粗糙集傅立叶神经网络及在故障诊断中应用

Rough Set Fourier Neural Network and Its Application in Fault Diagnosis

提出了粗糙集和傅立叶神经网络相结合的方法,进行粗糙集布尔逻辑离散化,并在此基础上求取初始隶属函数,以提高隶属函数准确性;再使用傅立叶神经网络进行诊断网络训练。以连续搅拌反应釜故障诊断为实例,研究结果表明,此方法可以减少网络训练时间并提高诊断精度,有效进行故障诊断。

A method which combines rough set and Fourier neural network was presented in this paper. In this method, initial membership function is determined on the foundation of Boolean logic discretization which improves the veracity of membership function. Then Fourier neural network was used for training diagnoses networks. Finally, this method was applied in CSTR fault diagnoses. The simulation result shows good performance of the proposed method in the learning speed and accuracy.