基于多级神经网络的桥梁结构模态参数识别方法  

Modal Parameter Identification Method of Bridge Structure Based on Multi-level Neural Network

作　　者：鹿婧 LU Jing(School of Information and Architectural Engineering,Anhui Open Universit,Hefei Anhui 230022,China)

机构地区：[1]安徽开放大学信息与建筑工程学院,安徽合肥230022

出　　处：《广州航海学院学报》2023年第3期50-54,共5页Journal of Guangzhou Maritime University

基　　金：2022年安徽省高等学校自然科学研究项目(2022AH052684);2021年安徽省高等学校自然科学研究项目(KJ2021A1256)。

摘　　要：桥梁结构模态参数识别方法在识别过程中难以获得完整阶次,构造的基准模型不完整,导致识别结果出现误差,因此,设计一种基于多级神经网络算法的桥梁结构模态参数识别方法.使用多级神经网络对识别算法进行优化,建立交配池,利用交叉和变异算子对交配池中的参量个体进行识别,建立新的群体,利用信号匹配识别结构模态参数,选择模态确信准则(MAC)以及相位共线性指标(MPC)作为模态区分的辨别指标并计算,最后优化整体模态参数识别流程.方法性能测试结果表明,设计的基于多级神经网络的桥梁结构模态参数识别方法在不同采集方式下得到的参数误差更小,可靠性更高.The identification method for modal parameters of bridge structures is difficult to obtain complete orders during the identification process,and the constructed benchmark model is incomplete,resulting in erors in the identification results,Therefore,design a method for identifying modal parameters of bridge structures based on multi-level neural network algorithms.Use multi-level neural networks to optimize the identification algorithm,build an interchange pool,use crossover and mutation operators to identify individual parameters in the pool,and establish a new population,a method of bridge structural modal parameters identification based on multi-level neural network algorithm.The multi-stage neural network is used to optimize the recognition algorithm.By establishing the mating pool,the crossover and mutation operators are used to identify the parameter individuals in the mating pool,and a new population is established.The structural modal parameters are identified by signal matching,and the modal assurance criteria(MAC)and phase collinearity index(MPC)are selected as the discrimination indicators of modal discrimination and calculated.Finally,the whole modal parameter identification process is optimized.The performance test results show that the modal parameter identification method based on multilevel neural network can obtain less error and higher reliability under different acquisition methods.

关 键 词：多级神经网络 桥梁结构 模态参数 模态辨别指标 

分 类 号：U448.35[建筑科学—桥梁与隧道工程]