基于BP神经网络算法的特大跨度斜拉桥施工过程中的线形预测  

Application of BP Neural Network in Alignment Predicition of Super-span Cable-stayed Bridges Construction

作　　者：卜一之[1] 李玉耀[1] 

机构地区：[1]西南交通大学土木工程学院,四川成都610031

出　　处：《现代交通技术》2009年第1期24-27,共4页Modern Transportation Technology

基　　金：国家科技支撑计划项目(2006BAG04B03)

摘　　要：随着斜拉桥跨度的增大,其几何非线性因素影响越来越明显,这就需要寻求考虑几何非线性影响的特大跨度斜拉桥施工线形预测方法。文章分析BP神经网络算法的基本原理及改进措施,借助Matlab语言建立预测模型,探讨其在特大跨度斜拉桥施工控制中的应用。以某主跨为1088m的特大跨度钢箱梁斜拉桥为工程背景,验证该方法的合理性和可行性,为同类型桥梁的线形控制提供参考。With the span of cable-stayed bridge increasing, the geometric nonlinear influence becomes more and more obvious,which need to seek a method of alignment prediction considering the geometric nonlinearity in super-span cable-stayed bridge construction.In this paper,the basic principles and measures are analyzed for improvement of BP neural network with the establishment of Matlab language model to explore its application in construction control of large-span cable-stayed bridge.Taking a steel girder cable-stayed bridge whose main span is 1 088 m as engineering background,the rationality and feasibility of the method are vertified.The results could provide a reference to construction control of the similar bridges.

关 键 词：斜拉桥 BP神经网络 线形预测 施工过程 

分 类 号：U448.27[建筑科学—桥梁与隧道工程]