基于RSM的连续刚构桥预应力损失不确定性分析  

Uncertainty analysis of prestressing loss of continuous rigid frame bridge based on response surface method

作　　者：周建庭[1] 辛景舟 王宇[1] 李志刚[1] 

机构地区：[1]重庆交通大学土木工程学院,重庆400074

出　　处：《江苏大学学报（自然科学版）》2016年第1期103-109,共7页Journal of Jiangsu University：Natural Science Edition

基　　金：国家杰出青年科学基金资助项目(51425801);国家自然科学基金资助项目(51278512)

摘　　要：针对现行规范计算得到的预应力损失与工程实际存在偏差的问题,考虑施工因素、材料性能和环境条件的随机性对预应力损失的影响,将响应面法引入到预应力损失的预测中.以某特大跨连续刚构桥为工程背景,选择建设期随机误差较为常见的8个主要结构参数,进行Bucher试验设计,构造不含交叉项的二次多项式响应面模型,将预应力损失与各随机参数之间的复杂隐性关系通过显式函数近似地表达出来;基于逐步回归进行基函数的显著性分析与筛选,运用F检验和修正的决定系数进行响应面模型的显著性和精度检验;同时,采用秩相关系数,分析各参数对预应力损失的敏感性;最后,结合Monte Carlo抽样技术,对预应力损失进行概率意义上的预测,得到一定置信概率下的预应力损失区间.The prestressing loss calculated by the current code of China can＇t fit actual engineering ideally. To solve the problem, with the considerations of construction factor, materials performance and environment condition, the response surface method was employed for uncertainty analysis of prestressing loss of large span continuous rigid frame bridge. Taking a continuous rigid frame bridge as project background, eight main structure parameters and Bucher design method were adopted to conduct the experimental design with three levels. The response surface models without influence of cross-terms were established to express the complex implicit relationship between prestressing loss and random parameters by approximate explicit functions. The significance test and the accuracy test on response surface models were carried out using F test and corrected determination coefficient. The significance analysis of basis function based on stepwise regression and the sensitivity analysis of parameters were conducted by the Spearman rank correlation coefficient. The prediction of prestressing loss in the sense of probability was realized by Monte Carlo sampling analysis.

关 键 词：预应力损失 不确定性分析 响应面法 Bucher设计 秩相关系数 MONTE Carlo抽样 

分 类 号：U448.27[建筑科学—桥梁与隧道工程]