论文部分内容阅读
针对美国IASC-ASCE的结构健康监测科研组提出的基准结构进行结构自振频率识别研究.神经网络训练时使用的数据为有限元程序计算所得出,将有损伤结构在环境激励下某点的加速度响应,通过快速傅立叶变换得到的离散频率响应函数作为神经网络的输入;将损伤结构的自振频率作为神经网络的输出.通过对在不同噪声水平下训练的神经网络的识别结果进行分析比较,结果表明:应用人工神经网络进行结构自振频率识别是切实可行的.
Aiming at the reference structure proposed by ISTR-ASCE’s Structural Health Monitoring Research Group, the natural frequency of structures is studied.The data used in neural network training is calculated by finite element program and will have the acceleration of a damaged structure under certain environment excitation Response, the discrete frequency response function obtained by fast Fourier transform is used as the input of neural network, and the natural frequency of damage structure is taken as the output of neural network. By analyzing and comparing the recognition results of neural networks trained under different noise levels, It shows that using artificial neural network to identify the natural frequency of structures is feasible.