摘要
为了能够准确、快速地对离心压缩机的各类故障进行诊断,将小波理论与神经网络相结合,建立了离心压缩机的故障诊断模型,分析了Harr小波函数的定义和特性。针对离心压缩机的故障特征,建立了对应的小波分析神经网络,输入层的神经元为5个小波变换获得的故障特征向量;输出层神经元为5个离心压缩机的故障模式;隐含层有3层,神经元数总共有24个,该方法具有较高的计算精度和计算速度。
In order to carry out the complete fault diagnosis for centrifugal compressor correctly and quickly,the fault diagnosis model of centrifugal compressor was established combining the wavelet theory and neutral network.The Harr wavelet function was founded,and there five neurons in the input layer,and these neutrons were fault eigenvectors obtained through wavelet transform.There were five neurons in the output that denoted the fault mode.There were three implication layers with 24 neurons.The method had high calculation precision and efficient.
出处
《煤矿机械》
北大核心
2010年第10期246-248,共3页
Coal Mine Machinery
关键词
小波神经网络
离心压缩机
故障诊断
wavelet neutral network
centrifugal compressor
fault diagnosis