摘要
为保障大跨度桥梁在施工期间的抗风安全性,以拉林铁路藏木大桥(主跨430 m的中承式钢管混凝土拱桥)为工程背景,提出了适用于工程推广应用的风速超前概率预测方法。即采用数据驱动的变分模态分解(VMD)将原始风速进行多尺度分解,通过特征选取技术重新组合多阶分量,形成仅包括可预测信息和干扰预测精度信息的两组序列;采用ARIMA-GARCH组合模型获得确定性的风速预测结果;最后基于纯随机性检验的循环迭代方法准确获得风速序列中蕴含的不确定性信息,并采用单变量核密度估计方法得到具有概率意义的风速区间预测结果。将所述方法风速预测结果与风速实测值进行对比,结果表明:所述预测方法风速预测精度高、区间带宽窄,95%置信度的上包络值可为桥梁施工期风灾辅助决策、防灾减灾提供更为可靠的参考依据。
Based on the engineering practice of Zangmu Bridge on Lhasa-Nyingchi Railway(a half-through concrete-filled steel tubular arch brdge spanning 430 m),a probabilistic wind speed prediction method is proposed,aiming to ensure the wind-resistant safety of long-span bridge during construction.In this method,the original wind speed series can be decomposed into several subseries by data-driven variational mode decomposition(VMD)and the decomposed multi-level subseries are reassembled by feature selection technique to be two series,including the predictable information series and prediction accuracy interference series.The ARIMA-GARCH model was established to obtain deterministic wind speed prediction results.Finally,the uncertainty information embedded in the original wind speed series is accurately determined by a cyclic iteration method based on pure randomness test,and a probability interval prediction results are acquired by univariate kernel density estimation.The wind speed prediction values gained by the proposed method were compared with the measured values.It is shown that the proposed method can predict wind speed with high accuracy,the bandwidth is narrow,and thus with a confidence level of 95%,the upper envelopes can provide more reliable references to aid wind disaster decision-making,disaster prevention and mitigation during bridge construction period.
作者
苏延文
曾永平
SU Yan-wen;ZENG Yong-ping(China Railway Eryuan Engineering Group Co. , Ltd. , Chengdu 610031, China)
出处
《桥梁建设》
EI
CSCD
北大核心
2021年第4期45-52,共8页
Bridge Construction
基金
四川省科技计划资助项目(2018GZ0052)。
关键词
铁路桥
钢管混凝土拱桥
风速概率预测
ARIMA-GARCH组合模型
随机性检验
单变量核密度估计
风灾辅助决策
railway bridge
concrete-filled steel tubular arch bridge
probabilistic wind speed prediction
ARIMA-GARCH model
randomness test
univariate kernel density estimation
wind disaster decision-making support
