结构混合控制的神经网络方法

Hydrid Control of Structure using artificial neural networks

介绍结构混合控制的神经网络方法．以模型结构振动台实验记录为训练样本，训练一个作为控制器使用的多层前向ＢＰ 网络来模拟结构的控制准则，并用来模拟未来作动器控制力的大小；训练另一个的神经网络来模拟混合结构控制系统，并用该网络预测结构未来在地震作用下的反应；最后进行数值仿真分析和实验分析的对比研究，结果表明人工神经网络可以在结构混合控制中发挥作用．

In this paper a hybrid control system of structure using neural networks is introdced,a multi layer feed forward neural network with BP algorithm has been trained as an emulator to simulate the hybrid control system of structure and to forecast the future responses of the structure under ground mtion.Another controlled neural network has been trained to simulate hybrid control system of structure and it has been used to control the responses of the structure under ground motion.The analysis of numerical simulation and real test are compared and the results show that the using of neural networks in hybrid control of structure can improve the control efficiency of controlled system.