利用ART1网络进行故障诊断的方法研究

ART1 NEURAL NETWORK APPROACH TO FAULT DIAGNOSIS

此文系统地阐述了自适应共振理论及 ART1神经网络模型的基本概念、特点、结构和算法。在对故障诊断进行理论分析的基础上 ,结合 ART1神经网络对输入的要求 ,提出了一种阈值法结合 ART1神经网络进行故障诊断的方法 ,并对此方法进行了理论分析和仿真实验。

This paper expatiates the ART (Adaptive Resonance Theory) and the conception,the characteristics,the structure and the algorithm of ART1 Neural Network(ART1 NN)model.On the basis of the fault diagnosis and the requirement of the ART1 NN's input,a method presented by this paper uses the sill-method integrating ART1 NN to diagnose the system failure.The paper analyses its theoretic and has simulating experiments,which indicate that the method can detect and diagnose the failures and then it can classify the failures.