基于切片谱免疫系统的旋转机械故障诊断

Rotating machinery fault diagnosis based on slice spectrum-AIS

根据旋转机械故障振动信号的特点,本文基于免疫学机制,提出了切片谱免疫故障诊断方法,建立了基于该方法的诊断模型,介绍了该诊断方法的实现过程。切片谱免疫故障诊断方法利用双谱对先验故障样本进行分析,获得与旋转机械故障类型一一对应的特征样本,将特征样本归一化作为人工免疫系统(AIS)模式识别的特征向量,利用AIS中的阴性选择算法(NSA)进行自己-非己匹配运算得到检测器集,然后将建立(训练)好的检测器集应用于实时故障诊断。通过轴承故障诊断的实验,结果表明,本方法是可行和有效的。

A method of rotating machinery fault diagnosis, combing diagonal spectrum with artificial immune system （AIS） , is presented. As a fitted tool for processing non-Gaussian signal and nonlinear system, slice spectrum is able to eliminate additive Gaussian measurement noise, boost signal-to-noise ratio. Belonging to AIS, negative-selection algorithm （NSA） inspired by the mechanism of human immune system is capable of discriminating between the self （body elements） and the non-self （foreign pathogens）. However, in our application, the self is defined to be the normal status and the non-self the fault category of rotating machinery. As input feature vector, diagonal spectrum, calculated from vibration rotating machinery, is classified by the detectors generated by NSA, and finally determines the fault modes of rotating machinery. The experiment of rolling element bearing shows that the approach based on diagonal spectrum-AIS is practical and effective.