signals and normal signals. Because two dimensional wavelet could effectively extract frequencies information from signals, the features of coupling information in complex bispectrum were extracted by 2-D wavelet, then the features were input into support vector machine to diagnose fault. Experiment results show that the correct rates of the same fault are also different with the different coupling modes.
展开▼
