铁路钢板梁桥的损伤识别与诊断研究  

Damage Detection and Diagnosis of Steel Plate Girders

作　　者：范连合[1] 王根会[1] 

机构地区：[1]兰州交通大学土木工程学院,甘肃兰州730070

出　　处：《兰州交通大学学报》2005年第6期35-39,共5页Journal of Lanzhou Jiaotong University

摘　　要：根据结构在损伤前后其固有频率的变化能够反映出结构的损伤信息这一特征,以洛河大桥为工程背景,通过BP神经网络模型在考虑无噪声和有噪声两种情况下分别对钢板梁桥整体损伤程度进行识别与诊断的研究.通过利用ANSYS软件提取出该桥在损伤前后的固有频率,并将其变化率作为网络的输入向量对网络进行训练,同时将该桥固有频率的实测值作为网络模型的测试向量,从而对钢板梁桥整体损伤程度进行准确的诊断.According to the characteristic that the shift of natural frequencies before and after structure damage can reflect the structural damage information,the detection and diagnosis of whole damage degree of steel plate girder of Luohe Bridge are studied, considering the condition of BP neural network without or with disturbance of noise. Software-ANSYSito distill natural frequency of steel plate girder before and after its damage, variance ration of natural frequency as input vectors tO train：the network, and the measured values of natural frequency of the bridge as testing vectors are applied to diagnose damage degree of the bridge exactly.

关 键 词：BP神经网络 铁路钢板梁桥 噪声 损伤 识别与诊断 

分 类 号：U441.4[建筑科学—桥梁与隧道工程]