摘要
为了解决刚构-连续梁桥施工过程中的风险问题,提出了一种神经网络与有限元结合的定量风险分析方法(F-R-M法)。以八盘峡黄河特大桥为工程背景建立有限元模型,运用参数敏感性分析确定出最不利施工阶段的结构强度主要风险随机变量,然后以蒙特卡洛原理为基础,通过径向基(RBF)神经网络计算结构失效概率,最后对桥梁施工风险进行定量分析。结果表明:该桥在最大悬臂状态下,结构应力仿真次数超过1000万次时,梁体在最大悬臂端应力失效概率仅约为1.17×10-5,失效概率极低。基于参数敏感性的F-R-M法可以快速、较高精度地对高速铁路刚构-连续梁桥施工期结构强度风险进行定量分析。
In order to solve the risk problem during the construction process of rigid frame-continuous beam bridge,a quantitative risk analysis method (F-R-M method) combined the neural network with the finite element was proposed.A finite element model was established by taking the Bapanxia Yellow River bridge as the engineering background,and the main risk random variables of structural strength in the most unfavorable construction stage were determined by using the parameter sensitivity analysis.Then the structural failure probability was calculated through the radial basis function (RBF) neural network.Finally,the construction risk of the bridge was quantitatively analyzed.The results showed that when the bridge was under the maximum cantilever state and the number of structural stress simulation was more than 10 million times,the stress failure probability of the beam at the maximum cantilever end was only about 1.17×10 -5 ,and the failure probability was very low.The F-R-M method based on the parameter sensitivity can be used to quantitatively analyze the structural strength risk during the construction process of rigid frame-continuous beam bridge for the high-speed railway rapidly and accurately.
作者
冯莉
樊燕燕
王力
李子奇
路韡
FENG Li;FAN Yanyan;WANG Li;LI Ziqi;LU Wei(School of Civil Engineering,Lanzhou Jiaotong University,Lanzhou Gansu 730070,China;National and Provincial Joint Engineering Laboratory of Road & Bridge Disaster Prevention and Control, Lanzhou Jiaotong University,Lanzhou Gansu 730070,China;School of Civil Engineering,Northwest University for Nationalities,Lanzhou Gansu 730030,China)
出处
《中国安全生产科学技术》
CAS
CSCD
北大核心
2019年第9期135-140,共6页
Journal of Safety Science and Technology
基金
长江学者和创新团队发展计划滚动支持项目(IRT15R29)
关键词
施工风险
刚构-连续梁桥
敏感性分析
RBF神经网络
蒙特卡洛原理
定量分析
construction risk
rigid frame-continuous beam bridge
sensitivity analysis
radial basis function (RBF) neural network
Monte Carlo principle
quantitative analysis
