Elman神经网络在非线性模拟电路故障诊断中的应用

Elman神经网络把隐层输出反馈回输入层,是一种动态递归神经网络,改进的Elman网络能逼近任意动态非线性系统。本文研究了Elman网络的结构和学习算法,并把改进的Elman网络应用于非线性模拟电路故障诊断激励信号参数优化中,以提高故障正确诊断率。首先用改进的Elman神经网络建立故障电路和无故障电路系统模型,然后用遗传算法搜索电路故障诊断的最佳激励信号参数。

Application of Elman neural network in nonlinear analog circuits diagnosis

Elman neural network whose hidden layer outputs were fed back to input layer is a dynamic recurrent neural network.The modified Elman network can approximate any dynamic nonlinear system.This paper researched the structure and the learn arithmetic of Elman neural network.In this paper,Elman network was used to optimize the parameters of testing stimulus of nonlinear analog circuits diagnosis.Elman network was used to develop the models of the fault circuits and fault free circuit.,then searched the optimum stimulus based on genetic algorithm.