广义动态模糊神经网络及在轴承故障诊断中的应用

提出一种结合小波包分解和广义动态模糊神经网络的故障诊断方法,该方法首先采用小波包分解与重构提取各频带的能量作为故障特征向量,并以此向量作为输入,再利用广义动态模糊神经网络建立轴承故障诊断模型,该模型不仅能对模糊规则而且能对输入变量的重要性做出评价,使每个输入变量和模糊规则都可根据误差减少率进行修正。实验结果表明:该方法对识别和预测轴承的状态具有较高的精度和效率。

Generalized Dynamic Fuzzy Neural Network and Application in Bearing Fault Diagnosis

A novel method for fault diagnosis combining wavelet packet decomposition and generalized dynamic fuzzy neural network(GD-FNN) is proposed.The eigenvectors are extracted with energy of each band by wavelet packet decomposition first.Then the eigenvectors are taken as input sample of GD-FNN to create model of bearing fault diagnosis.This model can not only make the evaluation on the importance of the fuzzy rules,but also that of input variables,so that each input variables and fuzzy rules can be amended in accordance according to the error reduced rate.The experimental results prove this method has higher accuracy and efficiency to recognize and predict the state of bearing.