考虑施工不确定性的斜拉桥张拉方案优化

Study of the backfill effect on tunnel lining segments property

为了有效解决在斜拉桥施工中各种不确定因素的影响,合理制定索力张拉方案.本文提出了一种基于统计回归和非线性优化的结构可靠度优化方法.利用统计回归拟合得到结构可靠度指标与设计变量之间的拟合函数关系,形成约束函数和目标函数,建立优化模型.为提高计算效率,在拟合之前对影响结果的随机变量进行了可靠度分析和灵敏度分析,筛选出拟合所需的最小样本容量和最小随机随机变量个数.通过对一个刚架模型的可靠度分析,比较了BP神经网络响应面法和多项式响应面法两种拟合方法的优劣性,并确定使用BP神经网络响应面法进行函数拟合,同时验证本文所述方法的可行性.并将该优化方法应用于一座带有水平索的双斜塔斜拉桥,通过优化张拉方案使边墩支座的脱空概率减小至可接受的范围.

In order to effectively solve the influence of various uncertainties in the construction of cable-stayed bridge,the cable tension plan is reasonably developed.A reliability optimization method based on statistical regression and nonlinear optimization is proposed in this paper.The fitting function relationship between structural reliability index and design variables is obtained by statistical regression fitting,forming constraint function and objective function,and establishing optimization model.In order to improve the computation efficiency,reliability analysis and sensitivity analysis were carried out before the fitting of random variables,and the minimum sample size and minimum random variables needed for fitting were screened out.Based on a frame model of reliability analysis,the advantages and disadvantages of two kinds of surface fitting method and polynomial response surface method BP neural network response were compared and identified using BP neural network response surface fitting.Meanwhile,the feasibility of the method described in this paper is proved.The method presented in this paper is applied to a double tower cable-stayed bridge with horizontal cables.By optimizing the cable tension plan,the probability of unseating of the side pier is reduced to acceptable range.