基于小波包无量纲免疫检测器的故障诊断模型

Model of Wavelet Packet and Dimensionless Immune Detector for Fault Diagnosis

为可靠地检出并识别旋转机械设备轴承故障,提出了一种基于小波包分解和无量纲免疫检测器的轴承故障模式识别方法.该方法采用小波包对原始时域信号进行预处理,分别提取原始时域信号和各频带范围内时域信号的无量纲指标,并计算其敏感因子,根据敏感因子选取敏感特征,结合人工免疫阴性选择算法,生成多个敏感特征无量纲免疫检测器,实现对故障进行识别和分类.仿真实验结果表明,所提方法能有效地识别各种轴承故障.

In order to check and identify bearing fault of rotating machines reliably, a pattern recognition method for bearing fault diagnosis based on wavelet packet decomposition and dimensionless immune detector was presented. Wavelet packet decomposition was presented as a pro-processing tool for the original time domain signal, then the dimensionless indicators of the original time domain signal and the various bands signals were extracted separately, and the sensitivity factor of each indicator was calculated, and the sensitive feature was determined by their value. Combined with artificial immune systems negative selection algorithm, several dimensionless immune detectors of sensitive features were constructed to identify and classify bearing fault. The simulation experiment results showed that the methods were effective to identify various kinds of bearing fault.