小波神经网络在往复泵故障诊断中的应用

小波分析是一种新的信号处理技术， 具有良好的时频局部化特征。将小波包分析应用于往复泵的故障特征提取， 建立了往复泵小波神经网络故障诊断模型。该模型使用小波包分析的特征提取及神经网络的非线性映射特性， 可完成从测量信号到往复泵故障的分类。以此为基础开发出的故障诊断软件系统具有使用的特征量少， 建造故障诊断系统较为简单等优点。通过对往复泵液力端多故障诊断实例的检验表明， 该系统完全能够较准确地识别出往复泵液力端多种常见故障。

Fault diagnosis for reciprocating pump with wavelet neural network

Wavelet analysis, a new technique for signal processing, possesses excellent characteristic of time frequency location and is suitable for analyzing the time varying or transient signals. The authors offer an approach to the feature extraction of reciprocating pumps, and establish a wavelet packet neural network fault diagnosis model of reciprocating pumps, which can carry out signal measuring and fault classification by means of feature extraction of wavelet packat analysis and nonlinear mapping of neural network. Based on this model, a fault identification software system of reciprocating pumps is developed. Diagnosis of faults of fluid end on a reciprocating pump proves the software system being capable of making accurate identification of various kinds of faults with less feature variables.