摘要
与一般公铁桥梁不同,磁悬浮轨道梁的温度场会受到周围磁场与功能件升温的影响。由于我国规范暂未对磁浮轨道梁均匀温度和温差分量的极值作出明确规定,研究具有概率保证的温度分量代表值具有重要意义。根据长沙磁浮运营线某混凝土箱梁1年多的现场温度监测数据,分析梁体均匀温度与竖向等效线性温差的时变特征。研究分别采用单变量线性回归法(ULR法)和梯度提升回归树法(GBRT法)建立基于气象因素的结构温度分量预测模型,对比发现GBRT法不仅能提高均匀温度的预测精度,还能克服ULR法无法预测竖向等效线性温差的局限性。而后将1960—2019年的长期气象数据输入到建立好的GBRT预测模型中,得到磁浮箱梁各温度分量60年的长期预测值,经整理后得到均匀温度和竖向等效线性温差的日极值、年极值样本。结合最大熵极值模型和Fourier函数得到各温度分量的日极值时变曲线,对年极值样本建立最大熵模型并确定百年重现期下的温度分量代表值。研究结果表明:磁浮箱梁均匀温度呈现以年为周期的季节性变化,竖向等效线性温差受太阳辐射影响,呈夏季高、冬季低的特点;百年重现期下磁浮箱梁均匀温度和竖向等效线性温差的代表值分别为44.4℃、8.6℃。该方法结合GBRT算法和最大熵极值模型,解决了极值分析中年极值样本不足的问题,结果可为磁浮轨道梁的温度变形设计提供参考。
Unlike conventional road and railway bridges,the temperature field of maglev track beams is affected by the surrounding magnetic field and the heating of functional components.Given the current absence of explicit regulations in Chinese standards regarding extreme values for the uniform temperature and temperature difference component of maglev track beams,studying representative values for temperature components with probabilistic assurance holds significant importance.Based on the on-site temperature monitoring data collected over one year for a concrete box girder of the Changsha maglev operational line,this study analyzed the time-varying characteristics of the uniform temperature component and vertical equivalent linear temperature difference component of the girder.The Univariate linear regression(ULR)and Gradient Boosting Regression Tree(GBRT)methods were separately used to establish predictive models for the temperature components based on meteorological factors.Comparisons showed that the GBRT method not only improved the prediction accuracy of the uniform temperature but also overcame the limitation of ULR in predicting the vertical equivalent linear temperature difference.Subsequently,the long-term meteorological data from 1960 to 2019 were input into the established GBRT predictive models,obtaining the 60-year long-term forecast values for temperature components of the maglev box girder.After organization,Samples of daily extreme values and annual extreme values of the uniform temperature and vertical equivalent linear temperature difference were obtained.Combining the Maximum entropy extreme value model with the Fourier function,the time-varying curves of daily extreme values for each temperature component were derived.The Maximum entropy extreme value models were established for the annual extreme value samples,determining representative values for temperature components under a hundred-year return period.The research findings are as follows:the uniform temperature of the maglev box girder exhibits
作者
戴公连
罗泽江
王芬
何昌林
张昂
DAI Gonglian;LUO Zejiang;WANG Fen;HE Changlin;ZHANG Ang(School of Civil Engineering,Central South University,Changsha 410075,China;National Engineering Research Center of High-speed Railway Construction Technology,Changsha 410075,China)
出处
《铁道科学与工程学报》
EI
CAS
CSCD
北大核心
2024年第6期2359-2370,共12页
Journal of Railway Science and Engineering
基金
国家自然科学基金资助项目(51708560)
中国铁路总公司科技研究开发计划重点课题(2017G006-N)
湖南省自然科学基金资助项目(2018JJ3672)。
关键词
磁悬浮轨道梁
现场试验
均匀温度
等效线性温差
GBRT算法
最大熵模型
极值分析
maglev track beam
field test
uniform temperature
equivalent linear temperature difference
GBRT algorithm
maximum entropy
extreme value analysis
