钢桥塔中锚室焊接变形预测与工艺优化  

Welding deformation prediction and process optimization for the central anchor chamber in the steel bridge tower

作　　者：强斌[1,2] 谢云杰 雷电 杨洪 陈虎 郭红艳 李亚东 QIANG Bin;XIE Yunjie;LEI Dian;YANG Hong;CHEN Hu;GUO Hongyan;LI Yadong(School of Civil Engineering,Southwest Jiaotong University,Chengdu 610031,China;National Key Laboratory of Bridge Intelligence and Green Construction,Chengdu 611756,China;Ning’an Railway Co.,Ltd.,Hefei 230041,China;China Railway Jiujiang Bridge Engineering Co.,Ltd.,Jiujiang 332000,China)

机构地区：[1]西南交通大学土木工程学院,成都610031 [2]桥梁智能与绿色建造全国重点实验室,成都611756 [3]宁安铁路有限责任公司,合肥230041 [4]中铁九桥工程有限公司,九江332000

出　　处：《东南大学学报(自然科学版)》2025年第2期459-467,共9页Journal of Southeast University:Natural Science Edition

基　　金：国家自然科学基金资助项目(52378198).

摘　　要：为优化钢桥塔中锚室的焊接工艺,减小焊接变形,采用Abaqus有限元软件对马鞍山公铁两用长江大桥钢塔中锚室的焊接过程进行了仿真模拟。与三维激光扫描所测实际焊接变形进行对比,仿真结果相对误差基本小于20%,验证了模拟的合理性。进而系统分析了焊接方向、焊接顺序及边界约束对中锚室焊接变形的影响,建立了涵盖128组焊接变形的数据库。通过深度神经网络算法,对焊接工艺参数进行优化筛选,获得最优工艺组合,最大变形值较初始焊接工艺下的变形值降低约40%。优化后的焊接工艺能够有效减少焊接变形,提高制造精度,可为钢桥塔的实际焊接提供可靠的理论依据和技术支持。To optimize the welding process of the central anchor chamber in the steel tower of the Ma’anshan Yangtze River Bridge for road and rail traffic and to reduce welding deformation,the welding process was simulated using the Abaqus finite element software.The simulation results were compared with the actual welding deformation measured by 3D laser scanning with a relative error basically within 20%,which proves the rationality of simulation.Then,the influence of welding direction,welding sequence,and boundary constraint on the welding deformation of the central anchor chamber was systematically analyzed,and a database covering 128 sets of welding deformation was established.Through the deep neural network algorithm(DNN),the welding process parameters were optimized and screened,and the optimal process combination was obtained,and the maximum deformation value was reduced to about 40%of the original value.The optimized welding process can effectively reduce the welding deformation,improve manufacturing accuracy,and provide reliable theoretical basis and technical support for the actual welding of steel bridge towers.

关 键 词：钢桥塔中锚室 焊接变形预测 三维激光扫描 深度神经网络 焊接工艺优化 

分 类 号：TU391[建筑科学—结构工程]