大跨度曲面钢屋盖变形施工监测与预测方法研究  

RESEARCH ON MONITORING AND PREDICTION METHODS FOR DEFORMATION CONSTRUCTION OF LARGE-SPAN CURVED STEEL ROOF

作　　者：董聿霖 巴盼峰 杨玉涛[2] 刘松杰 胡成 DONG Yu-lin;BA Pan-feng;YANG Yu-tao;LIU Song-jie;HU Cheng(School of civil engineering,Tianjin Chengjian university,300384,Tianjin,China;BCEG NO.4 Construction Engineering Co.,Ltd.,100024,Beijing,China)

机构地区：[1]天津城建大学土木工程学院,300384,天津 [2]北京建工四建工程建设有限公司,北京100024

出　　处：《建筑技术》2024年第23期2847-2851,共5页Architecture Technology

摘　　要：为探究大跨度长悬挑曲面钢屋盖施工阶段的应力应变及位移变化规律,以北京中关村论坛永久会址为研究对象进行施工监测研究。通过直接的现场试验模型、试验系统和试验数据分析,总结大跨度长悬挑曲面钢屋盖施工阶段中的变化规律,并提出基于长短期记忆神经网络(Long Short-Term Memory Neural Network, LSTM)数据预测方法,对钢结构在施工阶段内力与变形的发展趋势作出科学预测,并针对施工监测系统中涉及位移变形、应力应变等数据缺失情况利用模型预测值进行数据修复。最后结合现场实测数据分析与LSTM模型预测结果,提出基于形变主动控制的大跨度钢结构施工关键技术。In order to explore the stress,strain,and displacement change laws of large-span long-cantilevered curved steel roofs during the stage,this study conducted construction monitoring research on the Beijing Zhongguancun Forum permanent site.By directly using on-site experimental models,experimental systems,and data analysis,the change laws of large-span long-cantilevered curved steel roofs during the construction stage were summarized.Additionally,a data prediction method based Long Short-Term Memory Neural Network(LSTM)was proposed to scientifically predict the development trend of internal forces and deformation of steel structures during the construction stage Furthermore,the missing data related to displacement deformation and stress strain in the construction monitoring system were repaired using model predicted values.Finally,combining on-site measurement data analysis LSTM model prediction results,key technologies for the construction of large-span steel structures based on deformation active control were proposed.

关 键 词：大跨度 长悬挑 钢屋盖 主动监测 变形预测 

分 类 号：TU74[建筑科学—建筑技术科学]