波形钢腹板连续刚构桥架设过程的智能化施工控制  

Intelligent Construction Control of Continuous Rigid Frame Bridge with Corrugated Steel Webs During Erection Process

作　　者：白云腾 王晓明[1,2] 录哲元 张嘉鼎 BAI Yun-teng;WANG Xiao-ming;LU Zhe-yuan;ZHANG Jia-ding(Institute of Bridge Engineering,Chang an University,Xi'an 710064,Shaanxi,China;Key Laboratory of Transport Industry of Bridge Detection Reinforcement Technology,Chang'an University,Xi'an 710064,Shaanxi,China;Yellow River Engineering Consulting Co.,Ltd.,Zhengzhou 450003,Henan,China;CCCC Second Highway Engineering Co.,Ltd.,Xi'an 710065,Shaanxi,China)

机构地区：[1]长安大学桥梁工程研究所,陕西西安710064 [2]长安大学旧桥检测与加固技术交通行业重点实验室,陕西西安710064 [3]黄河勘测规划设计研究院有限公司,河南郑州450003 [4]中交第二公路工程局有限公司,陕西西安710065

出　　处：《建筑科学与工程学报》2022年第4期137-145,共9页Journal of Architecture and Civil Engineering

基　　金：国家自然科学基金项目(52178104);陕西省交通科技项目(19-30K)。

摘　　要：针对波形钢腹板连续刚构桥架设过程中腹板剪切变形所致挠度预测问题,建立了采用思维进化算法(MEC)优化BP神经网络(MEC-BP)的挠度预测框架,研发了基于可视化管控系统的智能化施工控制模块。挠度预测框架主要包括三部分:首先,获取用于建立预测模型的数据集,其中输入样本和输出样本分别由拉丁超立方体采样(LHS)技术和高精度实体有限元模型生成;然后,基于所获得的数据集,建立了基于MEC-BP神经网络的挠度预测模型,并根据统计准则对模型性能进行评估;最后,利用训练良好的MEC-BP模型对波形钢腹板连续刚构桥架设施工的挠度进行预测。智能化施工控制模块主要包括两部分:首先,根据规范建立分级预警指标体系,分别为红色、黄色及绿色预警;其次,基于可视化管控系统研发了施工控制模块,并开发了手机端APP,在此基础上,结合MEC-BP预测模型,实现了波形钢腹板连续刚构桥架设过程的智能化施工控制。以梁渠沟大桥为背景,将挠度预测值与现场实测值进行对比验证。结果表明:MEC-BP预测模型的挠度预测值与现场实测值吻合较好,预测误差均满足要求;智能化施工控制系统在梁渠沟大桥架设过程中成功预警3次,助力梁渠沟大桥的建设取得圆满成功;研究成果可为波形钢腹板连续刚构桥架设施工的智能化控制提供理论依据。Aiming at the deflection prediction problem caused by web shear deformation during the erection of corrugated steel web continuous rigid frame bridge, the deflection prediction framework of BP neural network(MEC-BP) optimized by mind evolutionary algorithm(MEC) was established, and an intelligent construction control module based on visual control system was developed. The deflection prediction framework mainly included three parts. Firstly, the data set used to establish the prediction model was obtained, in which the input samples and output samples were generated by Latin Hypercube Sampling(LHS) technology and high-precision solid finite element model respectively. Then, based on the obtained data set, the deflection prediction model based on MEC-BP neural network was established, and the performance of the model was evaluated according to the statistical criteria. Finally, the well-trained MEC-BP model was used to predict the deflection of corrugated steel web continuous rigid frame bridge. The intelligent construction control module mainly included two parts. Firstly, a hierarchical early warning index system according to the specifications was established, which were red, yellow and green early warning respectively. Secondly, the construction control module was developed based on the visual control system, and the mobile APP was developed. On the basis, combined with the MEC-BP prediction model, the intelligent construction control of the erection process of corrugated steel web continuous rigid frame bridge was realized. Taking Liangqugou bridge as the background, the deflection prediction results were compared and verified with the field measured values. The results show that the deflection prediction values of MEC-BP prediction model are in good agreement with the field measured values, and the prediction errors meet the requirements. The intelligent construction control system successfully gives early warning three times during the erection of Liangqugou bridge, which helps the construction of Liangqugo

关 键 词：架设过程 波形钢腹板 连续刚构桥 挠度预测 MEC-BP神经网络 智能化施工控制 

分 类 号：TU745[建筑科学—建筑技术科学]