小波包-神经网络在斜轴泵故障诊断中的应用研究

Fault Diagnosis of a Bend Axis Piston Pump by Wavelet Packet and Neural Networks

针对斜轴式柱塞泵零部件故障信息被泵本身的流体冲击、机械振动所淹没的问题 ,采用小波包将振动信号分解到不同的频带以提取有关部件的故障信息 ,并将小波包分解在不同频带反映斜轴泵工作状况的振动特征信息作为故障样本 ,研究人工神经网络结合小波分析对斜轴泵进行故障诊断的方法 ,建立了相应的 BP神经网络。研究结果表明训练成功的 BP网络可作为智能分类器对斜轴泵的常见故障进行识别和诊断。

The fault diagnosis of a bend axis piston pump by neural networks based on wavelet packet is studied,and the BP neural networks are established in the paper.According to the fault features extracted from the pump vibration via wavelet transform,the working conditions and the faults of the pump can be classified and identified by the BP neural networks.