论文部分内容阅读
考虑到建筑结构的不确定性和非线性,提出了基于神经网络的结构系统辨识方法,构造了具有一个隐含层的神经网络,用误差反向的传播算法训练,模拟了一个多层结构在地震作用下的动力响应特性。计算机仿真表明了该神经网络辨识结构的有效性。
Considering the uncertainty and non-linearity of building structure, a method of structural system identification based on neural network is proposed. A neural network with a hidden layer is constructed and trained by the error propagation algorithm. A multi- Dynamic Response Characteristics under Earthquake. Computer simulation shows the effectiveness of the neural network identification structure.