基于小波神经网络的油泵故障诊断

依据小波神经网络技术的各种优点，提出采用三层BP小波神经网络构造故障诊断模型，对油泵进行故障监测和诊断。该故障诊断方法对神经网络训练、故障特征参数提取和对应神经网络状态输出等均实现了数据库管理，对油泵多种常见故障取得了满意的诊断效果，不仅具有特征自动提取以及较强的自学习和自适应功能，而且操作维护简便。研究结果表明：信号的小波分析和神经网络识别的融合将为油泵状态监测与故障诊断系统的建立提供新的方法和更简便的途径；对油库安全维护与故障诊断具有重要意义。

THE OIL PUMP FAULT DIAGNOSIS BASED ON THE WAVELET NEURAL NETWORK

The oil pump is the important equipment of the oil deport.It′s very significant to assert and diagnose the pump fault.Adopting the fault diagnosis technology to examine the pump,finding out the fault and deal with it are the important aspects to the oil depot′s safety.According to the advantage of the wavelet neural network,in the paper，the three layers BP wavelet neural network is used to set up the fault diagnosis model and to diagnose the pump fault.This method can abstract the feature and selfstudy and selfadoption.It is easy to assert and diagnose the pump fault.It can realize the database management of the neural network study fault feature abstraction and the output of the neural network，the good effect is acquired in the oil pump fault diagnosis.As a result,a new way is provided to supervise and diagnose the pump fault by combining the wavelet analysis and the neural network.It is very significant to the oil deport.