摘 要

   随着高强度结构材料的发展、施工技术方法的进步，斜拉桥桥梁向着轻巧化、柔性化、大跨度的方向发展。斜拉桥属于高次超静定结构，施工阶段其结构内力和主梁线形变化相当复杂，且施工过程中受到的各种不确定性影响因素繁多。因此，斜拉桥的施工控制是十分重要的，特别是线形控制目标的实现直接关系到整个桥梁的结构安全状态。

   本文首先详述了国内外斜拉桥建设及施工控制发展的现状和发展趋势，然后对斜拉桥线形施工控制的重要性、主要内容、思想方法、影响因素、施工控制系统进行了概述，并对在线形施工控制中常见的四种预测与控制理论进行了对比分析，探讨了人工神经网络应用于斜拉桥线形施工控制的合理性。以自适应控制法为指导方法，以某斜拉桥的线形施工控制具体的实施方案为基础，建立了某斜拉桥的线形施工控制系统和主梁轴线偏差的 BP 神经网络预测模型，该控制系统包括施工控制组织管理分系统、结构状态监测分系统、数据管理分系统、安全预警管理分系统、结构仿真计算与分析分系统五个组成部分，重点介绍了此控制系统主要模块的服务功能；同时运用所建立的轴线偏差预测模型，并以部分节间主梁的实际监测数据为样本数据和测试数据对其主梁轴线偏差进行了预测，验证了线形施工控制系统和 BP 神经网

络在斜拉桥线形施工预测与控制中的应用价值。

关键词：斜拉桥；线形控制；控制系统；BP 神经网络；轴线偏差；自适应控制法

Abstract

   With  the  employment  of  highstrength  structural  materials  and  the  development  ofconstruction technique method,the Cable-Stayed Bridge will be more light and flexible. And its span is larger. The Cable-Stayed Bridge is high-Statically Indeterminate Structure.During its construction, the variation of structural internal force and main beam linearisvery complicated. Meanwhile, there is much uncertain factor. Therefore, the constructioncontrol  is  very  important  to the  CableStayed  Bridge.  In  particular,  its  successful  linearcontrol is related to the Structural safety of the stateof the whole bridge.

   First,  the  thesis  introduced  the  present  situation  of  the  CableStayed  Bridge’sconstruction  and  construction  control.  Second,  it  summarized  the  linear  constructioncontrol  of  the  CableStayed  Bridge,  including  importance,  main  contents,  supervisingmethod, factor, precontrol theory and construction control system, and Artificial Neuralnetworkis  help  for  the  linear  construction  control  of  the  CableStayed  Bridge.  Third,Taking the adaptive control method for the supervising method, based on the plan of theconstruction control of a CableStayed Bridge, the thesis established a BP neural networkprediction  model  for  axis  deviation  of  main  girder,  and  put  forward  a  construct  controlsystemwhich included construction control organization management, data management,structural  states  monitoring,  structural  simulation  analysis,  and  safety  early  warningmanagement.  Finally,  the  thesis  designed  a  software  for  the  bridge’s  linear  construction control system, and introduced the function of its main modules. After that, the thesis used

the prediction model of axis deviation for a CableStayed Bridge, the application value ofBP  neural  network  and  the  linear  construction  conctrol  system  for  the  Cable-Stayed Bridge’s pre-control is verificated.

   Key  words:  Cable-stayed  Bridge;  Linear  Control;  Control  System;  BP  Neural  Network; Axis Deviation；Adaptive Control Method