摘要
为了有效的进行刀具状态监测,采用小波神经网络的松散型结合对刀具进行故障诊断。通过小波变换提取刀具磨损声发射(AE)信号的特征,即对AE信号进行小波分解,提取了5个频段的均方根值作为神经网络的输入.来识别刀具磨损状态。试验表明,均方根值完全可以作为刀具磨损过程中产生AE信号的特征向量。仿真结果表明,基于小波神经网络的刀具故障诊断对刀具磨损状态的识别效率高。该方法是有效的。
In order to improve cutting tool condition monitoring, a method of cutting tool fault diagnosis based on wavelet and artificial networks with relaxed structure is proposed in this paper. Mean square roots of five frequency segments are extracted from wavelet decomposition of AE signals,which is used as the inputs of neural networks to detect tool wearing conditions. Experiments indicate that the mean square roots can serve as the eigenvectors of the AE signals and the methed is effective.
出处
《振动．测试与诊断》
EI
CSCD
2006年第1期64-66,共3页
Journal of Vibration,Measurement & Diagnosis
关键词
小波变换
神经网络
智能故障诊断
刀具状态监测
wavelet transform neural networks intelligent fault diagnosis cutting tool condition moni-toring
