【摘要】 利用ANSYS软件对建筑中常用的单层单跨刚架进行建模、加载、求解,得到刚架未受损和有损伤的固有频率,再利用MATLAB软件建立Elman神经网络模型对单层单跨刚架损伤识别,利用求得的固有频率作为学习样本对网络进行训练,应用训练的网络进行损伤识别。说明以固有频率作为标识量的神经网络模型在结构损伤识别方面的可行性。更多还原

【Abstract】 Using ANSYS software,the analysis model is built,the frequency of the frame with injury or without injury is obtained after loading and calculating the model.In this article,using the MATLAB software,indentifying the injury of single rigid frame with one-story by Elman neural network,getting the natural frequency,which is used as the training network.It can show that using neural network to identify the damage of structure is feasible.更多还原