摘要
利用神经网络和Kalman滤波技术,提出了一种直接识别结构物理参数的方法,用Kalman滤波技术训练网络。在贮仓振动台实验的基础上,用贮仓在动载作用下的位移、速度作为网络的输入,激振加速度和响应加速度作为网络的输出。仿真计算表明,本文方法是可行的。
A method to identify the physical parameters of structure system has been developed by using neural network. The Kalman filtering technique is applied to modify the weight matrices of neural network. Based on the shaking table test,we compose a neural network in which the input signals are the response displacement and velocity of the silo, and output signals are the response acceleration of the silo plus input acceleration of the shaking table at time step t . Results from computer simulation studies show that this method is valid and feasible.
出处
《世界地震工程》
CSCD
北大核心
2005年第3期116-121,共6页
World Earthquake Engineering
关键词
识别
神经网络
KALMAN滤波
identification
neural network
the Kalman filter algorithms
