基于神经网络的可控整流主电路故障诊断实现的研究

<title>Controllable rectifier main circuit fault diagnosis based on neural network

介绍了如何构建故障诊断目的的神经网络,模拟故障的实际采集数据和采用BP算法实现了神经网络的学习即神经网络的权值和阈值参数的确定,测试样本的输出结果表明构造网络的实用性和学习训练参数的正确性.

<Abstrcat>In this paper,it stresses how to structure the proposition of fault diagnosis.Study of neural network are completed applying BP algorithm and practice collecting datum.Power and value are decided according to the train.The conclusion of measure sample shows that the structure of neural network is practice and the parameter of train is right.