论文部分内容阅读
根据混凝土斜拉桥的结构特点,以双连体独塔斜拉桥为工程依托,通过建立实桥动力有限元模型,以模态分析结果为桥梁健康状态评估提供数值模拟试验模型及试验数据。借助Matlab图形用户界面,选择神经网络参数(网络的类型、网络拓扑结构、各层传递函数、特征参数、数据前处理方法及训练样本个数),构建BP神经网络,在对混凝土斜拉桥结构健康状态评估模型研究的基础上,提出了基于人工神经网络的评估技术,并通过实桥应用,完成对样本矢量的输入,以及对所建网络的训练,为实现混凝土斜拉桥健康状态评估提供合适的工具。结果表明,该方法能够有效反映在用混凝土斜拉桥结构的健康状态。
According to the structural characteristics of concrete cable-stayed bridges, double-pylon single-pylon cable-stayed bridges are relied on. Through the establishment of dynamic finite element model of the actual bridge, numerical simulation test models and experimental data are provided for assessment of bridge health status by means of modal analysis results. With the aid of Matlab graphical user interface, BP neural network was selected by selecting the parameters of neural network (network type, network topology, transfer functions of various layers, characteristic parameters, data preprocessing methods and the number of training samples) Based on the research of health assessment model, this paper puts forward an evaluation technology based on artificial neural network. Through the application of real bridge, the input of sample vectors and the training of the constructed network are completed, so as to provide the assessment of health status of concrete cable stayed bridges The right tool. The results show that this method can effectively reflect the health condition of the cable stayed bridge structure.