基于神经网络的连续梁桥结构损伤识别

The damage evaluation of continuous bridge structures based on neural network

探讨了神经网络技术对连续桥梁的损伤位置和损伤程度识别的方法,用有限元软件ANSYS的模态分析技术对连续梁结构的损伤情况进行模拟,得到损伤标示量并将其作为神经网络训练样本,数值仿真结果表明,该方法对于实际工程结构的损伤识别具有一定的参考意义。

Based on neural network the methods for evaluation of damage position and degree of continuous bridge are discussed in this paper. Then, the damage simulation of the structures has been done by using FEM modal technology and ANSYS sofaware, in which the damage variables would be as the training samples of the neural network. All the results show that the methods are important for the damage evalution of practical bridge structures.