RBF神经网络在旋转机械故障诊断中的应用  

Application of RBF Neural Network to Fault Diagnosis of Rotary Machines

作　　者：陈松[1] 

机构地区：[1]安徽建筑大学机械与电气工程学院,合肥230601

出　　处：《三峡大学学报（自然科学版）》2014年第4期101-104,共4页Journal of China Three Gorges University：Natural Sciences

摘　　要：以频谱能量和轴心轨迹特征作为机械故障的特征量,通过RBF神经网络对机械故障进行诊断,使用Matlab软件编程并得出识别结果.与BP神经网络、Elman神经网络相比,RBF神经网络在机械故障诊断中具有训练次数少,速度快,准确度高,识别误差小的优点.The frequency spectrum energy and axis orbit characteristics are used as the characteristics of me-chanical failure.Matlab software is used for programming,and then to obtain the recognition results.Using radial basis function（RBF）neural network in mechanical fault diagnosis,compared with the BP neural net-work and Elman neural network,the former has higher speed,higher accuracy and small identification error.

关 键 词：RBF神经网络 故障诊断 MATLAB 旋o转机械 转子故障 

分 类 号：TH165.3[机械工程—机械制造及自动化]