气象数据驱动的拉林铁路简支T梁温致变形预测  

Prediction of Temperature-Induced Deformation of Concrete T Beams of Lasha-Linzhi Railway Based on Meteorological Data

作　　者：李奇 揭崇清[2] 罗红英 柳斌[1] LI Qi;JIE Chongqing;LUO Hongying;LIU Bin(Research Center of Civil,Hydraulic and Power Engineering of Xizang,Xizang Agricultural and Animal Husbandry University,Linzhi Xizang,860000;School of Civil Engineering,Tongji University,Shanghai 200092,China)

机构地区：[1]西藏农牧学院西藏土木水利电力工程技术研究中心,西藏林芝860000 [2]同济大学土木工程学院,上海200092

出　　处：《高原农业》2024年第2期117-126,共10页Journal of Plateau Agriculture

基　　金：西藏自治区自然科学基金重点项目(XZ202301ZR0040G),高原环境下轨道平顺性变化规律及其车载动态监测方法研究。

摘　　要：为快速、准确预测日照环境下高原铁路桥梁的温度变形效应,基于长短期记忆神经网络(LSTM)提出了一种气象数据驱动的桥梁温度变形智能预测模型。以拉林铁路简支T梁为背景,采用热力耦合有限元仿真分析,构建了“气象数据—温度变形”映射的样本数据库,以此训练预测模型并对桥梁温致变形进行预测。结果表明,LSTM模型表现出了较高的精度和优势,其梁体竖向挠度预测的决定系数(R2)超过0.97,平均绝对误差(MAE)和均方根误差(RMSE)较反向传播神经网络(BP)模型提升超过70%,较随机森林(RF)模型分别提升了24%和27%。预测挠度与真实值在趋势和数值方面均基本一致,表明所提出的预测方法性能优异,为高原铁路轨道平顺性变化规律研究及动态检测数据评价提供参考。To rapidly and accurately predict the temperature deformation of plateau railway bridges in sunny environments,an intelligent prediction model for bridge temperature deformation driven by meteorological data was proposed based on the Long Short-Term Memory(LSTM)neural network.Taking the simply supported T-beam of the Lasa-Linzhi Railway as an example,a sample database mapping the meteorological data and temperature deformation was constructed through thermo-mechanical coupled finite element simulation,which was then used to train the prediction model and predict the temperature-induced deformation of the bridge.The results show that the LSTM model exhibits high accuracy and advantages,with a determination coefficient(R2)exceeding 0.97 for the prediction of vertical deflection of the beam.Compared with the Back Propagation(BP)neural network model,the Mean Absolute Error(MAE)and Root Mean Square Error(RMSE)of the LSTM model are improved by more than 70%.Compared with the Random Forest(RF)model,the MAE and RMSE are improved by 24%and 27%,respectively.The predicted deflection is basically consistent with the true value in both trend and numerical value,indicating excellent prediction performance of the proposed method.This provides a reference for investigating the variation patterns of track irregularity of plateau railways and evaluating dynamic detection data of the track irregularity.

关 键 词：高原桥梁 气象数据 简支T梁 预测模型 温致变形 长短期记忆神经网络 

分 类 号：U448[建筑科学—桥梁与隧道工程]