小波包神经网络在离心压缩机故障识别中的应用

为了能够提高离心压缩机故障识别的准确性和效率,构造了小波包模糊神经网络对离心压缩机进行故障识别。研究了小波包分析提取特征向量的方法。利用小波包分析获得了离心压缩机的特征向量后,利用模糊神经网络对其进行故障识别。利用小波包模糊神经网络对离心压缩机常见的5种故障进行了识别,结果表明该方法具有较高的故障识别能力。

Application of Wavelet Package Neutral Network on Fault Identification of Centrifugal Compressor

In order to improve the correctness and efficient of the centrifugal compressor,the wavelet package fuzzy neutral network was established.The extracting eigenvectors method depending on wavelet package analysis.After the obtaining the eigenvectors of the centrifugal compressor,the fault identification were carried out by using fuzzy neutral network.The five faults of the centrifugal compressor were identified by using wavelet package fuzzy neutral network,and the results showed that this method had better characteristic.