防灾减灾工程学报2025，Vol.45Issue(6)：1411-1420,10.DOI:10.13409/j.cnki.jdpme.20250430099
基于机器学习的连续刚构桥施工线形预测研究
Research on Construction Alignment Prediction of Continuous Rigid-frame Bridges Based on Machine Learning
摘要
Abstract
Given the time-varying characteristics of material parameters caused by complex environ-mental factors during the construction of continuous rigid-frame bridges,the traditional finite element analysis method is limited by the assumption of a single design parameter,and it is difficult to accurate-ly simulate the actual changing state of the influencing parameters during the construction process,which leads to key technical problems such as insufficient alignment control accuracy during bridge construction.Therefore,this study proposed an intelligent prediction method for construction align-ment of continuous rigid-frame bridges based on machine learning.First,parametric finite element modeling was used to simulate the construction deflection response under the coupling of multiple fac-tors,and a complete deflection prediction sample database was constructed.On this basis,four con-struction-alignment prediction models based on support vector machine(SVM),random forest(RF),long short-term memory network(LSTM),and particle swarm optimization backpropagation neural network(PSO-BP)were established,and their prediction accuracy was compared and analyzed.The results showed that the prediction accuracy of RF and ELM models was relatively low.The LSTM model exhibited better prediction accuracy,with a maximum prediction error of 1.2 mm.The PSO-BP model showed the best prediction performance.Its R² values for the training set and the test set were 0.996 and 0.992,respectively.The absolute error of the predicted deflection was only 0.55 mm,and the relative error was less than 10%.The PSO-BP neural network enables accurate prediction of construction deflection,effectively improves the alignment control level of continuous rigid-frame bridge construction,and enhances the closure alignment accuracy of the bridge.The findings provide an important technical reference for the intelligent construction of bridge engineering.关键词
连续刚构桥/桥梁施工/机器学习/施工线形Key words
continuous rigid-frame bridge/bridge construction/machine learning/construction align-ment分类
建筑与水利引用本文复制引用
ZUO Hongpeng,WANG Bing,CHEN Peng,ZHANG Hongyan,WANG Li..基于机器学习的连续刚构桥施工线形预测研究[J].防灾减灾工程学报,2025,45(6):1411-1420,10.基金项目
甘肃省联合科研基金重点项目(24JRRA869)、山西交通控股集团科技创新项目(23-JKKJ-6)资助 （24JRRA869）
