运用粗糙集与小波神经网络诊断往复泵故障

Fault diagnosis on reciprocating pumps by using rough sets and wavelet neural networks

提出一种粗糙集理论与小波神经网络集成的往复泵故障诊断方法。首先利用小波包对采集信号进行分解和重构能量特征向量。然后应用SOM网络对故障诊断数据中的连续属性进行离散化,根据粗糙集理论,借助遗传算法进行故障诊断决策系统约简,获得最优决策系统。在最优决策系统的基础上,设计RBF神经网络对往复泵故障进行诊断。试验结果显示,该方法可以有效提高往复泵故障诊断的精度和效率。

A method integrating rough sets with wavelet neural networks is proposed for fault diagnosis on reciprocating pumps. First a wavelet package is used to disassemble the collecting signals and reconstruct the energy characteristic vectors. And then SOM network is deployed to discretize the continuous attributes in fault diagnosis data. Based on the rough set theory, the diagnostic decision system is reduced by using genetic algorithm, thus an optimized decision system is obtained. On the basis of optimized decision system, RBF neural network is designed for diagnosing the faults in reciprocating pumps. The test result indicates that the method can be used to effectively improve the accuracy and efficiency of the pump diagnosis.