基于径向基函数神经网络的转子系统裂纹故障诊断

结合基于模型的转子系统诊断技术和径向基函数(RBF)神经网络在辨识非线性系统动态时的优势,本文提出了一种新的转子系统裂纹故障诊断方法.该方法采用RBF神经网络对裂纹转子系统的未知动态进行辨识,实现部分神经网络权值收敛到最优值以及神经网络对系统未知动态的局部准确逼近;诊断过程中利用已辨识的信息实现转子系统裂纹故障的快速检测与分离.所提方法尤其适用于微小裂纹的在线检测与定量识别.最后,以Jeffcott转子系统裂纹故障诊断为例进行仿真,仿真结果验证了所提方法的有效性.

Fault diagnosis of cracked rotor systems based on radial basis function neural networks

Combining the model-based fault diagnosis method of rotor systems and the advantage of radial basis function (RBF) neural networks on identifying nonlinear system dynamics, we propose a new fault diagnosis scheme for cracked rotor systems. In the proposed scheme, RBF neural networks are used to identify the unknown dynamics of cracked rotor systems; the convergence of partial neural network weights to their optimal values as well as the locally accurate approximation of the unknown system dynamics are achieved. In the diagnosis process, the identified information is reused so that crack faults can be detected and isolated quickly. Small cracks can be detected online and recognized quantitatively with this method. The example of the crack diagnosis of Jeffcott rotor systems is illustrated to demonstrate the effectiveness of the proposed method.