针对工程上齿轮箱实时监测和故障诊断的需要,对JZQ 250型齿轮箱展开研究,提出了状态监测特征值和故障特征向量相结合的分级诊断方法。通过时域参数分析提取监测特征值作为一级诊断,利用能量比重谱的方法,提取故障特征向量作为神经网络的输入向量进行故障诊断为二级诊断。实验结果表明,该方法不仅满足工程上实时在线诊断的要求,而且能够准确进行齿轮箱故障定位。
This paper carries out research on JZQ250 type gearbox for the engineering requirements of real-time detection and fault diagnosis of gearbox,and puts forward a method which combines the eigenvalue of condition monitoring and the eigenvector of fault.The eigenvalue of condition monitoring as first order diagnosis is extracted by time domain analysis.The eigenvector of fault as the input vectors of neural network is extracted by the way of energy-proportion spectrum for second order diagnosis.The results show that the method not only can meet the requirements of the real-time and on-line diagnosis in engineering,but also can accurately locate the fault of gearbox.