基于模型修正与灰色系统理论的梁桥损伤识别研究

Research on damage identification of girder bridge based on model correction and gray system theory

对既有梁桥结构开展损伤识别分析时,需要依托有限元软件建立理想的梁桥数值模型,然而依据设计图纸建立的梁桥有限元模型尚未考虑到建模过程的不确定性,有限元模型计算与既有梁桥实测存在较大偏差。利用基于遗传算法优化BP神经网络的模型修正方法,对修正后更加贴近实际梁桥结构的有限元模型展开数值分析,通过挠度影响线进行灰色系统相关性分析与损伤识别。修正结果表明:基于遗传算法优化BP神经网络的梁桥模型修正方法可使有限元模型贴近实际结构模型;结合现场模态测试,结果频率误差从未修正的30.5%降至修正后的9.8%。损伤识别结果表明灰色系统理论相关系数与损伤程度呈线性关系,灰色系统相关系数越小,损伤程度越大。

In order to carry out the damage identification analysis of existing girder bridge structure,it is necessary to rely on finite element software to establish an ideal bridge numerical model.However,the finite element model of the girder bridge established based on the design drawings has not yet taken into account the uncertainty of the modeling process,and the calculation of the finite element model has a large deviation from the actual measurement of the existing bridge.Using the model modification method based on BP neural network optimized by genetic algorithm,numerical analysis is carried out on the modified finite element model which is closer to the actual girder bridge structure,and gray system correlation analysis and damage identification are carried out through the deflection influence line.The results show that the proposed model modification method based on genetic algorithm optimized BP neural network makes the finite element model close to the actual structural model,and combined with the field modal test,the resultant frequency error decreases from the uncorrected 30.5%to the corrected 9.8%.The damage identification results show that the gray system theory correlation coefficient is linearly related to the damage degree,and the smaller the gray system correlation coefficient is,the greater the damage degree is.